Exploring the Relationship Between Environmental Scarcity and Violent Conflict: A Case Study of Burundi
Shannon England
I. Introduction

Since the overthrow of President Melchoir Ndedaye in October of 1993, Burundi has been locked in a gruesome civil war. Although estimates vary considerably, it is acknowledged that, at a minimum at least 160,000 Burundians, mainly civilians, have been killed in the massacres. A brief literature search using 'Burundi' and 'Violence' as key words turned up over a thousand news articles over the past two years. A glance at some of the more recent headlines about Burundi from the past year illuminates the magnitude of violence continuing to affect Burundians today:

The natural question for anyone reading these articles or for anyone affected by the tragedy of war is always 'Why?' Attempting to understand the causes of any particular violent conflict is not an easy undertaking and there is never a simple explanation for violence. Attempts to clarify will almost inevitably lead to oversimplifications, nonetheless, the effort to understand - and hopefully to prevent - violent conflict is important. An interesting question about violence was once posed to public health practitioners in the United States by the Surgeon General, and it changed the way we think about violence in our own society. The reality is that the question applies equally well to any society plagued by violence: "If violence is not a public health issue, then why is it killing so many people? It is a good question and the issue deserves serious attention by scholars.

Recent scholarship on international conflict has focused on the interaction between environmental scarcity and violence. Thomas Homer-Dixon is a professor of Political Science at the University of Toronto and co-director of the Project on Environmental Change and Acute Conflict. To date the project has researched several different countries and regions as case studies for the many hypotheses generated about the relationship between environmental scarcity and violent conflict. These have included studies of Bangladesh, Egypt, Senegal, El Salvador, Honduras, China, and Haiti. More recent work has focused on the Gaza strip, Mexico, Pakistan, South Africa, and Rwanda.

I propose to examine Burundi as a case study for applying these models of environmental scarcity. I will also incorporate ideas of transition theory as explained by William Drake, Professor of Public Health at the University of Michigan, to develop an understanding of the complex relationships evident in Burundi's current and past crises.

II. Overview of Burundi


Burundi is a small nation in Eastern Africa surrounded by Rwanda, the Democratic Republic of Congo, and Tanzania. (See Map 1.) The total area of Burundi is 27,830 square kilometers with 25,650 square kilometers of land, approximately equal in size to the state of Maryland. The country's climate is temperate with a hilly and mountainous terrain. There is a plateau in the east and plains along the coast of Lake Tanganikya. Arable farmland accounts for only 43% of the total land. Approximately 35% of the land is used for pasture and raising cattle. Soil erosion is growing problem in Burundi due to both overgrazing of cattle and expansion of agriculture into marginal areas. Soil exhaustion is also becoming a concern in some areas. Deforestation of Burundi's land is almost complete with less than 2% of the country currently covered by forests, primarily in designated national parks. Trees are cut down mainly for firewood fuel.


The most recent estimates for Burundi's population place it at 6,088,000 as of mid-1996. This represents an almost 250% increase in population since 1950. The population growth rate for 1996 is estimated at 2.9%. (See Chart 1.)

Burundi's extreme population density of 237 people per square kilometer places it fifth in Africa just behind Rwanda and the island nations of Comoros, Reunion, and Mauruitius in terms of population density. If only arable land is taken into account, Burundi's population density approaches 551 people per square kilometer. Comparative statistics with some other countries noted for high population densities highlight the magnitude of current population pressures in Burundi. (See Charts 2, 3, Map 2) (Click here for a larger version of Map 2.) Population pressures are also not evenly divided in Burundi, some areas are much more densely crowded than others. (See Map 3.)


Burundi's population is very similar ethnically and culturally to its neighbor Rwanda to the north. There are three main ethnic groups in Burundi: the Hutu, who comprise approximately 85% of the population, the Tutsi, about 14% of the population, and the remaining 1% consisting of Pygmy Twa, South Asians and Europeans.

Burundi's ethnic divisions, like Rwanda's, are often depicted as representing centuries old animosities, originating with distinct tribal divisions that inevitably must lead to war. In reality, however, the divisions between the Hutu and Tutsi are not nearly so simple to explain and any reference to "tribalism" as a explanation for conflict in Burundi is essentially meaningless. Burundi is an anomaly in a continent of states with artificially constructed borders in that it was a 'kingdom' long before colonization and has been a "national entity" for centuries. Moreover, both Tutsi and Hutu speak the same language in both Rwanda and Burundi- in Burundi the language is known as Kirundi. Hutu and Tutsi also share the same social organizations and have never been separated geographically. In fact, Hutu and Tutsi have lived literally side-by-side, and mostly peacefully, for centuries. The issue is further complicated in Burundi due to the fact that Burundi's kingdom was ruled, not by the Tutsi as in Rwanda, but by a ruling class of royalty known as the Ganwa, considered neither Tutsi nor Hutu, but rather seen as having origins and legitimacy with both groups.

The distinction between Tutsi and Hutu, historically, was based much more on socio-economic divisions than upon ethnic lines. Considerable intermarriage between "Hutu" and "Tutsi" was common, as was movement between groups. The term "Hutu" originally had more to do with social status and hierarchical relationships than with ethnic identity. In Kirundi the word was traditionally used to connote one who was in a subordinate position to another.

Because of the importance of ethnic distinctions in Burundian society today, much discussion by both Tutsi and Hutu centers around perceived physical differences between the groups. Tutsi are often described in the media as 'tall, thin, beautiful people, descendants of a Northern Hamitic population of pasturalists' and the Hutu as 'darker, shorter, Bantu stock, traditionally agriculturalists.' The fact of the matter is that it is not always possible for Burundians to distinguish a 'Hutu' from a 'Tutsi' on physical characteristics alone. Many Burundians do not fit the stereotypes of their ethnicity and many are killed for it when violence erupts. The historical distinction of the Hutu as agriculturalist and the Tutsi as pastoralist probably has a bit more truth to it, but the reciprocal nature of cattle tending in traditional Burundian society was more of a uniting factor for Hutu and Tutsi than a divisive one.

A full discussion of the pre-colonial relationships between Hutu and Tutsi is beyond the scope of this paper, but the important point is that ethnic conflict in the region is not the historical reality that many assume. For more information about pre-colonial society in Burundi, interested readers should refer to Rene Lemarchand's book Burundi: Ethnocide as Discourse and Practice, Woodrow Wilson Press, (1994).


