The Origin and Persistence of State Fragility In: Burundi

APRIL 2018
The origin and

persistence of state
fragility in Burundi
Janvier D. Nkurunziza
Abstract
State fragility in Burundi has been a cause, and consequence, of the country’s
political instability. Since independence, Burundi has endured six episodes
of civil war, two major foiled coup d’états, and five coup d’états that have led
to regime change. The root cause of state fragility is traced back to divisive
practices introduced by the colonial power, which have since been perpetuated
by post-colonial elites. This political volatility has generated persistent cycles of
violence, resulting in the collapse of the country’s institutions and economy, even
after the negotiation of the Arusha Agreement. This has led to mass migration of
Burundi’s people and the emergence of a large refugee population, dispersed
among neighbouring states and far away. Therefore, state fragility in Burundi is
first and foremost the result of the strategies and policies of its political leaders,
who are motivated by personal interests. Political capture calls into question the
legitimacy of those in power, who feed state fragility through rent extraction,
corruption, and mismanagement. This has had vast economic consequences,
including slow growth, an underdeveloped private sector, an unstable investment
landscape, and severe financial constraints. For reconciliation to be achieved,
justice needs to be afforded to those who have encountered repression from the
state, thereby breaking the cycle of violence. What’s more, Burundi needs strong
and long-term engagement of the international community for the successful
implementation of reforms, as well as the provision of technical and financial
resources, to embark on a prosperous and peaceful path.
Janvier D. Nkurunziza
United Nations Conference on Trade and Development (UNCTAD)
[email protected]
About the commission

The LSE-Oxford Commission on State Fragility, Growth and Development was
launched in March 2017 to guide policy to address state fragility.
The commission, established under the auspices of the International Growth

Centre, is sponsored by LSE and University of Oxford’s Blavatnik School of
Government. It is funded from the LSE KEI Fund and the British Academy’s
Sustainable Development Programme through the Global Challenges Research
Fund.
Front page photo: SuSanA | CC BY-SA (with modification)
2 The origin and persistence of state fragility in Burundi

Contents
Introduction 1
Origins of state fragility in Burundi 7
Regional and international dimensions of 13

fragility
Strategies and policy choices that feed state 19

fragility
Economic consequences of state fragility 23
Conclusion 31
References 34

Introduction
State fragility in Burundi
1
State fragility in Burundi displays all five dimensions that generally characterise
fragility:
1. For most of the country’s post-colonial period, the country has been
ruled by illegitimate regimes that came to power through coups d’états or
constitutional fraud. Since independence in 1962, Burundi has recorded seven
major coups d’états, of which five led to regime change.
2. The state has not been effective in steering the country through a
development process, as illustrated by recent statistics listing Burundi as the
second poorest country in the world.1
3. As political elites have primarily focused on controlling the state, and

capturing the associated “rents to sovereignty” (Nkurunziza and Ngaruko,
2008), the private sector has been neglected and remains underdeveloped. 2
4. State fragility in Burundi has been narrowly associated with insecurity due
to a long cycle of violent conflict. Over the last 50 years, Burundi has gone
through six episodes of civil war. The state has not only been unable to keep
peace, but most of the violence has been caused by the state itself, because
of acts of “cold blood by the deliberate fiat of holders of despotic political
power” (Toynbee, 1969).
5. The combination of these fragilities has made Burundi highly vulnerable to

political and economic shocks.
1 In terms of the country’s GDP per capita measured in Purchasing Power Parities (current
international United States (US) dollars) (World Bank, International Comparison Program database,
2017. Accessible: http://data.worldbank.org/indicator/NY.GDP.PCAP.PP.CD?view=chart)
2 The concept “rents to sovereignty” refers to the spoils associated with the control of political
power, including the capture of part of foreign aid, international borrowing, and tax revenue;
allocation of public investment and public employment to benefit members of a specific group; and
organization of the economy with a view to generating rents that are enjoyed by political elites at the
expense of those not affiliated with them. This attitude creates inter-group tensions and grievances
that eventually lead to political violence.

This paper analyses the origins, manifestations, and consequences of
Burundi’s state fragility, with a focus on the country’s recurring political instability.
State fragility has been both a cause and consequence of political instability.
A fragile state has little capacity to absorb the shocks that lead to political
instability.
Instability and violence

In 1993, the assassination of a newly elected president through a coup d’état led
to the longest episode of Burundi’s civil war. In contrast, in the 1980s and 1990s,
the presidents of Burkina Faso and Niger, Sankara and Mainassara, respectively,
were assassinated by members of the army who subsequently took power, but
these shocks did not result in civil wars. This suggests that Burundi was less able
to absorb this political shock than Burkina Faso and Niger. Conversely, recurring
instability as observed in Burundi destroys or drastically weakens the state,
making it extremely fragile.
The recurrence of political violence in Burundi should not be understood as
a succession of several civil wars. Cyclical political violence has been the result
of the failure or unwillingness of the state and political elites to address the root
causes of violence since the early years of independence.
Violence has been the outcome of asymmetric fights between political elites
from the two major groups, the Hutus and Tutsis.3 Each group fights to capture
the state and its spoils, as discussed in some detail below. The group in power
has a clear advantage over the other as it uses all state resources to annihilate
its challenger. This explains why state repressions have been so ferocious.
Criminal acts are committed with the disproportionate use of force, while those
controlling state institutions have no incentive to punish their members involved
in the illegal killings, perpetuating impunity.
This profile of political instability and violence that is immune to accountability
means that each conflict leaves many victims with grievances that remain
unaddressed. This sows the seeds of future conflict, as the victims find that it
is only by ousting their repressors that they can get justice. As each episode of
violence feeds into the next one, Burundi’s political instability can be understood
as due to one unresolved conflict that manifests itself in cyclical episodes of
extreme violence.
3 The concept of ethnicity in the context of Burundi differs from its empirical definition as used in
other contexts (e.g. Horowitz, 1985). Burundi’s population is divided into four groups: the Hutu, Tutsi,
Ganwa and Twa. The Hutus are thought to represent most of the population, even though there are no
credible statistics giving the proportion of each group, followed by the Tutsi group. The Ganwa are
a relatively small group, mostly made up of descendants of the country’s traditional monarchs. The
Twa are another small group that is poorly integrated with the rest of the population. All four groups
have shared, over several centuries of coexistence - apparently since the 11th century - one single
language, Kirundi, one culture, and live mixed in the same geographical areas. Hundreds of years of
coexistence under a common value system forge a common identity. Therefore, using differences
in ethnicity to explain Burundi’s political violence is a tenuous argument, despite its popularity in the
literature on Burundi.

Governance and impunity
This paper will devote special attention to the issue of governance, particularly
the impunity of criminals involved in state crimes. This is an aspect of fragility
that will need to be addressed if Burundi is to emerge out of its fragility trap. In
this light, state fragility is analysed primarily through the lenses of insecurity and
conflict.
The economic implications of state fragility are also explored. Burundi’s failure
to reconstruct its economy and institutions have been the result of choices
made by its leaders over the years. Indeed, contrasting the cases of Burundi and
Rwanda, two countries with comparable initial conditions in terms of history,
culture, social mix, institutions, and economic development, illustrates that
Burundi could have followed a different trajectory had its leaders made the right
choices.
Since 1965, Burundi leaders have been unwilling and, to some degree,
incapable of transforming their country’s institutions, including economic
institutions, in a way that would strengthen the country’s resilience to shocks
and help avert future violence. Every episode of violence is a reminder that
Burundians have been locked into a situation “where the past does not pass”
(Manirakiza, 2002).
The paper proceeds as follows:
1. Section 2 draws from Burundi’s history to help understand the nature and
origins of state fragility.
2. In Section 3, the paper analyses the regional and international dimensions of

state fragility in Burundi.
3. The motives driving policy choices that led to fragility are discussed in
Section 4.
4. Section 5 highlights some major consequences of state fragility.
5. Section 6 concludes with a discussion of the lessons learned from this case
study and makes suggestions on how Burundi can break out of its fragility
trap and rebuild an economy and institutions that would help avert future state
collapses.
Credit: Bruno De Hogues/Getty

Origins of state
fragility in Burundi
This Section draws on the history of Burundi to identify the origin of state fragility.
2
It goes on to briefly discuss the different episodes of civil war, highlighting how
they are inter-related.
Historical origins of state fragility
Colonisation
The history of Burundi does not provide accounts of any serious ethnic conflict
that occurred before the country was colonised at the end of the 19th century.
Instead, during the pre-colonial period, the state is described as well-structured,
centralised and strong, with power controlled by a secular monarchy whose
authority was generally uncontested throughout the country.
The kingdom of Burundi was one of the strongest kingdoms in the African
Great Lakes region for several centuries, until the end of the 19th century when
it became a German colony until the end of the First World War and thereafter, a
Belgian colony. As an illustration of the strength of Burundi’s state, when an army
of Arab slave traders penetrated the country in 1884 in search of slaves, they
were inflicted a humiliating defeat by the country’s army (Gahama, 2001). Hence,
Burundi never experienced slave trade, unlike some other kingdoms in the region.
Belgian colonists’ attempts to subjugate the population of Burundi faced
similar resistance. Several missionaries, who were among the first Europeans to
penetrate into Burundi, were killed (Bonneau, 1949). Therefore, weakening the
traditional state became the Belgian colonists’ modus operandi to stamp their
authority on the country. They undermined the traditional system of governance
by introducing “divide and conquer” policies that broke the secular identity of
the people of Burundi.
Divide and conquer

