Burning Up and Burning Down - Tracing the Flames of Ethnic Conflict in Côte d’Ivoire and Burundi

Once praised as West Africa’s “beacon of stability,” Côte d’Ivoire shocked the world when its bloody civil war erupted in 2002. The unrest ultimately killed over 1,000 people, according to Freedom House. What sparked this conflict and propagated the violence? I argue that the main cause of violence in the 2002 Ivorian Civil War was elites’ strategic manipulation of ethnic identity—specifically, the concepts of ivoirité and nordistes. I present the constructivist view, showing identity redefinition by elites from both the incumbent party and rebel forces, and demonstrate the inadequacy of the primordialist counter-interpretation to explain this specific civil war. The case of Côte d’Ivoire exemplifies how top-down factors—such as identity manipulation by elites—can cause civil war violence. However, not all cases can be investigated on this national level. In many cases, bottom-up factors and local-level dynamics like grassroots mobilization or local community tensions may better explain the causes of violence. Burundi’s civil war from 1993 to 2005, for instance, is better understood through this bottom-up lens. To juxtapose these two differing frameworks, I will also examine the causes of Burundi’s civil war, showing how using solely national-level analyses would be insufficient to fully comprehend the factors that drove Burundi’s civil war.

Constructivists view identity as malleable, contestable, and socially created by both internal and external factors. One way that identity is constructed is through what James Fearon and David Laitin in their 2000 article termed “strategic action by elites,” in which leaders redefine ethnic boundaries to achieve their personal goals. This method occurred in Côte d’Ivoire when both incumbent and rebel elites manipulated the concept of ivoirité, a formulation of identity that privileges certain traits as being quintessentially Ivorian.

In 1995, President Henri Bedie began to change ivoirité from a common national identity to one that distinguished between “pure-blooded” Ivorians in the South from “immigrant, foreign” Ivorians in the North. Since northern Côte d’Ivoire contained large immigrant populations from Burkina Faso and Mali, Bedie used xenophobic rhetoric to publically degrade Northerners as “strangers,” unlike their indigenous Southern counterparts. He bolstered his interpretation of ivoirité by changing electoral rules, requiring both parents of presidential candidates to have been born in Côte d’Ivoire (thereby excluding Alassane Ouattara, the primary Northern representative, from the presidential races). Though Bedie was overthrown in 1999 via military coup, his successors also “used ivoirité to consolidate governmental power in the south at the expense of the northerners,” according to Moya Collett’s 2000 article in Nations and Nationalism.

Insurgent elites responded to ivoirité by creating a counter-identity as “nordistes,” merging various marginalized ethnic groups. As Matthew Kirwin notes in his 2006 study in the Nordic Journal of African Studies, “The identity of nordistes is a construct [...] nordistes formed from the image of solidarity among the rebels and alliances based upon regional ethnic groups.” One main rebel elite, Guillaume Soro, used the nordistes identity to unite disparate rebel groups into one large group, titled Forces Nouvelles. United as nordistes, insurgent elites organized protests and launched a coup d’etat on September 19, 2002, propelling Côte d’Ivoire into violence and civil war.

Ethnic labels were thus the central cause for conflict. Identity construction led to violence through the action of “ethnic entrepreneurs,” elites who constructed exclusionary identities to force individuals to take sides and thereby compete with one another for scarce resources and privileges. Elites then rallied support in two ways. The rebels mobilized around the shared grievances of discrimination and oppression, while the incumbent elites promised jobs and services to Southern voters.

A positive implication of the constructivist explanation for the 2002 Ivorian civil war is its implications for conflict resolution. If elites can construct conflicting identities, then they can also construct identities that facilitate peace. Alex De Waal, for example, argues for the construction of non-violent identity formations and overarching identities that encompass smaller identities to encourage inclusivity in his 2005 African Affairs article. There is thus constructivist optimism for lasting peace if elites reconstruct ivoirité as a form of modern nationalism that encourages national unity.