Burundi is a resource poor country, described by some as being in an early stage of economic development. The economy is dominated by the agricultural sector and 90% of Burundians depend upon subsistence level farming. The main crops for farmers are coffee, tea, sorghum, sweet potatoes, bananas, manioc (cassava), meat (cattle and goats), and milk.

Burundi's per capita income is one of the lowest in the world at $210. The UNDP 1994 Human Development Report lists Burundi as second only to Rwanda for having the highest percentage of people living in poverty: eighty four percent of Burundians are estimated to live in absolute poverty. (See Chart 4.)

Foreign exchange for the government is dependent upon the nation's coffee crop which accounts for 80% of the country's foreign exchange earnings. This dependence on one crop leaves Burundi's economy extremely vulnerable to both climate change and to market changes in world coffee prices. (See Map 4 for a depiction of Burundi's main agricultural zones.)

Health Indicators

The overall poverty rate in Burundi is reflected in several health status measures: the most notable of them being average life expectancy. A Burundian's average life expectancy is only 50.2 years, compared with 76.2 for North Americans and 64.7 for the world wide average. (See Chart 5.) Many Burundian children are subject to malnutrition resulting in growth problems. According to the UNICEF report on the State of the World's Children, almost half of the children in Burundi are suffering from malnutrition induced wasting and stunting.
(See Chart 6.) Burundi is, in fact, among the worst fed countries in Africa: only about 84% of the country's daily per capita food supplies are currently being met. (See Chart 7.) Maternal mortality in Burundi is also very high, estimated at about 1,300 deaths per 100,000 live births. Contraceptive prevalence is extremely low with less than 1% of married women reporting the use of any contraceptives at all.

III. Theories of Environmental Scarcity and Violent Conflict

The central questions posed by scholars interested in the link between the environment and conflict began first and foremost with speculations about population growth. Given past growth rates, within the next fifty years or so, the total population of the earth will probably surpass nine billion people. Unless radical changes are implemented, global economic output is expected to grow exponentially alongside of total population. If we accept the notion that the earth has finite resources, the next question becomes this: what happens to human populations when resources become scarce?

In his article "On the Threshold: Environmental Changes as Causes of Acute Conflict" Thomas Homer-Dixon poses the question of what happens when renewable resources of agricultural land, forests, water resources, and fisheries become scarce: In his words: "If such 'environmental scarcities' become severe, could they precipitate violent civil or international conflict?" Based on research of specific cases, he presents his own answer to the question in his follow-up article, "Environmental Scarcities and Violent Conflict: Evidence from Cases."

The concept of environmental scarcity defined by Homer-Dixon refers to three main sources of scarcity of renewable resources: environmental change, population growth, and unequal social distribution of resources. Environmental change occurs when human induced activity produces a decline in the quantity or quality of natural resources at a rate faster than it can be renewed by natural processes. Population growth creates a situation where per capita availability declines as resources are split amongst a growing number of people. Unequal resource distribution causes environmental scarcity by concentrating resources in the hands of a few at the expense of the many.

These three causes of environmental scarcity can and do interact to produce even more extreme results. Two such patterns of interaction, "Resource Capture" and "Ecological Marginalization," are described in Figure 1. "Resource Capture" is described as a situation where environmental changes resulting in a reduction in available renewable resources combine with population growth to encourage elites within a society to shift resource distribution in their favor. Severe environmental scarcity is the result for the majority. "Ecological Marginalization" describes a situation where unequal access to resources combines with increased population growth to force migrations to regions that are ecologically unstable. High population densities and a lack of capital or knowledge necessary to protect natural resources, in turn, results in increasingly more severe environmental changes leading to greater environmental scarcity for all.

How does environmental scarcity cause violent conflict? A model of the sources of and possible consequences of environmental scarcity is presented in Figure 2. First, environmental scarcity may cause population movement which can lead to "group-identify conflicts" conceivably leading to violence. Migrations of newcomers may also place new demands upon states and decrease their legitimacy. Environmental scarcity can also lead to violence by contributing to increased economic deprivation, resulting in a decreased legitimacy of the state.

Homer-Dixon is careful to note that economic deprivation in and of itself does not lead directly to violence. Notions of 'deprivation' are subjective and poor people do not necessarily revolt. First, they must believe that they have a right to economic well being. Societies with fatalistic attitudes about deprivation are unlikely to be as prone to violence as are those where the poor believe they have a right to economic and social justice. Second, the economic deprivation must be severe and persistent enough to erode the moral authority of the state. The probability of civil violence is also increased when groups centered around social cleavages (class, ethnicity, religion) are already in existence and can mobilize aggrieved individuals to action.

The theory further speculates that conflict induced in this manner is likely to be sub-national, diffuse and persistent. Environmental scarcity has cumulative impacts as it can lead a society into a continuous spiral of decline by encouraging widespread population movements and increased economic despair. Violence is also likely to increase in intensity as environmental scarcities increase. For states affected by environmental scarcity, Homer-Dixon predicts one of two outcomes: fragmentation and a decreased ability to control their regional territories or, alternatively, the state may become authoritarian and militarized in an attempt to control internal challenges to authority.

It should be noted that Homer-Dixon and his colleagues are not without their critics. The critics argue that simply because environmental scarcity and violent conflict may happen to coincide in a particular region this does not imply that you can prove causality. Many possible causes of conflict can usually be identified and focusing on one particular cause (environmental scarcity) is unlikely to explain very much. Moreover, even if causality could be proved, by only studying cases where conflict and environmental scarcity are both in evidence, scholars in the field have left themselves open to the charge that the research does little to add to the overall body of knowledge. After all, ideally research would be directed at finding solutions to the problem of violence and conflict engendered by environmental scarcity- and not simply at proving that such a connection exists. By not examining a "control group" of countries or regions that have avoided conflict despite severe environmental scarcity, potentially useful "model cases" are overlooked.

Thomas Homer-Dixon addresses some of these issues in a rebuttal specifically addressing both methodological techniques and the problems inherent in studying environment and population dynamics. He points out that all too often those analyzing violent conflict have focused on politics and have simply failed to consider the independent limits inherent in some environmental systems, as well as the resulting effect of those limits upon rational actors responses. In addition, "threshold" effects evident in many systems may lead to a situation whereby environments pass a point of irreversibility. Environmental scarcity in these systems will have long reaching effects upon human systems regardless of economic or political responses to the problem. He argues that ignoring such an important variable when analyzing violent conflict is not the answer.