The Belgian colonists opposed the two major groups, the Hutus and Tutsis,
along what they considered to be ethnic lines. Historians are of the view that
the colonists “racialised” the Hutu and Tutsi categories that had formerly been

considered as social groups (Chrétien, 2000a). This policy not only weakened
traditional state structures but also created, over time, polarisation between the
Hutus and Tutsis. This polarisation has persisted to date.
More specifically, between 1928 and 1934, the Belgian colonists introduced
far-reaching administrative reforms (Gahama, 2001) that favoured the Tutsis,
who were considered as superior and born to rule, at the expense of the Hutus
who were described as backward peasants (Sandrart, 1953:2). For example,
traditionally, the chiefs that were appointed by kings as regional governors were
drawn from the Hutu, Tutsi and Ganwa groups. Colonial administrative reforms
replaced all sitting Hutu chiefs with Tutsis and Ganwa. The proportion of Hutu
chiefs went from 20% in 1929, to zero in 1945 (Reyntjens, 1994).
This interference with traditional leadership practices not only marginalised
the Hutu political elite but also instituted a rigid system of domination of the Hutu
and Twa by the Ganwa and Tutsis. As expected, the policy created resentment
among the Hutus, inducing them to make several unsuccessful attempts to
capture power from the Tutsis and Ganwa by force. In contrast, the Tutsi elite
acted to strengthen and perpetuate the system as it favoured them. Moreover,
the Tutsis used Hutu attempts at capturing power as an excuse to mercilessly
repress them, which enabled the Tutsis to tighten even further their political
control over the country.
Independence
Political splits and incessant fights continued into the early years of
independence, which was established in July 1962. There were two groups, one
called Casablanca Group, favouring pan-Africanism and deep integration of
the continent as the best way forward for Burundi, while the other, the so-called
Monrovia Group, was also for pan-Africanism but not at the expense of national
statehood (Manirakiza and Harroy, 1990).
Political elites in the first few years after independence were so absorbed by
political fights, as well as plotting and counter-plotting, that governments were
highly unstable. Between independence and the end of 1966, no government
was in place for more than one year. This instability eventually led to the first
large-scale political violence in 1965. Amidst these political fights, other aspects
of governance, such as economic development, were of little interest to the
leadership.
While the Belgian colonial authority is attributed with the creation of state
fragility in Burundi, Burundian political elites, particularly those within the Tutsi
group who ruled the country for most of the post-colonial period, perpetuated
it through their leadership. Months before the country’s independence, Prince
Louis Rwagasore, the highly respected national hero who fought for the country’s
independence, was assassinated by political opponents “who seemed to have
acted with the tacit approval of Belgian authorities” (Stapleton, 2017: 67). Within
the Unité pour le Progrès National (UPRONA) party, Rwagasore had been able to
unite Hutus and Tutsis behind his independence project, which was coupled with
a clear development vision. After his assassination in October 1961, Burundi went
through a period of turmoil, as Hutu and Tutsi political leaders were locked into
leadership disputes.

Political volatility
In this context of political volatility, in October 1965, a group of Hutu leaders,
frustrated by what they perceived as their marginalisation, attempted to topple
the government but failed.41Then, under the instigation of some Hutu leaders,
Hutu peasants in Muramvya province killed several hundred Tutsis. In response,
the Tutsi-led army used this as an excuse to purge the country’s leadership of its
most influential Hutu members. The army killed about 5000 Hutus in Muramvya
to avenge the Tutsi killings (Stapleton, 2017). This marked the first large-scale
political killings in the country.
In November 1966, with the Hutu leadership having been decimated a year
earlier, a group of Tutsi officers mostly from the southern province of Bururi,
carried out a coup d’état, overthrowing the centuries-old monarchy. This spelt
the end of the post-independence political system that had to some degree
included Hutu and Tutsi leaders from all regions of the country, albeit in unequal
proportions.
The small Tutsi élite from Bururi installed an illegitimate system concentrating
powers - political, military and economic - in their hands. The marginalisation of
Hutus and, to some extent, non-Bururi Tutsis, polarised the country even further.
The 1966 coup d’état turned out to be the first of a long series. Others that led to
regime change were staged in 1976, 1987, 1993, and 1996. Most recently, in May
2015, a group of officers also attempted but failed to seize power by force. A
bloody repression ensued, plunging Burundi into its current turmoil.
The most enduring legacy of political instability in Burundi is that crimes and
other injustices committed by state agents and institutions since the 1960s have
never been acknowledged, let alone punished. The rule of law has been crippled,
allowing political elites to still engage in bloody fights to control the state, further
deepening state fragility. The failure of successive governments to acknowledge
past state crimes has alienated large sections of the population, particularly the
victims of such crimes.
The Arusha Agreement

The first attempt ever undertaken to find a lasting solution to Burundi’s conflict
has so far been the Arusha negotiations and their outcome, the Arusha Peace
and Reconciliation Agreement for Burundi—The Arusha Agreement. The
negotiations brought together participants from a large spectrum including
political parties, civil society, the army, the government, and the international
community. They were organised in the Tanzanian town of Arusha, resulting in
the Peace Agreement that was signed in August 2000, and it entered into force in
November 2001.
The Agreement succeeded in bringing an end to the longest episode of civil
war, between 1993 and 2003. It also allowed the current political Hutu elite,
4 Hutus in Burundi were of the view that their population majority should guarantee them de facto
control of power, as occurred in Rwanda in 1959, which served as their model of political leadership.
In contrast, the Tutsi elite in Burundi considered the Rwandan 1959 Social Revolution as an anti-
model to be prevented at any cost. Hence, all attempts by Hutus to take power by force in Burundi
were mercilessly repressed, fueling constant tensions between political elites in the two groups
(Nkurunziza, 2016a).

coming from the group that had long been persecuted by former regimes, to
come to power after a long bloody war. In this regard, the political settlement
was radical in the sense that it brought an end to the domination of Burundi’s
politics by the Tutsi elite since the 1930s. Considering the political imbalances
that the Agreement sought to redress, it seemed to be tilted towards one group,
that of the traditionally marginalised.
State collapse
The new political elite found it impossible to resist the temptation of absolute
power and, in some cases, the urge to avenge past humiliations, partly because
of their limited political experience. Power sharing, as envisioned in the Arusha
Agreement, all but collapsed as the new political leadership failed to respect the
power sharing provisions reached through tough negotiations. The unravelling
of the Arusha Agreement entrenched the fragility of the state, leading to its
collapse.
Since April 2015, Burundi has been in a state of political violence that has
led to about 1,200 deaths and rendered more than 400,000 people refugees.
The economy has been hit very hard, contracting by 4% in 2015 and 1% in 2016.
Projections over the medium term are bleak. Between 2018 and 2022, gross
domestic product (GDP) growth is expected to vary between 0.0% and 0.5%
(IMF, 2017).
The failure of Burundi’s political elites to consider the implementation of the
Arusha Agreement as their best chance for long-term peace and stability seems
to have returned Burundi to the old path of cyclical violence. It is in this context
that the current debate about bringing back stability to Burundi systematically
refers to the return to the text and spirit of the Arusha Agreement.
Burundi’s internal armed conflict
State repression
The first episode of civil war in 1965 was followed by others in 1972, 1988,
1993-2003, and from 2015 to date (Table 1). With the exception of the 1993-2003
episode, almost all the casualties of the civil war have been due to repression by
state institutions.
The modalities have been the same. Leaders of the group out of power,
often the Hutu leadership, try to capture it by force, prompting those in power,
often the Tutsi leadership, to unleash maximum violence against the Hutu group,
irrespective of individual responsibilities. State institutions such as the army,
police and intelligence service carry out these repressions. Given the asymmetry
of the forces on the ground, the number of victims of political repressions dwarfs
the casualties caused by the initial attacks.
In 1972, a group of Hutu politicians, frustrated by the repression endured by
their group in 1965 and their political marginalisation, organised a rebellion with
the intention of overthrowing the Tutsi government. They killed thousands of
unarmed Tutsi civilians in the Southern province of Bururi, the same region where
the political elite of the time was from.