Despite this strength, the argument has two main weaknesses. First, it does not explain why followers follow. Even if elites attempt to reconstruct identity, what motivates the populace to support elite decisions? One potential explanation for this tendency is the ethnic security dilemma; each group’s increase in economic and political power is interpreted by the other group as an escalating threat, prompting both parties to build up defenses and military capacity and therefore propagating violence. A second weakness of the identity explanation is that ethnic labels can obscure investigations of political and economic factors. For example, violence was more likely to erupt when identity divisions were reinforced by systematic political exclusion or economic marginalization. Ivoirité was institutionalized in laws that denied Northern Ivorians land tenure, public sector employment, and public services; the government often refused to issue certificates of nationality to many people from the north on the grounds that they were not Ivorians.

In contrast to constructivists, primordialists view identity as defined by innate, pre-determined characteristics that do not evolve over time. For instance, Kaplan suggests that humans adopt fixed characteristics of their respective ethnic groups, while Samuel Huntington argued, in his seminal 1993 Foreign Affairs article “The Clash of Civilizations,” that natural cultural differences between civilizations “will not soon disappear.” In the Ivory Coast case, primordialists claim that the North and South have forever hated each other due to historic tensions, and that this inevitably escalated into violence.

While the primordialist approach is appealing in its simplicity and expediency, being particularly useful for journalists who seek easily understandable narratives for a mass audience and for policymakers who must produce quick decisions in high-pressure environments, ultimately there is a crucial weakness in this approach. Ethnic diversity does not always lead to civil war. In fact, in Côte d’Ivoire, there are sixty ethnic groups which peacefully coexisted for years and did not harbor hatred towards one another. As Kirwin notes,“Despite societal friction based on class, religion, and region of origin, Côte d’Ivoire never suffered from frequent civil wars or military coups.” The primordialist view fails to explain why, givens historical ethnic diversity, civil war erupted in 2002 as opposed to other times.

Elites from both the governing and rebel entities strategically manipulated identity in pursuit of personal political power; incumbent elites such as President Bedie, General Guei, and Laurent Gbagbo reconstructed the ivoirité identity to exclude their Northern opponents, while insurgent groups reconstructed the nordistes’ identity to solidify their response. This process of exclusionary identity formulations created antagonist identities, and led to violence. The incentives that lead elites to establish overarching identities, and the best strategies for using ethnic identity manipulation for the sake of peace, remain open questions.

Now I turn to an investigation of Burundi, and with it to a new analytical framework. Burundi’s devastating civil war lasted from 1993 to 2005 and killed over 200,000 people. What explains the onset and perpetuation of this violence? Collier and Hoeffler, in their 2000 paper “Greed and Governance in Civil War,” present a combined model of national-level factors to explain the causes of civil war. I argue, though, that the Collier and Hoeffler (CH) model insufficiently explains Burundi’s civil war because it omits micro-level dimensions at the village and individual level. I pair the village micro-level lens with economic explanations to argue that villages were more prone to violence when they lost agricultural income from decreased rainfall and the individual micro-level lens with Weinstein’s “initial endowments” theory, as established in his 2000 book Inside Rebellion, to argue that individuals were drawn to rebels’ “activist” strategies because they desired camaraderie and the chance to create change.

The combined Collier and Hoeffler model, which integrates elements from the opportunity and grievances models, “predicts poorly the risk of [Burundi’s] civil war,” according to Floribert Ngaruko and Janvier Nkurunziza in their 2005 article. I find the CH model insufficient because it relies solely upon national narratives of Hutu-Tutsi hatred. As Ngaruko and Nkurunziza elaborate, the CH combined model claims that the 1993 Tutsi-led assassination of Burundi’s first democratically elected Hutu president, Melchior Ndadeye, rekindled historic tensions between Hutus and Tutsis in Burundi and thus sparked the onset of war. Furthermore, war was prolonged because the postcolonial Tutsi elite, who came mostly from the southern province of Bururi, excluded Hutus from political power.

The CH combined model therefore interprets conflict as mere manifestations of the cleavage between Tutsis and Hutus; it not only ignores the bottom-up influences that fueled violence but also ignores the impact of alliances. According to Kalyvas, “cleavage” is the dominant, nationwide dimension used to explain conflict, and “alliances” are mutually beneficial links between local and supralocal levels through which local actors advance their local agendas and central actors gain support and national power. Macro-level explanations of Hutu-Tutsi conflict show overarching patterns of civil war violence, but leave many questions unanswered. Why did certain villages experience more violence and rebel recruitment than others? Why did followers follow the elite leaders’ rhetoric? A fuller picture of Burundi’s civil war must consider periphery events that unfolded independently of the overarching ethnic divides, and contributed to the perpetuation of violence at the micro-level.