Homer-Dixon also argues that identifying "control countries" to match up to specific "experimental countries" would be virtually impossible given the highly complex nature of environmental and political interactions. His argument is that case studies focusing on areas where clear relationships between violence and environmental scarcity exist may help scholars identify potential relationships that can then be used to study conflict in other areas, and that case studies are a necessary first step to understanding those relationships.

IV. Is Environmental Scarcity Related to Violence in Burundi?


A brief introduction to Burundi's current political crisis will perhaps help to clarify a very complex situation of ethnic tension. An understanding of the historical situation is also crucial to interpreting any of the relevant data on environmental scarcity to be presented. The question remains, however, "Is environmental scarcity related to ethnic violence in Burundi? And if so, how?" In attempting to answer these questions in light of Homer Dixon's theories, it is perhaps best to focus on the availability of different renewable resources in Burundi over time. Water resources are not scarce in Burundi. For the most part, people have access to enough water to meet all of their personal and agricultural needs. Renewable resource sectors to examine then become the following: agriculture, fisheries, and forests. Because population growth is also necessarily related to environmental scarcity, it will additionally be useful to look at trends in Burundi's population growth over time. Transition theory, as described by William Blake in his article, "Towards Building a Theory of Population-Environmental Dynamics: A Theory of Transitions," will be used to speculate on Burundi's stage in these four different sectors over time. Finally, the history of recent violence in Burundi will be analyzed in light of this data and its relationship to current theory about environmental scarcity and conflict.

Historical Overview of Recent Violence in Burundi

Patterns of Recent Violence
Since independence from the Belgian administration in 1961, Burundi has experienced several major periods of ethnic turmoil and violence. As mentioned previously in the introduction, Burundi's most recent period of conflict began in October of 1993 with the assassination of the first democratically elected Hutu President in Burundi's history. During this most recent period of conflict, there have been upwards of 160,000 deaths with perhaps as many as 1,600,000 people dislocated from their homes.

In 1991, a period of violence claimed between 4,000-6000 lives. There were over 5,000 Burundians killed by ethnic violence in 1988, and in 1972 over 100,000 Burundians were killed by their fellow countrymen in one of the most gruesome massacres in African history. In 1965 a wave of killing took the lives of approximately 5,000 Burundians. Charts 8 and 9 depict very conservative estimates of the number of those killed and the number of those forced from the country due to violence since 1961.

The conflict in Burundi today is explained as a simple conflict for power between Hutu and Tutsi, but it cannot be understood completely without also considering intra-group rivalry between the ruling Tutsi elite. Each of the aforementioned outbreaks of violence involved to some degree a grappling for power of the state amongst different factions of the ruling Tutsi classes. Each wave of violence, too, is necessarily connected intimately with the ones preceding it. Indeed, violence and the memory of violence permeates the political consciousness and the meaning of ethnicity for all Burundians. One must also consider the history of Rwanda to understand Burundi. Due to cultural and historical similarities, the political situation and crises in Rwanda have always had a profound effect on subsequent events in neighboring Burundi.

The political and historical events leading up to each of Burundi's massacres have been duly noted elsewhere and the story is not a pleasant one. A brief summary in this paper cannot do justice to the subject, but the story perhaps begins with the assassination of Burundi's first ever Prime Minister designate the Prince Louis Rwagasore on October 13th, 1961.

Post-Independence Burundi
Prince Rwagasore, a member of the UPRONA party was assassinated in a plot by the rival PDC party (perhaps in conjunction with the departing Belgian regime) only one month after winning the election on the eve of Independence in 1961. The tragic death of Prince Rwagasore sent Burundi down a path of violence that some beleive could have been avoided. Rwagasore's popularity, charm, and perhaps more importantly, his legitimate claim of decent from the Ganwa royalty may have served as a unifying factor in a country that was destined to become very ethnically divided.

Rwagasore's death marked the beginning of a crystallization of ethnic identity among Burundian elites. The Rwandan revolution and the establishment of a ruling Hutu government in Kigali also served, for both Hutu and Tutsi elites, as a demonstration of what could potentially happen in Burundi. Majority rule with a distinctly Hutu government in place was a prospect looked upon favorably by Burundian Hutus and decidedly not so by Tutsis. An influx of Rwandan Tutsi refugees escaping from violent conflict poured into the Northern provinces of Burundi. These refugees, as many as 50,000 in number, directly spread ideas of ethnic hatred in Burundi that could not help to fan the flames of a rising sense of an identification based upon ethnicity in Burundian elites.

In 1963 and 1964, soon after independence, the Tutsi and Hutu elite were jockeying with the monarchy (made up of the 'Ganwa') for control of both the government and the UPRONA party. (Burundi was created as a constitutional monarchy at the time of independence from Belgium in 1961. UPRONA was the dominant political party.) The intrusion of external cold war factors tipped the balance of power to the Tutsi with the "Congo Rebellion" in Eastern Zaire. The Chinese wished to use Burundi as a base to transfer ammunitions into Zaire. In exchange, they agreed to provide financial support to the Tutsi elite in Bujumbura. Tutsi refugees from Rwanda also entered into the equation because they wanted to use Burundi as a "safe-haven" from which to wage a war against the Hutu government in Rwanda. Rwandan Tutsi thus had a vested interest in helping the Tutsi in Burundi to gain power over both the Hutu and the Royal Crown and they soon became pivotal in the events that would unfold.

On January 18th, 1965 a Rwanadan Tutsi shot Burundi's Hutu Prime Minister Pierre Ngendandumwe. The event sparked outrage amongst the Hutu elite. The monarchy, sensing its legitimacy quickly slipping away, called for general elections to be held for both a new Prime Minister and National Assembly. The Hutu emerged from a legislative victory in May of 1965, capturing 23 out of a total of 33 seats in the National Assembly. Hutu elites naturally expected that a new Hutu Prime Minister would be placed into power. Instead, the monarchy selected a famous "Ganwa," Leopold Biha, to the position. Meanwhile the government, increasingly controlled by the monarchy, decreed that the number of communes (subsets of provinces) would decrease from 181 to 78 and that the elected leaders of the communes (most of whom were Hutu) would be replaced by functionaries appointed by and responsible to the court. This blatant attempt to concentrate and increase power into the hands of the ruling monarchy was not well received by either the Hutu or Tutsi elites.