Table 1: Characteristics of Burundi’s internal armed conflict
Characteristic 1965 1972 1988 1991 1993–2003 2015 to date
Duration (months) 2 4 2 1 120 27
Deaths (thousands) 5 200 15 1–3 300 1.2
Refugees (thousands) 0 300 50 38 687 416
Deaths & refugees 0.2 14.0 1.3 0.7 17.1 3.8

to population (%)
Years from – 6 16 3 2 12
previous conflict
Provinces affected Muramvya Whole Ngozi, Cibitoke, Whole Whole country

country Kirundo Bubanza, country but mostly
Bujumbura Bujumbura
Source: Nkurunziza (2016b); Data in the last column (2015 to date) reflects the latest available
information. Data on total deaths is from ACLED (2016), effective April 2016; information on the
number of refugees is from UNHCR, at (https://data2.unhcr.org/en/situations/burundi), effective
July 2017
In retaliation, Bururi-dominated government forces engaged in ferocious

killings of Hutus, not just in Bururi, where the new-born rebellion operated,
but across the whole country. As a result, about 200,000 innocents, mostly
Hutus, were killed. By targeting almost all Hutus with some level of education,
the objective, it seems, was to eradicate the Hutu political movement. Another
300,000 people were forced to flee the country as refugees. The same pattern
was observed in 1988 and 1991, albeit to a more limited extent.
Cycles of violence
The 1972 episode affected the country so deeply that it became a defining
moment for future political violence. Many subsequent Hutu leaders were either
former refugees who had come back to Burundi after leaving the country in
1972 as young adults, or children of the Hutus who were killed in 1972. The
most prominent members of this group include the current president, Pierre
Nkurunziza, who lost his father in the 1972 political repression.5 Melchior
Ndadaye, the first civilian president elected democratically in 1993, was a former
1972 refugee. Two of his successors, Sylvestre Ntibantunganya, and Domitien
Ndayizeye, were also former 1972 refugees.
In 1993, Burundi organised the first ever democratic elections since the early
1960s. Melchior Ndadaye, a former Hutu refugee, was democratically elected. In
October of the same year, after just three months in power, he was assassinated
by members of the Tutsi-dominated army, triggering the longest episode of the
civil war.
The assassination of the Hutu president triggered a massacre of Tutsis by their
Hutu neighbours throughout the country. The reaction by members of the Hutu
5 Later, during the 1993-2003 conflict, he escaped two assassination attempts before deciding to
join the rebellion.

group was so swift that some have argued that it had been planned by the Hutu
leadership. As in the past, the Tutsi-dominated army resorted to repression of the
Hutus, but this did not end the violence. About 17% of the population were either
killed or fled to become refugees.
The Hutu leadership considered that the October assassination of the Hutu
president was an attempt by the old Tutsi elite to recapture the power they had just
lost through the elections. For the first time, the Hutu group organised a strong
rebellion that waged war for ten years. Many of its military and political leaders
were orphans of the 1972 government repressions.
The pursuit of peace

As the war dragged on, the international community brought together the warring
parties in negotiations that culminated in the signing of the Arusha Agreement.
Elections were organised in 2005, and won by the Conseil National pour la
Défence de la Démocratie-Forces pour la Défence de la Démocratie (CNDD-
FDD), dominated by the Hutu rebel movement.
The transfer of power to Hutu leaders, the victims of repressions by past
Tutsi-dominated regimes, was greeted by many Burundians as ushering in a new
political era where political violence would have no place. However, the recent
past has shown that the current political elite has adopted the same practices
that they decried. Many independent reports have pointed to the role of state
security institutions in many acts of violence since April 2015, when the current
president decided to seek a third term in office, a move considered by many
as illegal. At least 1,200 people have been killed, mostly by state agents, while
400,000 people have left the country to seek refuge abroad. Scores have been
tortured and thousands of Burundians are languishing in prison.
While non-state actors involved in violence are often killed during state
repressions, very little has been done to identify and prosecute the state actors
involved in illegal acts of violence. This has constantly been the biggest problem
hindering a peaceful resolution of the cycle of violence.
Indeed, “more serious problems arise when the body responsible for
threatening and causing death is - or is in complicity with - a State itself”
(Whitaker, 1985, paragraph 18). The fact that state agents can kill and torture
with impunity sends the message that the law applies only to the powerless.
As impunity persists, today’s victims will likely seek taking justice into their own
hands in the future, perpetuating the conflict.

Regional and
international
dimensions of
fragility
State fragility in Burundi has had important regional and international
3
repercussions. Three different aspects are discussed: the flow of refugees
to neighbouring countries and beyond; regional initiatives, as well as the
involvement of the African Union and the United Nations in finding a solution to
Burundi’s conflict.
Refugees
As Table 1 illustrates, the conflict in Burundi has generated a cumulative number
of 1.5 million refugees. These have been concentrated in neighbouring countries,
particularly Tanzania. It alone housed more than half a million refugees from
Burundi at the height of the refugee crisis. Other countries that have hosted large
numbers of Burundian refugees at some stage of the conflict include Rwanda
- particularly in relation to the 1972 and 2015 conflicts - and the Democratic
Republic of Congo (DRC).
Starting with the 1993 conflict, a sizable number of Burundian refugees have
also reached faraway places, including Europe (e.g. Belgium, Netherlands,
Switzerland), and Canada, most prominently. Burundian refugees are now found
in every region of the world.
Migration
The refugee question has been a sensitive political issue in the context of
relations between Burundi and its neighbours. In the 1970s and 1980s, Rwanda
hosted a large number of Hutu refugees from Burundi following the 1972 state
repression. Since the Social Revolution in 1959, up until the genocide in 1994,
Rwanda was run by politicians from the Hutu group. Burundi, in contrast, was led

by Tutsi politicians from independence until 2003, save for the period between
1993 and 1996. As a result, the Hutus who were persecuted in Burundi until
1993 considered the regime in Rwanda as friendly, so many fled to this country,
particularly in 1972.
Conversely, among the Tutsis, who fled Rwanda from 1959 to the genocide
in 1994, many migrated to Burundi, where the Tutsi regimes were considered
friendly. Both countries perceived the presence of large numbers of their
refugees - Burundian Hutus in Rwanda and Rwandan Tutsis in Burundi - as a
political threat, generating tensions between the two countries. In the 1980s, the
refugee issue pushed Burundi and Rwanda to the brink of war.
Political tensions
Political fortunes in the two countries have reversed. From 1994 onwards,
Rwanda has been led by members of the Tutsi group, while a Hutu elite has been
at the helm of Burundian politics since 2005. Relationships between the two
regimes were friendly until 2015, when scores of Burundian refugees started to
flee to Rwanda after the eruption of violence in April 2015. Several opposition
leaders, army officers, civil society leaders, journalists, business people,
professionals, as well as normal citizens went to Rwanda to seek refuge.
Allegedly, some officers who were involved in a May 2015 foiled coup in
Burundi fled to Rwanda. As a result, Burundi has been openly accusing Rwanda
of harbouring its enemies and training them to overthrow the current government
in Burundi, which Rwanda has vehemently denied. Political tensions are so high
that trade between the two countries has been suspended.
The DRC has also been drawn into Burundi’s conflicts. In 1972 for example,
the group that attacked the Southern part of Burundi came from DRC, then called
Zaire. Until today, the unruly and mountainous Eastern part of DRC is said to host
Burundian rebels, alongside many other rebellions. Consequently, the Burundian
army has been crossing the border into DRC in pursuit of these groups, causing
frictions between Burundian and DRC governments.
Tanzania
Tanzania has traditionally hosted the largest number of Burundian refugees over
the decades of conflict. The largest number of refugees went to Tanzania in 1972
and during the 1993-2003 conflict. Interestingly, tensions between Burundi and
Tanzania over the refugee issue never were as high, as compared to the case of
Rwanda, despite Tanzania hosting more refugees than Rwanda.
The 1993-2003 rebellion that later came to power in Burundi used rear
bases in Western Tanzania, particularly in Burundian refugee camps. The Tutsi-
dominated government of Burundi had long suspected Tanzania of being pro-
Hutu, especially during the Arusha negotiations that were facilitated by Julius
Nyerere, former president of Tanzania (Boshoff, 2010). Despite this suspicion,
the negotiations took place and were concluded in Tanzania, partly under the
facilitation of a Tanzanian political leader (see discussion below).

The international community
Beyond the African region, the refugee issue associated with the current political
crisis has caused serious diplomatic frictions between Burundi and Belgium, the
former colonial power. The latter has also attracted the wrath of the government
of Burundi for hosting a number of prominent refugees from Burundi.
Diplomatic relations between the two countries reached an all-time low when
both governments recalled their respective ambassadors. The government of
Burundi accuses Belgium of being behind the drive to freeze international aid to
the country. Being the main donor to Burundi in terms of bilateral aid, the freezing
of Belgian direct aid to the government of Burundi has increased the country’s
state and economic fragility.
Beyond Europe, Canada has received large numbers of refugees from
Burundi, predominantly starting with the 1993-2003 crisis. In light of the
acceptance rates of applications from Burundi citizens for refugee status in
Canada in 2016 being high, combined with a high volume of cases, Canada
included Burundi in the group of countries for which the processing of cases is
expedited (as of 1 January 2018). Burundi joins Afghanistan, Egypt, Eritrea, Iraq,
Syria, and Yemen.7
African-led peace negotiations
The Arusha negotiations

The conflict in Burundi has involved prominent African leaders who have invested
their personal and political capital into helping to find a permanent solution. As
stated earlier, the Arusha negotiations were the first serious attempt by Africans
to find a lasting solution to Burundi’s political problem. Julius Nyerere, former
president of Tanzania and a highly respected international figure, facilitated the
negotiations until his death in 1999. Thereafter, another highly respected African
statesman, Nelson Mandela, former president of South Africa, took over the
negotiation process. 8
This exceptionally high caliber of chief facilitators to the negotiations drew
the attention of the international community to the crisis in Burundi. At the signing
of the Arusha Agreement in August 2000, international personalities, including
Kofi Anan, Secretary-General of the United Nations, were present to show their
support.
The Arusha Agreement had shortcomings, as highlighted in Nkurunziza
7 Immigration and Refugee Board of Canada (2018). Expedited Processing of Refugee Claims by
the Refugee Protection Division. Accessible: http://www.irb-cisr.gc.ca/Eng/NewsNouv/info/Pages/
expedited-processing-refugee-claims.aspx
8 For a detailed analysis of these negotiations, including the reasons why they took place, as well as
their shortcomings, see Nkurunziza (2016a). It is worth noting that unlike the previous four episodes
of the conflict in Burundi, the 1993-2003 episode overlapped with the genocide in neighbouring
Rwanda, where the international community failed to act to stop the killings. It is possible that
collective guilt might have prompted the international community to do something in the case of
Burundi.

(2016a). However, it formed a basis on which a credible and inclusive political
system in Burundi could be built. In its opening paragraphs, the Arusha
Agreement acknowledges that its signatories are “Aware of the fact that peace,
stability, justice, the rule of law, national reconciliation, unity and development
are the major aspirations of the people of Burundi” and that the signatories are
“Determined to put aside our differences in all their manifestations in order to
promote the factors that are common to us and which unite us, and to work
together for the realization of the higher interests of the people of Burundi”.9
National Truth and Reconciliation Commission

In the Arusha Agreement, national reconciliation is the cornerstone of
peacebuilding. It calls for the establishment of a National Truth and
Reconciliation Commission with the following main mandate: investigate the
killings and all human rights abuses committed; arbitration and reconciliation;
clarification of history to offer one reading of Burundi’s history (Protocol I,
Article 8). The period to be covered is from independence on 1 July 1962, to 4
December 2008 (Protocol I, Article 6).
Implementation
The Arusha Agreement has valuable provisions if implemented objectively. In his
analysis, Nkurunziza (2016a) shows that the implementation has been selective,
favouring provisions that are in the interest of the different ruling elites. Some
of the most important provisions of the Agreement, such as the creation and
facilitation of a credible National Truth and Reconciliation Commission, were
never taken seriously.
A possible reason is that none of the successive regimes entrusted with the
implementation of the Arusha Agreement had interest in revisiting a past where
their own crimes might come to light. All the elites that have ruled the country
seem to benefit from “historical silences” (Lemarchand, 2002). The decision
of the ruling party in 2015 to field the same candidate for a third presidential
term, which seems to violate the Arusha agreement, is another illustration of the
selective approach to the implementation of the Arusha Agreement.
International involvement
Muted reactions
Overall, the involvement of the international community in addressing Burundi’s
state fragility has been wanting. The 1965 conflict was considered as a simple
local conflict, attracting little interest from the international community. In
9 Arusha Peace and Reconciliation Agreement for Burundi (2000). Accessible: http://peacemaker.
un.org/sites/peacemaker.un.org/files/BI_000828_Arusha%20Peace%20and%20Reconciliation%20
Agreement%20for%20Burundi.pdf

1972, despite the large-scale nature of the killings that decimated the Hutu
intelligentsia, the reaction of the international community remained muted.10
The international community seems reluctant to act forcefully in extreme
cases of violence, as observed in Rwanda in 1994. Moreover, in Burundi in 1972,
the then Tutsi-led government that was responsible for the massacres engaged in
a diplomatic crusade across the world to thwart any serious condemnation of the
violence and potential intervention by the international community.
Commission of inquiry
A muted response to grave violations of human rights in Burundi was not limited
to 1972. After the assassination of the Hutu president in 1993 that triggered the
1993-2003 civil war, it took two years (28 August 1995) for the United Nations
Security Council (UNSC) to approve Resolution 1012, requesting the Secretary-
General to establish an international commission of inquiry to:
“establish the facts relating to the assassination of the President of Burundi
on 21 October 1993, the massacres and other related serious acts of violence
which followed” and to “ recommend measures of a legal, political and
administrative nature, as appropriate, after consultation with the Government
of Burundi, and measures with regard to the bringing to justice of persons
responsible for those acts, to prevent any repetition of deeds similar to those
investigated by the commission and, in general, to eradicate impunity and
promote national reconciliation in Burundi” (USIP, 2004).
The commission produced a damning report. High-ranking officers from the
army and other senior politicians were identified by name as being involved in
the assassination of the president and fueling the violence that ensued. These
individuals refused to cooperate with the commission and, in the end, none were
held to account.
The United Nations

The UNSC’s failure to hold to account those responsible for such serious
crimes provoked doubts about the willingness of the international community to
help bring justice to Burundi. The United Nations helped with a peacekeeping
mission that supported the transition to the current regime. This was useful
but insufficient in realizing lasting peace. As the point is made below, Burundi
needs a long and well-funded international presence that should oversee the
establishment of a more resilient political and economic system.
Recently, since the eruption of violence in April 2015, the United Nations has
debated the case of Burundi several times, both in the Security Council and the
Human Rights Council. Fact-finding missions have been sent to Burundi and
neighbouring countries. They have all come up with the same results, attesting
to grave and widespread violations of human rights committed by government
forces and, to a lesser extent, opposition groups.
10 Later, the 1965 and 1972 episodes of violence were acknowledged as cases of genocide
(Whitaker, 1985).

International diplomacy
The situation in 2015 was judged to be so grave that the African Union’s Peace
and Security Council voted to send a 5,000-strong peacekeeping force into
Burundi in December 2015. The government of Burundi opposed this move and
in January 2016, the African Union backtracked on its pledge. In July 2016, the
UNSC voted to send a modest police force of 288 officers to Burundi to monitor
the situation. The government of Burundi refused to cooperate, claiming that its
own national police force had the situation under control.
The exercise of international diplomacy, at least as recently applied
to Burundi, has failed partly due to its internal contradiction. Why should
international intervention be subjected to the agreement from the same “holders
of despotic political power” (Toynbee, 1969) who are committing the very crimes
the international community seeks to end? What incentive does such a regime
have to allow into its territory a force that is perceived as coming to destroy it?
Distressed populations in Burundi have learned from this experience that
the ‘Duty to Act’ and ‘Commission by Omission’ principles of international
criminal law cannot be considered as a reliable source of justice. On the other
hand, political powerholders in the country have learned that they can defy the
international community by easily thwarting international intervention. This bodes
ill for peacebuilding in Burundi.
Credit: SuSanA | CC BY-SA

Strategies and
policy choices that
feed state fragility
As discussed above, political and economic fragility in Burundi are not inevitable
4
consequences of the country’s destiny. On the contrary, they are results of
choices made by ruling elites who have favoured policies that serve their
personal interests rather than those of the population at large. This section
illustrates how state institutions have systematically been captured to promote
and protect the interests of the ruling elites, deepening state fragility.
Civil service and security fragility
Political capture
The role of the civil service is to support the government in the efficient design
and implementation of policies. In Burundi, however, to varying degrees, different
regimes have used employment in the civil service as an opportunity to reward
their faithful followers with positions of influence. Moreover, as an extension of
executive power, the civil service has been organised in such a way that power
remains in the hands of the “right” people, who would not use their positions to
challenge the government of the day.
From the mid-1960s until the early 2000s, most positions of influence within
the civil service were in the hands of people from the Southern province of
Bururi, the cradle of the political leadership of that period. This influenced wealth
distribution, given that the state has traditionally been the major employer in
Burundi (Ngaruko and Nkurunziza, 2000). The allocation of public goods was also
skewed in favour of this region. For example, most of the country’s best schools
and best teachers were concentrated in Bururi, an otherwise remote region
(Nkurunziza, 2012). Privileged access to education in this region had a long-term
effect on wealth, since access to quality education leads to better jobs with
higher pay (Pritchett and Filmer, 1999).
The coming to power in 2005, of a traditionally marginalised Hutu elite,

does not seem to have fundamentally changed the practice of political capture
initiated in the mid-1960s. Influential positions in the public service, the army,
police and intelligence services, have changed hands in favour of members of
the CNDD-FDD, the party in power. Many experienced and competent civil
servants have had to cede their positions to people with limited or no capacity to
carry out the functions associated with their positions, not just to reward CNDD-
FDD members for their loyalty, but to also ensure that the administration is in the
“right” hands. The result has been a generalised weakening of state institutions
and widespread corruption.
Illegitimacy and power