Hutu elites responded in anger with an attempted coup against the ruling monarchy in October of 1965, a response that was to prove disastrous for them later. The coup attempt failed and troops loyal to the Crown regained control as the King and his courtiers fled to Zaire in panic. Hutu elites were then accused by the Tutsi of being disloyal to the government and attempting to establish a Hutu revolution similar to what had unfolded in Rwanda. Thousands of Hutus were killed in the purges of the army and the government that followed. All Hutus with any political standing were killed, and with the monarchy in exile, power came to rest exclusively with the Tusti elite.

The repression was brutal, resulting in at least 5,000 Hutu civilians being killed, mostly in the provinces of Bujumbura and in neighboring Muramvya. Surprising though it seems now, most of the rest of the country remained relatively calm, and relations between Hutu and Tutsi were fairly peaceful. The fighting among the "elite" for power seemed to have little impact on the day to day lives of Burundi's rural population. In the urban areas, by contrast, polarization around ethnic identity was fairly complete.

Besides the complete elimination of the first generation of Hutu leaders, one other notable effect resulting from the failed coup was the extreme weakening of the monarchy. The Court and most "Ganwa" were still in exile and it was not until March of 1966 that the King decided to entrust power to his son the Prince Charles Ndizeye. The reign of the new king, crowned Ntare, was to be a short lived five months. The Tutsi led government and military moved against the new king while he was away in Zaire celebrating the first anniversary of President Mobutu Sese Seko's rise to power.

The abolition of the monarchy following the Tutsi led coup in 1966 probably sent the final blow to national unity in Burundi. In Burundi, in contrast to Rwanda, the ruling monarchy class or "Ganwa" had always served as a unifying factor for Hutus and Tutis. Both groups, although occasionally in direct conflict with the monarchy, generally acknowledged the perceived legitimacy of Ganwa rule. The Crown had thus served as an intervening factor in diminishing Hutu and Tutsi perceived ethnic differences.

The First Republic
The new Tutsi led government, known as the First Republic, was led by President Michel Micombero. President Micombero was a Tutsi of the Hima clan from the south of the country, Bururi. He filled his government with his fellow clansmen. Of the seventeen military officers in his newly formed National Revolutionary Council (NRC), 8 were from Bururi and only 3 were of Hutu origin. The National Assembly was abolished and a military rule established. The UPRONA party was called upon to be the "unifying force" for an increasingly divided Burundian society.

The new government, despite its rhetoric, was hardly democratic and far from a "unifying" force. The struggles that were to follow stemmed from a conflict between Tutsi and Hutu, but also between Tutsi and Tutsi. Different Tutsi clans, based both on regional and historical kinship lines, struggled for control of the state. The threat of Hutu insurrection, with the example of Rwanda ever present in their minds, was often used as a tool by the ruling government to bring any potential Tutsi competitors for power back into line. Specifically, reference to the 'Hutu peril' as used by those from the South to accuse competitors from the North as being too 'soft' on the Tutsi-Hutu 'problem.' As Muramvanya was the site of the deposed King, Tutsi from this region were also accused of seeking to re-establish the monarchy, which supposedly could not happen without Hutu support. Purges of the army and government of 'Hutu elements' continued. A particularly harsh repression began after an accused Hutu rebellion plot was "uncovered" in 1969. Whether or not there was any basis for the charge is unclear, as there was no documentation, but the state sanctioned killings that resulted certainly were clear. What few Hutu remained in the government were clearly "tokens" and the remaining Hutu elite began to feel increasingly insecure.

In the same way that the Hutu "problem" was related to intra-group conflict among the Tutsi, similarly, Hutu insurrection in Burundi was always preceded by perceived weaknesses in the ruling government. Hutus sought to exploit the opportunities they saw in the deepening rift between Tutsi elites from the North (Muramvanya) and those from the South (Bururi). Violence erupted full force in Burundi in 1972 after just such an attempt by Hutus to wrest power from the increasingly brutal Tutsi regime. The Hutu insurrection occurred in April of 1972 after the well publicized and much discussed trials of several Muramvanya Tutsi and Ganwa by their Southern Tutsi rivals from Bururi. The Hutu uprising began in the south of the country in Nyanza-Lac when Burundian Hutus invaded from across the border of Tanzania. Assisted by Zairean troops, Hutu insurgents attacked the military outposts in Nyanza-Lac and Rumonge. Simultaneous attacks also occurred in Cankuzo and Bujumbura. In Rumonge and Nyanza-Lac, Hutu began to slaughter every Tutsi in sight. Most of the victims were civilians and estimates are that as many as 3,000 Tutsi were killed. Also killed were any Hutu that attempted to resist the slaughter.

The official interpretation of the insurrection varied widely. Some claimed that it was a plot by the former ex-king Ntare, who had returned to Burundi from Zaire, and, it was claimed, had plotted with foreign mercenaries to reinstate the monarchy. Others claimed that official high-ranking Hutus had collaborated to finance a rebellion. The historical facts hardly matter, as the repression that followed the insurrection made enemies both of the suspected 'monarchists' and the Hutu officials. They were, in fact, accused of plotting together. In the slaughter that followed, Hutus throughout the country were executed in large numbers. The ex-king Ntare also was summilarily executed. The bloodbath was carried out over several months in a chilling and systematic manner. In an effort to wipe out the Hutu elite entirely, every Hutu male with any education at all was accused of conspiracy against the government. In a call to "rise up against the python in the grass" Tutsi were advised to slaughter all Hutu males down to even the grade school level.

In many areas of the country, Hutus fought the rebels alongside of the Tutsi, only to be later slaughtered in turn. If anything served to cement the hatred of Hutu for Tutsi, it was this systematic execution that took place throughout the country. Never again would the political struggles of the elites be seen by the rural population as irrelevant to their lives. The scale of the killings varies. Some conservative estimates say 100,000 people, others claim the number is closer to 200,000. In addition, at least 150,000 Hutu fled Burundi to neighboring Tanzania and Zaire to live in exile.