Firm control of security institutions by a military-political group ensures that
potential threats to power are kept in check. This was illustrated by the brutality
of the army, police and intelligence services, when dealing with the foiled coup
of May 2015, and subsequent unrest. Thousands of people have been killed,
tortured, or abducted by the country’s security apparatus, while close to half a
million people have fled the country. Traditionally, illegitimate regimes in Burundi
have solely relied on their repressive security apparatus to remain in power.
To ensure the loyalty of security institutions, politicians have purged the
army, police and intelligence services of most of their old members, replacing
them with new ones who are more malleable and answerable to the ruling party.
Thanks to the Arusha Agreement, the police and intelligence services were
deeply reformed from the outset since the coming to power of the new regime in
2005. They have been dominated by elements from the Hutu rebellion ever since.
Reforming the army generally followed the provisions of the Arusha
Agreement, as it was already well-structured, unlike the police and intelligence
services. The new regime also understood the risk of brutally dismantling the
army to install its own members. But maintaining the status quo was not a viable
option.
One key provision of the Arusha Agreement was to ensure that there were
50% Hutus and 50% Tutsis in security institutions. Considering the traditional
dominance of these institutions by Tutsis, restructuring them implied injecting a
large number of Hutus, and sending a number of Tutsis into early retirement. This
was the first big pay-back opportunity for former Hutu rebels and other members
of the new ruling elite. Hence, a slow “restructuring” process was set in motion,
putting loyalists in key positions and sidelining the old guard, one by one.
As a result of the restructuring, a large number of former combatants, with
insufficient military and general training, suddenly became very senior members
of the army. Many former rebels in their early 30s became generals, an age at
which those who went for formal training through Burundi’s officers’ academy
would be lieutenants or captains (Nkurunziza, 2016a). As expected, this policy
caused frictions within the army, but discontent was contained for several years.
The destructive effect of this policy became apparent in the 2015 political
instability. The witch-hunt that followed the foiled coup targeted members of the
traditional army who were perceived to be sympathetic to the Tutsis, accused
of seeking to overthrow the regime. Many were killed and several high-ranking
officers fled the country. The army, namely its command structure, is now

dominated by former rebels who were integrated in it after 2005, defeating
the letter and spirit of the Arusha Agreement that provides for a balanced
representation (ICG, 2016).
Feeding state fragility

These transformations, combined with a selection system guided by an
obsession to ensure that the security apparatus remain loyal to the individuals -
not institutions - controlling political power, have weakened the security sector
in Burundi. The recent instances of human rights violations, observed particularly
since 2015, might partly be attributed to the lack of professionalism of security
actors. As observed above, the situation will only worsen if the isolation of the
political leadership persists. As has been the case under past regimes, resorting
to violence risks becoming the only instrument that political leaders can use to
stay in power.
Economic fragility
State rent extraction

Control of political power by a small group has a strong economic motivation:
the control of “rents to sovereignty” (Ngaruko and Nkurunziza, 2008), as noted
above. In Burundi, he who controls political power also controls economic
power. In a country where the private sector is rudimentary, most of the rents
are concentrated in the public sector where they are extracted through corrupt
practices, including the awarding of public contracts. “To become rich in
Burundi, you need a government contract” (ICG, 2012, footnote 59).
All regimes have considered the public sector as a source of rents to be
shared among family members and loyal followers. In this light, appointments to
some positions are considered strategic, not for the prestige they procure, but
for the opportunities for corruption they offer. For example, positions from where
one controls the allocation of government contracts, are held by individuals
who are trusted members of the small political elite running the country because
of the pecuniary opportunities associated with the function. Hence, some
have characterised Burundi’s conflict as largely distributional (Ngaruko and
Nkurunziza, 2000, 2008; Nkurunziza, 2012).
The management of state firms created in the late 1970s and 1980s illustrates
Burundi’s extractive system. About 100 state-owned firms were created between
1977 and 1982. They operated in sectors such as retail, telecommunications,
tourism, and manufacturing. These were among the country’s largest firms.
On average, state firms were more than four times the size of the average firm
created in the private sector during the same period, or more than seven times
the average size of firms in operation in 2010 (Nkurunziza, 2016b).
These firms were managed by civil servants - not business people - belonging
to the ruling elite. They had no business skills, as their primary objective was
to generate and distribute rents among the ruling elite and their associates. To

do so, they used selective employment targeting members of the ruling class
and their associates, and offered generous packages to their employees, even
though most of these firms never made a profit.
Corruption and mismanagement

To continue operating, every year, state firms absorbed a large share of the
government budget in the form of subsidies. For example, in 1995, state firms
recorded a net loss equivalent to 6% of GDP or 14% of government revenue,
excluding grants. These amounted to transfers from the population to the ruling
class and their allies. Mismanagement led most of these firms to failure in the
1990s and 2000s (Nkurunziza, 2016b). These large state firms benefited the
individuals who ran them and their associates at the expense of society at large.
Recently, similar practices have emerged in other sectors. There is anecdotal
evidence that state contracts are allocated without any competition to a handful
of individuals working with or on behalf of powerful politicians or generals.
Several mining contracts11 have been signed in total secrecy with the public
having not a shred of information on the terms of the contracts.12
Moreover, in 2011, OLUCOME, a local anti-corruption organisation,
purported to have evidence about 50 persons who it claimed had embezzled up
to 80 million US dollars. According to the non-governmental organisation (NGO),
this information was compiled from cases analysed by the NGO itself, as well as
state institutions such as the State’s General Audit Service, the Police, and the
Anticorruption Court.13 The NGO implied that the authorities have been aware of
these cases but unwilling to act, most probably because some influential figures
might be among those involved.
International institutions have also ranked Burundi as one of the most corrupt
countries in the world. According to the index of perception of corruption
compiled by Transparency International, Burundi is ranked 159th out of 176
countries surveyed in 2016.14 As a result, corruption and mismanagement
represent another important dimension of state fragility in Burundi.
11 Contrary to widespread opinion, Burundi is very rich in minerals, including some of the world’s
largest deposits of nickel, gold, coltan, and rare earth metals.
12 See for example, a detailed article by a local anti-corruption NGO, OLUCOME, on several
irregularities surrounding the granting of a mining license to the Musongati Mining operation (2014).
Accessible (French): http://olucome.bi/IMG/pdf/lettre_au_president_republique_pierre.pdf.
13 See: http://www.arib.info/index.php?option=com_content&task=view&id=3546
14 Transparency International (2016). Accessible: https://www.transparency.org/news/feature/
corruption_perceptions_index_2016

Economic
consequences of
state fragility
State fragility in Burundi
5 Slow growth
The permanent state of political instability has not allowed the country to focus
on economic development. Given the recurrence of the conflict (see Table 1 in
Section 2), any economic progress that is achieved between two episodes of
conflict is undone by the next conflict. As a result, political fragility has led to
economic fragility.
Capital accumulation and GDP growth

One indicator of economic fragility is the slow rate at which Burundi accumulates
physical capital. As illustrated in Figure 1, during periods of conflict, physical
capital is destroyed or not maintained, explaining the slow and mostly negative
rates of capital accumulation. In turn, a small stock of capital slows economic
growth, and growth rates remain slow as the conflict persists. It is against this
backdrop that Burundi’s GDP per capita has been declining since the early
1990s.
Consequently, Burundi has one of the highest incidences of poverty in Africa.
In 2013, 75% of Burundians were considered poor, as their daily income was
less than US$1.9, measured in 2011 Purchasing Power Parities (PPP). This rate of
poverty is almost twice the African average of 41% of the population in the same
year.15
11 World Bank (2017). PovCalNet. Accessible: http://iresearch.worldbank.org/PovcalNet/

povOnDemand.aspx

Figure 1: Annual rates of GDP growth (right axis) and capital accumulation
(left axis): Rwanda and Burundi (percentage changes)
Source: Data on GDP growth from UNCTADstat, UNCTAD’s database. Data on capital stock
computed by Nkurunziza (2017).
Figure 1 illustrates the economic cost of political fragility. In Burundi, GDP

contracted by 6.4% in 1972, the first time the conflict affected the whole country.
When another big conflict broke out in 1993, GDP growth contracted by 6% and
remained negative, or oscillated around zero, for most of the years until the mid-
2000s.
Owing to the very low GDP growth rates, the trend of Burundi’s GDP per
capita has never recovered from its collapse in the early 1990s. Similarly, its
stock of capital contracted by 3.4% in 1972. Moreover, during the conflict of the
1990s, the stock of capital contracted continuously for ten years, between 1994
and 2004. Thereafter, at least until the late 2000s, the rate of capital formation
remained weak, illustrating what has been identified as the absence of a peace
dividend in Burundi (Nkurunziza, 2016a).
Comparative analysis: Rwanda

To put the case of Burundi in perspective, it is compared with Rwanda, a
realistic counterfactual for Burundi. In 1994, GDP collapsed by half because of
the genocide. However, the recovery was swift. Between 1995 and 2014, annual
GDP growth averaged 9.75%. In Burundi, average GDP growth during the post
conflict period - between 2004 and 2014 - was only 4.1%, less than half of the
rate in Rwanda. This rate was also lower than the rates in other African post-
conflict settings such as in Sierra Leone (7.5%), and Mozambique (7.4%), for
example. Capital formation in Rwanda followed the same pattern as GDP growth.
Between 1995 and 2014, the average annual rate of capital formation was 4.9%,
twice the rate in Burundi’s post-conflict period, where it was only 2.3%.
Rwanda’s recovery and economic success since the mid-1990s may be

considered as an illustration of the “phoenix factor” hypothesis, whereby
devastated economies are transformed into more efficient and competitive
systems. This is thanks to a process of post-conflict rebuilding that relies
on better resources and the most up to date technologies. The fact that the
Rwandan state had collapsed as a result of the genocide implied that the country
had to rebuild its institutions from scratch. This might have been easier than
having to operate in a context where pre-existing institutions that had led to the
genocide would constrain the reform process. In contrast, Burundi’s failure to
emerge out of its fragilities seems to illustrate the “war ruin hypothesis”, which
posits that armed conflict destroys a country’s economy and institutions and that
post-conflict reconstruction is very costly and messy (Nkurunziza, 2017).
Slow economic recovery in Burundi means limited opportunities “to reinsert
its large number of ex-combatants and returnees from exile, and offer jobs to its
youth. Most particularly, ex-combatants needed economic opportunity in order
to help them pursue a peaceful path” (Nkurunziza, 2016a). The absence of a
strong peace dividend, combined with the breaking down of the rule of law, as
discussed earlier, make war a tempting option for the youth in Burundi.
Undeveloped private sector

Weak capital formation resulting from weak private investment have contributed
to keeping the private sector in a rudimentary state. As discussed earlier, private
sector development has never been the priority of political leaders, whose main
interest has been to extract rents from the public sector that is under their direct
control.
Private sector corruption

To some extent, rent extraction has extended to the private sector, by virtue
of the latter’s relationship with political spheres. Many actors in the private sector
are surrogates of politicians or people with strong political connections who use
them to obtain public contracts, the main source of economic opportunities in a
country where the state is the main economic agent. As a result, when regimes
change, many old private sector actors go out of business as they lose access to
public contracts and other associated rents, such as access to foreign currency
at overvalued official exchange rates. Members of a new political class therefore
install their own proxies in the private sector.
Firm growth
The politicisation of business extends political fragility to the private sector,
preventing it from growing and consolidating over time. Hence, the private sector
is dominated by very small firms with a median size of seven permanent workers
per firm. A census of formal firms carried out in 2010, with data covering the year
2007, found that micro firms (those with less than 5 full time workers) represented
34% of total firms, while small firms (those with 5 to 19 full time workers)
represented 46% of the total number of formal firms. Therefore, micro and small

firms represented 80% of all formal firms in Burundi in 2007 (Nkurunziza, 2016b).
The main issue is not that firms are small, but that they do not have the
opportunity to grow due to the challenging political and economic environment
they operate in. As the literature on industrial organisation in developing countries
has shown, small firm size has negative implications for firm resilience, access to
resources, productivity and welfare. Moreover, firms in Burundi are not only small,
but also limited in number. The same census found that in 2007, Burundi had only
1361 formal firms. This suggests a ratio of about 1.5 firms for every 10,000 people.
Assets and demand

In addition to the weakness of the private sector, political instability has had a
devastating effect on firms through the destruction of assets and plummeting
demand. There is no recent data in Burundi that has attempted to measure more
directly these two effects. The only attempt that seems to have been made, was a
survey of a sample of 84 firms in 1994, one year into the 1993-2003 episode of civil
war. It showed that only 45% of the firms had kept their pre-crisis size or reduced
by less than 10%. About 15% of firms shrank by 10 to 25%, and 21% of them
contracted by more than 25%.
It was also found that commercial banks’ short-term lending to firms in the form
of working capital increased from 62%, to 78% of total lending to the private sector,
just a few months after the beginning of the 1993 conflict. At this time, firms were
struggling to remain afloat in a very difficult business environment (Nkurunziza and
Ngaruko, 2002).
During the political crisis that started in 2015, several businesses curtailed their
activities for security concerns. Mainly in the second half of 2015 and early 2016,
security in Bujumbura was so precarious that firms’ opening hours were reduced or
simply closed. As tourists and business travelers stopped travelling to Burundi due
to insecurity, hotels and, to a great extent, restaurants, as well as related activities,
were among the most affected sectors. Moreover, many businesses have been
forced to shut down or curtail their activities following the killing or fleeing or their
owners. Foreign direct investment (FDI), which was small even before the crisis, has
all but collapsed, with the exception of some mining contracts that are shrouded in
secrecy, as referred to earlier.
Risk and unpredictability

Political instability has caused macroeconomic instability, which has in turn
compromised the health of firms and the economy.
As shown in Figure 2, inflation has been generally high, particularly during
episodes of conflict. During the longest period of instability, between 1993 and
2003, inflation was, on average, 15% per year, peaking at 31% in 1997, and dropping
to an average of 10% per year after the war (between 2004 and 2014). The average
lending interest rate was 16.4% between 1993 and 2003. High inflation and high
interest rates, in an environment of slow economic growth, dampened domestic and
foreign investment, and long-term economic growth.

Figure 2: Macroeconomic instability and economic performance
Source: World Development Indicators.
Financial constraints
Burundi has very limited fiscal space that shrinks even more during periods
of political instability. It relies on two major sources of financing: tax revenue
and official development assistance (ODA), including concessional loans.
Comparable data on tax to GDP ratios of African countries is sparse, and data
coverage is uneven across countries.
Tax revenue
According to available data, Burundi’s tax to GDP ratio is 14%, on average, well
below the African average of 18%.16 This particularly low level of tax revenue
reflects low compliance, poor fiscal governance, and low state capacity. With
respect to low compliance, tax evasion and widespread discretionary tax
exemptions in favour of politically-connected individuals have been identified as
representing an important loss of fiscal revenue.
There are three types of tax exemptions:
1. Granted by law or through adherence to an international agreement, such
as duty-free imports by international organisations or imports relating to aid
projects.
2. Exemptions in Burundi’s investment code, designed to attract investment.
3. Discretionary exemptions granted by authorities.
In 2012, for example, total exemptions represented about one-fifth of the total
revenue collected by the Office Burundais des Recettes (OBR) - Burundi Tax
Authority - during the same year (Holmes et al., 2013). Not all the exemptions are
16 This average is based on a limited number of observations so should be taken with caution.

illegal. Of the three types, discretionary exemptions are the most problematic as
they are prone to abuse. Even though they are granted by authorities vested with
this power, their legitimacy may be questionable when their objective is to benefit
specific individuals in their private capacity, with no positive effect on society.
Investment-related exemptions have also been abused by so-called investors
who end up selling at full price their duty-free imports that are supposed to be
part of their capital investment, most of the time with the complicity of some
leaders. These practices take place either with the complicity of the authorities
tasked with overseeing the use of exemptions or because they have limited
follow-up capacity or competence to establish that exemptions are used
as legally intended. As noted by a former head of OBR, the first to head the
institution, the capacity and competence of staff dealing with revenue collection
needs to be developed over time (Holmes et al., 2013). Directing external aid to
the development of domestic resource mobilisation could be an efficient way of
using foreign assistance, as the example of Rwanda has shown.
Untapped resources and poor state capacity