The systematic nature of the killings can probably only be understood by noting the cycle of fear by now implicit in the relationship between Hutu and Tutsi. Many Tutsi saw the elimination of the Hutu as the 'final solution' necessary to protect their well-being. As had happened in the South, they feared a slaughter of every man, woman, and child of Tutsi origins. This fear was compounded by the anti-Tutsi violence that shaped Rwanda's history from 1959-1962. The Tutsi elite hoped to not only eliminate the perceived immediate threat to their well being, but also to inspire fear and terror that would be remembered for generations to come. It should also be noted that, although most of the victims were Hutus, many Tutsi were killed as well. Particularly targeted were any of the remaining Muramvanya Tutsi accused of monarchism. Many Tutsi also died attempting to protect their Hutu friends and neighbors.

While not all Tutsi were involved in the genocide, many clearly benefited from the deaths. Transfer of property to Tusti seems to have been a major inducement to violence in some areas. Those killed were robbed of their money, personal belongings, lands, and livestock.

After this ethnic cleansing, the only 'elites' in Burundian society were of Tutsi origin. A strict rule of 'Kirundization' of the schools was instituted whereby it was against the law to teach French in all government schools. This effectively served to prevent Hutu children from learning the language of governance, restricting it to the realm of the wealthier and ruling Tutsi. Because admission to higher levels of education was predicated upon a knowledge of French, Hutus were effectively blocked from rising to positions of prominence in society. The army became even more Tutsi dominated, to the extent that nearly 100% of the recruits and officers were Tutsi.

The Second Republic
On November 1, 1976, Burundi's state structure underwent another profound change with the coup d'etat placing into power Lt. Col. Jean-Baptiste Bagaza. The Second Republic was a time of consolidation of Tutsi rule and an attempt to codify into the national law and discourse an official view of 'National Unity,' meant to restore the moral legitimacy of the state. The official view of history propagated by the government at this time was mostly an attempt to legitimize Tutsi rule through the ruling UPRONA party. Hutu were almost entirely excluded from political participation. By late 1987 only 2 seats out of 65 in the Central Committee of Uprona's party were held by Hutus. To prevent claims of ethnic favoritism or discrimination in educational or public life, the Second Republic simply banned the public discussion of ethnicity. Anyone publicly referring to Hutu or Tutsi ethnic identities was charged with inciting 'racial hatreds' and arrested. Individuals were issued identity cards that had to be carried with them at all times. Rural Hutu were, for a time, organized into "villages" in a nationwide campaign of "villagization." Ostensibly to increase agricultural yields in a "development scheme," the move was, in reality, an attempt to exert greater control over the Hutu communes. The project ultimately failed, but it is an example of the lengths to which the Tutsi dominated regime went to officially denying the real motives for its actions.

Divisions and cleavages within the ruling Tutsi hegemony would once again lead to a renewed cycle of violence in Burundi. This time the church played somewhat of a role in the ensuing violence that was to come. The Catholic Church, as well as several Protestant denominations, had been establishing schools and organizing community-based projects for rural Hutu. These activities threatened to undermine the carefully constructed power structures that the Tutsi rulers had created. In addition, the UPRONA party leaders remembered the crucial role of the Catholic Church in the "emancipation" of the Hutu masses in Rwanda in the 1950s. Several Church affiliated leaders also began to speak up for Hutu labor groups and (illegal) political parties.

The state responded by severely restricting the activities of churches in the country. New rules were made restricting the appropriate days for worship and meetings. Religious broadcasts were banned and foreign missionaries were expelled. Religious primary schools were closed in 1977 and all religious secondary schools were closed in 1986. Informal 'catechism schools' organized for Hutu school children by the Catholic Church were also forbidden. Some 220,000 children were denied basic schooling in this manner. By 1987 the government had expelled approximately 550 foreign missionaries from the country.

Increasing levels of church-state antagonism, including arbitrary arrests of several Catholic clerics, aroused the attention of the international community. Belgium reviewed its technical assistance program to Burundi in light of the human rights abuses surrounding the Church. The French also began to take notice. Suddenly the antagonism of church and state began to threaten a large amount of foreign exchange currency. In addition, several high ranking Tutsi were affiliated with the Catholic Church. The usual and continued nepotism and corruption of the regime also made it vulnerable to challenges of its authority by other Tutsi elites. The 'trigger event' of the coup of 1987 appears to have been the early 'retirement' of several officers in the military due to economic austerity. One batch was dismissed in 1986, with another to follow in 1987. The members of this group, who saw no other options for employment in the civilian sector, orchestrated the bloodless coup d'etat of September 3, 1987.

The Third Republic
The Third Republic was established with Major Pierre Buyoya as the new President of Burundi. Like his predecessors, President Buyoya was a Tutsi from Bururi and so much of the state's apparatus remained in place. Important changes were beginning, however, first with the repeal of restrictions on the Church, and second, with the release of several hundred Hutu political prisoners. Hopes were raised among the Hutu population that real change and political liberalization was in order. Student strikes were held and mainly Hutu students participated. Hutu were said to be engaging in clandestine meetings in the North with Hutu from Rwanda. Ethnic tensions increased and minor conflicts between Hutu and Tutsi began to occur throughout the country. The growing tension and anxiety on the part of Hutus was based in part on fears of a repeat of the 1972 massacres.

Tensions were highest in the North in Ntega, Ngozi Kirundo. Proximity to Rwanda also played a role as there were many Tusti refugees living in the two provinces. Coffee was the main economic support for the region and the drop in world coffee prices in the mid 1980's heavily impacted this region. Many Hutu producers reaped large losses. Competition for scarce social service resources (schools, health care, etc.) in the region also heightened tensions. When a confidential document was discovered that showed that the central government in Bujumbura had requested Kirundo primary schools to report their ethnic breakdown of students, tensions flaired. In a climate where any mention of ethnicity was officially banned, the document seemed to support Hutu suspicions of admission requirements based on ethnicity designed to keep their children out of schools. Finally, local elections where the Hutus won 84% of the vote were seen to be meaningless in a context where the communal level administrator was appointed by the central Tutsi government. These so-called "sham elections" played a large role in exacerbating Hutu frustrations. As tensions began to mount the gendarmes were called in to patrol Ngozi, causing widespread panic among the Hutu peasants who remembered 1972 all too well. In Ntega, Reverien Harushingoro, a known Tutsi collaborator in the 1972 massacres, was seen showing groups of army men the way from hill to hill. This was the spark that set the killings in motion.