Capacity building and improvement in economic management, including
fiscal policy, as well as all areas of public sector management, will help the
country realise its economic potential. Burundi has been wrongly described
as a resource-poor country: it is instead a “resource-untapped” country with
considerable economic potential.17 However, its weak state institutions are not
conducive to development.
Indeed, Burundi’s overall Country Policy and Institutional Assessment (CPIA)
score has systematically been below the African average over the last ten years,
despite some improvements in specific sectors. In 2016, economic management,
and public-sector management and institutions, were the lowest performing
clusters relative to the Sub-Saharan average. These two clusters also regressed
in comparison with their values in 2008.18 Addressing low state capacity should
be a policy priority.
External aid
Burundi’s financial resource envelope has also been affected by movements in
aid flows, as summarised in Figure 3 below. Data covering the period between
1970 and 2014 shows that on average, Burundi recorded aid inflows amounting to
19% of its Gross National Income (GNI) per year, a value that is 73% higher than
the African average of 11%.
`17 The country has good arable land, high-quality coffee and tea, Lake Tanganyika, and minerals
such as nickel, cobalt, gold, and rare earth metals.
18 World Bank (2016). Burundi: Quick Facts, Country Policy and Institutional Assessment.
Accessible: http://documents.worldbank.org/curated/en/891501500349324004/122290272_2017062
00032524/additional/117514-REVISED-96p-WB-CPIA-Report-July2017-ENG-v16.pdf

Figure 3: Aid as a percentage of Gross National Income (GNI)
Source: Based on data from World Development indicators
As Figure 3 shows, aid flows have been highly unstable. One important feature of
aid disbursements to Burundi is that external aid has been used by donors as a tool to
influence the behaviour of political leaders at different stages of the country’s political
trajectory. Aid started to increase substantially in the second half of the 1980s, as
Burundi adopted structural adjustment measures, serving as an incentive to reform
the economy. From around 12% of GNI in the first half of the 1980s, the adoption of
structural adjustment in 1986 led to the doubling of aid, reaching 29% of GNI in 1992,
and 34% in 1994.
From 1996 to 2000, external aid dried up as a reaction to a military coup carried
out in July 1996, during the 1993-2003 civil war. A few days after the July 1996 coup,
Burundi was put under a total economic embargo and most of its donors suspended
their assistance to the country. This explains the drop in aid from 29% of GNI in 1995,
to 13% in 1996, and 6% of GNI in 1997. This move was intended to force the coup
plotters to cede power back to elected civilians. But it had a devastating impact on the
economy and social sectors.
The recovery of aid flows started in 2001, in response to the signing of the Arusha
Agreement by Burundian warring parties, and its entering into force in November 2001.
This heralded a new rapprochement between Burundi and the international community,
as the change in leadership was seen in Burundi and by the international community as
heralding a new political era of peace and prosperity in the country.
As a result, during its first term, the new regime that came to power in 2005 enjoyed
very high levels of external aid. Aid flows dropped after the 2010 elections that were
boycotted by the opposition, claiming that mass fraud had taken place. Worse still,
although no figure on aid flows after the 2015 contested elections is available, it is likely
that the ratio of aid to GNI has dropped back to its values during the embargo of the

1990s.
Indeed, all major traditional donors, including Belgium, the Netherlands, and
the European Union - the latter being main donor to the country (it had pledged to
disburse 432 million Euros between 2014 and 2020) - have suspended direct aid to the
government of Burundi as a way of forcing the leadership to adopt a more democratic
stance. The reduction of aid has severely affected the economy, as a large part of the
government’s budget, and most development projects, are funded through external
aid.
The drying up of external aid led to a massive devaluation of the Burundi franc, and
the reappearance of an active black market for foreign currency, at a time when the
official and parallel markets had but converged. One dollar was trading for about 1700
Burundi francs in May 2015, but it is now trading for about 2750 Burundi francs. This
represents a devaluation of about 62% in a period of two years.
Other consequences of the drastic reduction in external aid are a collapse in public
and private investment, and a drop in spending on social sectors, particularly health
and rural development. In late 2015, The World Bank noted that the 2015 crisis was
“turning a decade of good economic performance to a macroeconomic collapse.”

Conclusion
Lessons
6 Violence as a tool
This discussion on state fragility in Burundi has shown that fragility has been narrowly
associated with extreme political violence. The latter has been used by different
regimes to repress groups perceived to threaten their hold on power, precipitating the
whole country into civil war. Even though political violence in Burundi has its roots in
“divide and rule” policies introduced by the Belgian colonial power, the persistence and
recurrence of fights between political elites from the two major groups, the Hutus and
Tutsis, reflects the failure of Burundi’s post-colonial leaders to come to terms with these
divisive policies.
Unity, law, and legitimacy

It was the duty of politicians to re-instill into Burundians the sense of unity that had
characterised them centuries before the colonialists came to the country, late in the
19th century. On the contrary, all the regimes that have ruled the country since its
independence in 1962 have, to different extents, exploited the imagined differences
between the country’s two major groups as a strategy to keep a tight control on
political power and its spoils. The illegitimacy of these regimes has forced the ruling
elites to rely on violence as the only way of staying in power, further compounding the
fragility of the state.
The absence of the rule of law has been at the heart of the country’s persistent
insecurity. The use of state institutions as instruments of repression and the absence
of checks and balances in the way power has been exercised have alienated most
Burundians from their political leaders. Some victims of state violence have also
resorted to violence, either as a survival strategy – killing before you get killed – or as a
way of avenging past crimes. The most prominent members of the current leadership,
who came to power in 2005, owe their ascent to power to this strategy. One important
lesson from this experience is that the cycle of violence will likely continue, unless its
root causes, particularly the absence of the rule of law, are addressed.

Constraints on truth and reconciliation
Responsibility for past crimes must be established, and those responsible must
be prosecuted to allow the victims to reach some form of closure. This was at
the heart of the Arusha negotiations and the Arusha Agreement. However, this
aspect of the Agreement, one of the most important for peacebuilding, has not
been implemented.
The difficulty to serve justice arises from an incentive compatibility constraint:
past and current political leaders who are expected to make justice possible
might also be among the prime suspects. In this case, to make justice possible,
a third party is needed: the international community should play a prominent
role in this process. But judging by experience, there is little chance that the
international community will successfully help Burundi become peaceful, unless it
becomes more forceful in its engagement. It is not satisfactory that interventions
of the international community, whose mission is to end reprehensible actions by
the government, must rely on the goodwill of the same illegitimate regime.
Failure to intervene
Burundi can be considered as a test case as far as international intervention is
concerned. The failure to intervene in Burundi, a small African country, might
have been interpreted by illegitimate leaders elsewhere as permission to defy the
will of their people and get away with it. The failure to act has dented the moral
authority of the international community, which might restrict its capacity to act
elsewhere.
Prospects
Three steps are essential for Burundi to break out of the cycle of violence and
state fragility:
1. Affording justice
The country’s post-colonial history has shown that there will be no long-term
stability without justice. As provided for in the Arusha Agreement, the country
has no choice but to revisit its violent past and dispassionately rewrite the
country’s recent history more objectively. In doing so, the country’s leaders
should not shy away from establishing responsibility for past crimes and ensuring
that the victims feel that justice has been served. The argument that re-opening
past wounds would not be appropriate misses the point that such wounds have
never healed and will not heal unless “historical silences” are broken as they do
not serve the victims.
2. International intervention
As discussed above, judging by the past, it is not realistic to expect that
Burundian leaders will have the incentive or moral force to revisit the country’s

violent history on their own. The Arusha Agreement provided a good framework,
but different governments post-Arusha, shied away from implementing this
provision of the Agreement.
Therefore, only neutral international actors appear to be the right actors
who can help the country move forward. They will need to work hand in hand
with neighbouring countries, particularly from the East African Community,
who have strong influence on the leaders of Burundi. Regional and international
cooperation could, if necessary, force the leaders of Burundi to adopt and
implement a roadmap for peace and development. Such a roadmap could be
based on the Arusha Agreement or its variant.
3. Peacebuilding
It is also crucial to have an external guarantor for post-conflict institution-
building. Entrusting the implementation of the Arusha Agreement to the
government of Burundi, with its history of weak and poor governance, without
any monitoring mechanism proved to be a bad strategy. It is time to consider
entrusting this role to an external institution to oversee the implementation of an
agreement reached by the parties.
A clear and reasonable implementation timeline should be agreed on, with
the possibility of evaluating progress periodically. Building a democratic culture
and strengthening the rule of law would be at the core of this project. Critics who
would argue that this approach would put Burundi under the tutelage of a foreign
institution should admit that other “more reasonable” approaches have been
tried and failed. The Arusha Agreement is a case in point.
Going forward
These solutions will require sustained external financial assistance. Indeed, it
will be more worthwhile investing in Burundi’s long-term peacebuilding than
bearing the costs of its fragility. As discussed above, the cost of civil war
in Burundi includes half a million deaths; 1.5 million refugees; development
impediments such as malnutrition, poverty, and other ills associated with civil
war; as well as the destruction of human and physical capital. The conflict has
also had impactful negative spillovers, including the cost of caring for refugees in
neighbouring and faraway countries. It would be financially less costly to build a
conflict-resistant society than regularly pay the cost of a recurring conflict.
Special attention should be devoted to building a stronger and more resilient
economy. Most pressingly, it is important to develop the private sector to create
opportunities for sustainable and peaceful livelihoods for more Burundians. This
would aid in diversifying individuals’ pursuits beyond political leadership and the
civil service, diluting the benefits of controlling the state and rents to sovereignty.
In the reconstruction process, frontloading resources can build the foundations
of a stronger economy and generate a peace dividend, by incentivising those
who may be seeking revenge to pursue more peaceful ambitions.