On the night of August 14, 1988, a group of Hutu surrounded Harushingoro's house, clearly intending him harm. Harushingoro opened fire on the crowd, killing six people. Violence erupted and the Hutu killed Harushingoro and his family before fanning through the area killing Tutsi and burning their homes. Ethnic hatred caused a massacre of every Tutsi in sight and hundreds of civilian Tutsi lost their lives over the next few day. The army, of course, moved in to restore 'peace and order' and upwards of 15,000 Hutus were killed in retaliation. More than 50,000 Hutus fled to Rwanda to escape. Analysts agree that the fear of an apocalypse reminiscent of 1972 probably played a large role in the initial Hutu killings of Tutsi. The feelings of fear, related to earlier killings of Hutus, were sparked when troops began arriving in the area. Rumors, combined with the collaboration of Habushingoro in speaking with the army, led to the preemptive violence by Hutus.

Response to Massacres of 1988
The future of ethnic relations in Burundi was eerily predicted in the letter written by Hutu protesters to President Buyoya following the massacres: The predicted violence did not happen right away, however. Compared to 1972, the difference in the regime's response to the massacre of 1988 was quite positive and dramatic. The difference was due in large part to international pressures. Unlike in 1972, when the U.S. and the world turned a blind eye to violence in Burundi, changes in international perspectives on human rights meant that the killings in 1988 generated outrage on the part of the international community. The World Bank and the U.S. threatened to cut off financial assistance unless "national reconciliation" was achieved. International pressure did what it was supposed to do. On October 12, 1988 Buyoya agreed to a major reshuffling of his cabinet with the number of Hutu cabinet members increasing from 6 to 12. He also selected a Hutu to become Prime Minister.

Under President Buyoya, Burundi was to undergo a period of increasing liberalization and integration of Hutu members within the government. These efforts culminated with the election in August of 1993 of the first ever Hutu President in Burundi's history, President Melchoir Ndadaye. This did not happen without struggle. The army continued to be dominated by Tutsi 'hard-liners' who wanted no part of democratization. Additionally, revolts by Hutu in November of 1991 in selected parts of the north of Burundi (Cancuzo and Cibitoke) were systematically repressed and several thousand Hutu were once again killed. President Buyoya also survived an abortive coup by Tutsi hardliners in March of 1992.

The election held in August of 1993 pitted President Buyoya, a member of the traditional UPRONA party (primarily the party of Tutsi- but also claiming some Hutu who were encouraged by the liberalizing changes in the party) against the Hutu Melchoir Ndadaye, part of the FRODEBU party (Front de Democrates du Burundi) which came to represent the majority of Hutu in the country. The election was a landslide victory with FRODEBU clearly gathering the majority of votes. (See Map 4) President Buyoya stepped down gracefully and the Hutu people rejoiced at the fact that one of their own was in power. The rejoicing was to be short lived. In October of 1993 Burundi was once again thrown into the by now all too familiar downward spiral of ethnic violence and killing.

Ethnic Violence Begins Again
The violence started when several units of the Tutsi dominated army stormed the Presidential Palace on the night of October 21, 1993, and abducted, and subsequently killed, President Ndadaye. They also killed the President and Vice President of the National Assembly and three cabinet ministers. When news of the assassinations got to the communes, Hutu supporters of the FRODEBU party reacted with rage. Every Tutsi in sight, as well as Hutu UPRONA supporters, were seen as legitimate targets for attack. Tens of thousands of innocent Tutsi civilians lost their lives in fighting that was originally contained in the Northeast and Central parts of the country. Men, women, and children were hacked to death with machetes and Tutsi houses and fields were burned. The subsequent retaliation by Tutsi soldiers was equally horrific: entire Hutu communities were attacked by soldiers with machine guns. The coup of October 1993 ultimately failed but the damage had been done. An interim government ruled for a while and another military coup eventually displaced the ineffective 'transition government' in July of 1997.

Burundi is now locked in civil war and has been since 1993. Burundi is ruled by a military-controlled dictatorship lead by former President Buyoya, who came once again to power in the coup of July, 1997. Recently, the government has taken to requiring Hutu to live in 'regroupement' camps in the interior of the country. This aids in solving the problem the army had faced in controlling the rural population over the past few years. "Hutu rebels" had continued to prevail in the communes and fight against the military, which controlled most of the towns. Conditions in the rural camps are said to be "horrific" in the words of one aid worker. The prevalence of disease and malnutrition is high. Sporadic violence continues throughout the country and several hundred people are killed every day. Very conservative estimates place the number who have been killed in the range of 160,000- 200,000 people. The conflicts in neighboring Rwanda and Zaire have also affected the situation in Burundi. Rwandan Hutus' systematic massacre of over 500,000 Tutsi has made many Burundian Tutsi reluctant to ever reconsider liberalization or democracy. They literally define the "problem" as a matter of Tutsi ethnic survival. The possibilities for a solution to the current crisis between Hutu and Tutsi seem limited at best and bleak at worst.

The reasons that the Tutsi military ended Burundi's experiment with democracy in 1993 originate with two facts: 1.) Buyoya was planning to reduce the Tutsi presence in the military and to allow a more equitable representation of the population prevail, and 2.) Buyoya planned to return land to Burundian refugees who had fled Burundi in past years. The Tutsi hard-liners realized full well that the lands would come from current Tutsi occupiers and they felt that this was unacceptable. The actions of two units of the military were unfortunately enough to spark the violence that has consumed the country to this day.

Hypotheses Linking Environmental Scarcity & Conflict in Burundi
It is clear that a multitude of factors were evident in the pattern of violence that has dominated Burundi since independence. A look at several trends over time in Burundian society, with the political situation as a backdrop, can perhaps shed some light on the situation. Several hypotheses of how conflict might interact with environmental scarcity in Burundi should be presented at this time. They include the following: The evidence in support of these possible hypotheses will be examined in the sections to follow.


Burundi is in the early stages of the demographic transition. While death rates from disease have fallen over the past 50 years or so, birth rates have not slowed enough to prevent the population from growing exponentially. This increase in population has contributed greatly to the pressures on the land. As mentioned earlier, Burundi is one of the densest nations in Africa at 230 people per square kilometer. If only arable land is considered, Burundi's population density approaches 551 people per kilometer squared. The trends in crude birth and death rates over the past fifty years in Burundi are presented in Chart 10. This has resulted in the exponential growth noted previously in Chart 1. Burundi's population growth trend best fits an exponential curve. If current growth continues, Burundi's population will approach 12,000,000 by the year 2030. (See Chart 11). If this happens, Burundi will have a total population density approaching 470 people per square kilometer. Clearly Burundi must find a way to maintain a sustainable population by lowering its birth rate before this happens. To date, this has not been the case. Burundi's total fertility rate is very high at 6.8 children on average per woman. Both Burundi and Rwanda have had little success in lowering their fertility rates. As is evident in Chart 12, there has been little change in this rate since 1950. Today less than 1% of Burundian women say they use birth control at all. Changing this fact is crucial to Burundi being able to maintain a healthy population. The population is not divided evenly in Burundi. Map 3 shows population densities in Burundi by province.

Agricultural Sector: Food

Burundi has been able to maintain increases in food production since 1961. Chart 13 shows the food production index for Burundi between 1961 and 1995. The index is based on 1979-1980 crop production. As is evident from the chart, production in 1994 was 30% higher than in 1979. Burundi also compared favorably with both the World and with Africa in terms of raising production levels. This growth was obtained primarily through improvements in agricultural practices, such as increased us of fertilizers. Chart 14 depicts the growth in total fertilizer consumption in Burundi.

The growth in production is misleading, however, when you consider Burundi's population growth and its relationship to total available land. Chart 15 illustrates the fact that almost all of Burundi's arable land that is currently being used for farming. As is evident in the graph, there has been almost no increase in arable or permanent cropland since 1969. Chart 16 shows the decreasing farmland available per person. The decline began in 1970 and has been dropping ever since. The net result is that per capita food production has also been steadily declining since 1979. (See Chart 17.) The precipitous drop in 1995 probably reflects disruptions in the agricultural season due to ethnic conflict. Burundi, although historically self-sufficient in food, is becoming increasingly unable to feed its people. It is interesting to note also that although the world on a whole has been producing more food per person since 1979, Africa has done worse during that same time period.

Agricultural Sector: Coffee

As noted earlier, coffee makes up 80-85% of Burundi's foreign exchange currency. The crop is also the main source of income for the majority of Burundians. Fluxuations in price or climate quality for growing coffee beans can seriously threaten Burundi's economy. The world price of coffee dropped in the 1980's. The effect of this drop is clearly apparent in Chart 18 which shows a steady but very slow growth in the GNP from 1980 through to the present. Charts 19 and 20, Gross Domestic Product Per Capita and Gross National Product Per Capita, both illustrate that the economy as well, did not keep up with the country's population growth. Since the early 1980's, the average Burundian has been getting poorer and poorer over time. There is little evidence to suggest, however, that environmental degredation led to decreased coffee production. Chart 21 shows the production of coffee from 1971 to 1989. Although coffee production has fluctuated over time, in response primarily to changes in world coffee prices, there is no evidence to suggest a trend of decreasing coffee production over time.

Agricultural Sector: Cattle

Cattle production is important culturally to many Burundians. Cattle in Burundi generally are not exported, they are stored as wealth or they are eaten at festivals, weddings, and parties. Chart 22 examines cattle stocks in Burundi since 1961. In 1980, total stocks of Burundian cattle started to decline rapidly. The reason for the precipitous decline in cattle production is unclear. The government of Burundi states only that "counts are inaccurate" and that "people seemed to be slaughtering more of their animals for cash." One possible explanation might be the decrease in available land as more marginal grazing land was converted to food production. Another possible theory might be that as coffee incomes dropped, rural farmers began relying on their cattle as a "saving account" to support them in hard times. The data could, however, simply reflect a change in reporting or data collection methods. The data are somewhat interesting but it should be noted that 98% of Burundi's total food supply comes from crop foods, and only 2% from cattle. Although cattle are culturally important to Burundians, the issue of cattle production is unlikely to have affected the majority of Burundian in the same manner that crop food production would have


Fish are taken from Lake Tanganikya by Burundian fishermen at a fairly low rate considering the size of the population in Burundi. The data in Chart 23 shows a somewhat cyclic pattern to annual fisheries catches from the lake. There appears to be no evidence, however, that overall stocks have declined. Data from tow separate sources show very disparate findings, an illustration, perhaps, of the variability and inconsistency of the data. Fish are also not likely to play a large role in ethnic conflict in Burundi as they account for less than 1% of Burundian's food production or diet.


The final data I proposed examining showed that despite the fact that deforestation may be a problem in Burundi, it is one that happened well before independence. Forest cover is currently less than 2% of Burundi's land and it has stayed at that level for approximately the past 30 years. FAO data report that Burundi has approximately only 85,000 hectares of forest cover left, mostly in national parks that are poorly maintained and at risk for people seeking firewood. Changes in access to forest resources, however, are unlikely to have had much effect on ethnic violence in Burundi, as they have been low and relatively constant until quite recently.


The initial analysis showed that, if any renewable resources are likely to have played a part in Burundi's ethnic violence, that resource was access to land. The data also highlight the fact that population growth has played a tremendous role in environmental scarcity in Burundi.

Conservative estimates for ethnic violence, when plotted against population growth as in Chart 24, and cropland per person, Chart 25, do appear to show a somewhat positive correlation between violence and scarcity of land. The sharp increase in population that began in 1970 corresponds with the sharp increase in violence noted in 1972. Additionally, there was a marked decrease in available cropland that immediately preceded the violence of 1972.

What about the spatial data? Did violence occur in Burundi in areas where scarcity of land was a problem, or was there no association at all? Results from studying maps of outbreaks of violence in Burundi (See Map 5.) shows that in the earlier periods of violence, 1965 and 1972, there appears to be little correlation between where the violence began and where population densities were highest. In 1965, violence did erupt in the two most densely populated provinces, Bujumbura and Muramvya, but politics probably played a much greater role in this than environmental scarcity. Bujumbura was the capital city and Muramvya was the site of the deposed monarchy. In 1972, violence began when Hutus in exile from the repression of the late sixties attacked from across the border of Tanzania. The outbreaks of violence were initially focused, therefore, in the South and East of the country, areas with relatively low population densities. It should be noted, however, that the violence in 1972, although it began in the South and East, quickly spread to encompass all of Burundi. Some historical evidence indicates, moreover, that Tutsi were in fact appropriating land from Hutus during the massacres that occurred at this time.