References
Ajello, A. (2001). Cavalier de la paix: Quelle politique européenne commune pour
7
l’Afrique?, Editions Complexe. Collection Les livres du GRIP.
Bonneau, H. (1949). Comment le diable se mit le doigt dans l’oeil, Grands Lacs:
Revue Générale des missions d’Afrique, 64ème année, Nos 4-5-6, pages 9-10.
Boshoff, H. (2010). “Historical Overview”, in Boshoff, H., Vrey, W. and

Rautenbach, G. (eds). The Burundi Peace Process: from Civil War to Conditional
Peace. Monograph 171, Institute for Security Studies, 3-9.
Filmer, D. and Pritchett, L. (1999). “The Effect of Household Wealth on

Educational Attainment: Evidence from 35 Countries”, Population and
Development Review, 25 (1), 85-120.
Gahama, J. (2001). Le Burundi sous administration belge: la période du mandat

1919-1939. Second edition. Paris, Karthala. Karthala Editions, 1983 – Belgium.
Holmes, K., Ndihokubwayo, D., and Ruvakubusa, C. (2013). “For State and
Citizen: Reforming Revenue Administration in Burundi”, Africa Research Institute.
Policy Voices Series, Westminster, London.
Horowitz, D. (1985). Ethnic Groups in Conflict. Berkeley and Los Angeles:

University of California Press.
International Crisis Group (2016). Burundi: A Dangerous Third Term, Africa Report
N°235.
International Crisis Group (2012). Burundi: A Deepening Corruption Crisis, Africa

Report N°185. Bujumbura/Nairobi/Brussels.
International Monetary Fund (2017). World Economic Outlook, April 2017: Gaining
Momentum? Statistical Appendix. Washington, D.C.
Manirakiza, M. (2002). Burundi: quand le passé ne passe pas: (Buyoya I--

Ndadaye) 1987-1993. Editions Longue vue.
Manirakiza, M. and Harroy, J. P. (1990). La fin de la monarchie burundaise: (1962-

1996). Le Mât de Misaine.

Nkurunziza, J. (2017). Growth in Fragile States in Africa: Conflict and Post-conflict Capital
Accumulation, Framework Paper, Project on Growth in Fragile States in Africa, African
Economic Research Consortium, Nairobi.
Nkurunziza, J. (2016a). “Timing and Sequencing of Post-Conflict Reconstruction and

Peacebuilding in Burundi”, in Arnim Langer and Brown, Graham (eds), Building Sustainable
Peace: Timing and Sequencing of Post-Conflict Reconstruction and Peacebuilding, Chapter
13, Oxford, Oxford University Press, 223-242.
Nkurunziza, J. (2016b). “MSMEs Financing in Burundi and its Welfare Effect”, Belgian Policy
Research Group on Financing for Development (BeFinD), Working Paper No. 11.
Nkurunziza, J. (2012). “Inequality and Post-Conflict Fiscal Policies in Burundi”, in Langer, A.,
Stewart, F. and Venugopal, R. (eds), Horizontal Inequalities and Post-Conflict Development,
Palgrave Macmillan, 209-229.
Nkurunziza, J. and Ngaruko, F. (2008). “Why Has Burundi Grown so Slowly? The Political
Economy of Redistribution”, in Ndulu, B. J, O’Connell, S. A, Azam, J. P, Bates, R. H, Fosu, A.
K, Gunning, J. W, and Njinkeu, D. (eds). The Political Economy of Economic Growth in Africa,
1960-2000: Country Case Studies, Vol. 2, Chap. 2, Cambridge University Press, 51-85.
Nkurunziza, J. and Ngaruko, F. (2005). “War and its Duration in Burundi”, in Collier, P. and
Sambanis, N. (eds), Understanding Civil War: Evidence and Analysis, Volume 1: Africa, World
Bank Publications, Washington, D. C., 35-62.
Nkurunziza, J. and Ngaruko, F. (2002). Explaining Growth in Burundi: 1960-2000. CSAE

Working Paper Series, CSAE WPS/2002-03, University of Oxford.
Nkurunziza, J. and Ngaruko, F. (2000). “An Economic Interpretation of Conflict in Burundi”,

Journal of African Economies, 9 (3), 370-409.
Reyntjens, F. (1994). L’Afrique des Grands Lacs en Crise: Rwanda et Burundi, 1988-1994.
Paris: Karthala.
Sandrart, G. (1953). Le Ruanda-Urundi: Terre des rois-pasteurs, Dessart, C (ed.) Ruanda-

Urundi, Brussels: Plus Oultre.
Stapleton, T. (2017). A History of Genocide in Africa, Praeger Security International. Santa

Barbara, California, Praeger.
Toynbee, A (1969). Experiences. Oxford, Oxford University Press.
United States Institute of Peace (2004). International Commission of Inquiry for Burundi:
Final Report. Accessible: https://www.usip.org/sites/default/files/file/resources/collections/
commissions/Burundi-Report.pdf.
Whitaker, B. (1985). United Nations Whitaker Report on Genocide, United Nations

Economic and Social Council Commission on Human Rights. Thirty-eighth session, E/CN.4/
Sub.2/1985/6 — 2 July.

The LSE-Oxford Commission on State
Fragility, Growth and Development
was launched in March 2017 to guide
policy to address state fragility.
The Commission, established under

the auspices of the International
Growth Centre (IGC), is sponsored
by the London School of Economics
and Political Science (LSE) and the
University of Oxford’s Blavatnik School
of Government. It is funded from the
LSE Knowledge Exchange and Impact
(KEI) Fund and the British Academy’s
Sustainable Development Programme
through the Global Challenges
Research Fund.
www.theigc.org/fragilitycommission

The Origin and Persistence of State Fragility In: Burundi

Uploaded by

Copyright:

Available Formats

The Origin and Persistence of State Fragility In: Burundi

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Origin and Persistence of State Fragility In: Burundi

Uploaded by

Copyright:

Available Formats

APRIL 2018

The origin and

About the commission

The commission, established under the auspices of the International Growth

Front page photo: SuSanA | CC BY-SA (with modification)

2 The origin and persistence of state fragility in Burundi

Origins of state fragility in Burundi 7

Regional and international dimensions of 13

Strategies and policy choices that feed state 19

Economic consequences of state fragility 23

3 The origin and persistence of state fragility in Burundi

State fragility in Burundi

3. As political elites have primarily focused on controlling the state, and

5. The combination of these fragilities has made Burundi highly vulnerable to

4 The origin and persistence of state fragility in Burundi

Instability and violence

5 The origin and persistence of state fragility in Burundi

The paper proceeds as follows:

2. In Section 3, the paper analyses the regional and international dimensions of

4. Section 5 highlights some major consequences of state fragility.

6 The origin and persistence of state fragility in Burundi

Credit: Bruno De Hogues/Getty

Historical origins of state fragility

Divide and conquer

7 The origin and persistence of state fragility in Burundi

8 The origin and persistence of state fragility in Burundi

The Arusha Agreement

9 The origin and persistence of state fragility in Burundi

Burundi’s internal armed conflict

10 The origin and persistence of state fragility in Burundi

Characteristic 1965 1972 1988 1991 1993–2003 2015 to date

Duration (months) 2 4 2 1 120 27

Deaths (thousands) 5 200 15 1–3 300 1.2

Refugees (thousands) 0 300 50 38 687 416

Deaths & refugees 0.2 14.0 1.3 0.7 17.1 3.8

Provinces affected Muramvya Whole Ngozi, Cibitoke, Whole Whole country

In retaliation, Bururi-dominated government forces engaged in ferocious

11 The origin and persistence of state fragility in Burundi

The pursuit of peace

12 The origin and persistence of state fragility in Burundi

State fragility in Burundi has had important regional and international

13 The origin and persistence of state fragility in Burundi

14 The origin and persistence of state fragility in Burundi

African-led peace negotiations

The Arusha negotiations

15 The origin and persistence of state fragility in Burundi

National Truth and Reconciliation Commission

16 The origin and persistence of state fragility in Burundi

The United Nations

17 The origin and persistence of state fragility in Burundi

18 The origin and persistence of state fragility in Burundi

Credit: SuSanA | CC BY-SA

Civil service and security fragility

19 The origin and persistence of state fragility in Burundi

Illegitimacy and power

20 The origin and persistence of state fragility in Burundi

Feeding state fragility

State rent extraction

21 The origin and persistence of state fragility in Burundi