In the more recent outbreak of violence, there appears to be a higher degree of correlation between violence and population densities. (See Map 6.) This higher degree of correlation in 1993 would be appropriate, given that over time land has become increasingly more scarce (as per capita land availability has decreased). In 1991, violence erupted to a smaller degree in Cankuzo and Cibitoke, areas with medium population densities. One more interesting correlation may be noted in Map 7, which shows the results of the Presidential elections of 1993. There appears to be a loose association between population density and discontent with the ruling UPRONA party. Those provinces with the highest population densities were also the ones most likely to register discontent with the status quo and to vote for the opposition party of FRODEBU. Map 8 displays population density, election results and violence on one map.

I also studied a map of coffee production by province, as it has been suggested that in 1988 it was not, in fact, access to agricultural land which caused distress in Burundi's Northern regions, but in fact, declines in world coffee prices in the Eighties. The map presented shows that Burundi's two largest producers of coffee in 1988 were Karuzi and Kayanza, neither of which were the location of the outbreaks of violence in 1988. (See Map 9. ) Violence was most serious in 1988 in Kirundo and Ngozi provinces. It does seem probable to suppose, however, that decline in coffee prices in the Eighties, which led to governmental financial difficulties, was the cause of the Tusti elites fight for power and the Coup d'Etat of 1987 that placed President Buyoya in power. It was the "forced retirement" of several Tutsi military elites that sparked the coup. The uncertain period which followed probably contributed to the likelihood of violence and tension that resulted. Violence was highest, however, in areas where there was the most pressure for land.

Based upon the small amount of evidence presented, therefore, I believe there is support for Hypothesis # 1 and Hypothesis # 3, but not Hypothesis # 2. Most Tutsi were dependent upon the state for their income and resources. The financial viability of the state was and is dependent upon coffee production. Total production of coffee did not appear to be declining throughout the 1960 - 1993 time period. On the contrary, coffee production (See Chart 21), continued to rise throughout the 70s and 80s, although in a sporadic manner. The stagnation of income by the government, and hence the Tutsi, therefore, probably had less to do with environmental limits and much more to do with international market forces.

Hypothesis # 1 and Hypothesis # 3 may also have both been salient possibilities for explaining the violence in 1988, 1991, and in 1993. In 1993 this is especially the case because the violence began when the government was in the process of handing both land and positions within the military to long neglected Hutus. As Homer-Dixon's theory correctly predicts, Burundi has experienced an increase in the amount of violence as environmental scarcity has grown. The theory also predicts the nature of the violence that plagues Burundi today. Violence in Burundi is diffuse and sub-national. It also resulted in first a decline in the ability of the government to control its territory and now an even greater militarization of political and social life in Burundi.

V. Discussion

Policy Implications

The results of this analysis highlight, I believe, a number of important issues in attempting to break the cycle of escalating violence in Burundi. The first notes the role that the international community can play in Burundi. The fact that democratization and liberalization happened in response to international pressure must not be forgotten. The international community has disassociated itslef from Burundi since the coup d'etat placing President Buyoya in power once again. Imports and exports from Burundi have been stopped. I believe that this strategy is misguided in light of evidence that economic deprivation is only likely to heighten tensions in Burundi. Coffee exports in particular should be resumed.

The international community also, I believe, has a responsibility to actively prevent further tragedy in Burundi. It is not unrealistic to suppose that a similar level of violence to that of Rwand in 1994, where over half a million people were killed, might also happen in Burundi. An international peace-keeping force could prove invaluable in stopping the immediate violence from escalating. If Burundi's army continues to be controlled by the Tutsi, any future prospects for peace seem unlikely. Indeed, the prospects for peace decline with each new death. Violence and human rights abuses continue daily in Burundi and must be stopped by the international community if massive and large scale massacres are to be avoided. The fact that many Hutu males have been placed in "regroupement" camps by the army is ominous, especially in light of the 1972 massacres in Burundi.

Long term solutions to Burundi's problems include an increased focus on holding those responsible for the killings accountable. An effort to improve the judicial system in Burundi is essential, and a public and honest accounting for past crimes is the only way to ensure futrue security and to build trust in the political system.

An increased focus on population issues must also be incorporated to help solve Burundi's long term problems. Access to agricultural land is only likely to decrease, and ignoring the reality of demographic pressures is irresponsible. Efforts should be made to improve access to contraceptives as well as to improve the education of women in an effort to lower the very high fertility rates of Burundian women.

Finally, improvements in agricultural productivity are essential for Burundi's long term stability. It is clear that Burundi will be forced to improve the productivity of the land if the country is to survive. Although some improvements, such as fertilizer use, have occurred, there is tremendous potential to be explored. Research is urgently needed to develop integrated, sustainable systems capable of producing greater yields from the land. This research should take into account the physical limits inherent in Burundi's climate and geographical location.

Problems with the Analysis and Directions for Future Research

The most serious problems with the analysis in this paper stems from the fact hat accurate information about violence in Burundi is difficult to obtain. There have been very strong incentives for the government of Burundi to hide evidence of massacres and deaths. The international community has been unable to document adequately even past abuses from 1972. More recent violence, because it is still continuing today, poses its own problems in terms of locating accurate information.

Information about ethnicity and access to property or land would also be invaluable in an expanded analysis of this topic. For political reasons, discussion of ethnicity was banned for many years in Burundi. It is only recently that the Tutsi elite have begun to speak about ethnic differences, usually in an attempt to defend the need for a Tutsi dominated army to "protect" them as a minority in a majority Hutu country. Further research would look at issues of land ownership in Burundi and specifically at transfers of land titles from Hutu to Tutsi or vice versa.

Finally, although there does appear to be a correlation between environmental scarcity an the more recent violence in Burundi, it is difficult to prove causality. This problem was mentioned earlier in connection with Thomas Homer-Dixon's analysis of violence and environmental scarcity. Case studies of the correlation between environmental scarcity and violence are only ne first step towards proving the possible links.