According to Kalyvas, the micro-level approach asserts that central cleavages on a national scale may differ from cleavages that activate violence on local levels. At the village level, economic factors impacted the village’s magnitude of violence and rebel recruitment. Nillesen and Verwimp in their 2009 article analyzed the effects of local rainfall amounts on coffee production and total village income, since over 95 percent of the rural population relies on rain-fed agriculture as a primary source of revenue. After surveying 872 households across 16 provinces, the researchers discovered that villages experiencing below-average amounts of rain were significantly more prone to violent activity and recruitment by the Forces of National Liberation (FNL), the largest Hutu rebel group. The researchers also discovered that “a drop in a village’s coffee sales in one month substantially increased the incidence and intensity of the civil war in coffee-intensive areas the next month.” Poverty lowered the opportunity costs of participating in rebellions, which increased violence and rebel recruitment in the village. The micro-level approach shows how decreased rainfall generated decreased agricultural wealth, which lowered the risk to the village of violence with other villages, for instance through looting and stealing (often itself intended to compensate for poverty).

The second important micro-level lens is the individual level. A survey of 2,980 ex-combatants in Burundi found that individual fighters “expected emotional benefits played a significant role in convincing individuals to join the rebels,” and that 65 percent of Burundi ex-rebels stated one primary reason for fighting was wish that “they might play a pivotal role in winning the conflict for their side and helping rebuild their country,” according to Gwendolyn Taylor’s 2010 dissertation Choosing to Fight. In fact, many fighters believed they had no other way to establish a stable system, and “thought it worthwhile to fight for a meritocratic system.” This is also supported by the fact that, when the war ended, the establishment of elections gave individuals more peaceful opportunities to create change and express individual political desires in non-violent ways.

Another important factor at the individual level was the desire for camaraderie. Taylor’s surveys revealed that participation in war among boys was “almost entirely volunteered,” so much so that “the fundamental building blocks of the rebellion were friendship groups among Hutu boys and young men.” In one interview, reported in Cyrus Samii’s Journal of Peace report, an 11-year old boy shared a heart-wrenching story: after his parents and siblings were killed, the boy said, “I found myself completely alone, but my new family took me in,” referring to the FNL as his new family. Many more displaced individuals yearned for a sense of belonging and joined rebel groups due to these higher-level concepts.

Weinstein offers two organizational strategies, opportunistic and activist, to explain how individual desires led to large scale violence. In Burundi, since many of the fighters came from above-average incomes and education levels, rebel leaders utilized “activist” strategies to recruit young men and boys. In fact, in the rebel group CNN-FDD, Taylor argues that “individual fighters were, on average, more educated” and thus were drawn to non-material goals. One such “activist” strategy for non-material promises was elites’ strategic manipulation of identity. Burundi rebel elites acted as ethnic entrepreneurs to create a higher purpose for the war through and used identity as a rhetorical tool to rally support, much as did the elites in Côte d’Ivoire.

In response, followers followed because of their individual desires for belonging. In particular, boys and young men were, as Samii puts it, “enticed by the concept of solidarity among members of a shared ethnic group.” Even for the army (FAB), Taylor’s surveys found that “very few individuals believed that the FAB would give cash to individuals who joined,” and therefore joined not because of material benefits but because of higher-level promises. Through strategic manipulation of identity, elite leaders of both the FAB and rebel groups such as the FNL gave individuals a higher-level purpose, and individuals followed because of their desire for change and community, therefore perpetuating violence and prolonging war.

These interactions were further reinforced by alliances, which Kalyvas defines as mutually beneficial links between local and supralocal levels. For Burundi’s civil war, the mobilization process of the one of the largest rebel organizations, the CNN-FDD, consisted of a local “mobilization officer” who collaborated with an “intelligence network” that not only included mobilization officers from other communities, but national coordinators. While local leaders promoted their local agendas by receiving “external muscle” in weapons, prestige, and local power from national elites, in return national rebel leaders increased their own chances for achieving national power by growing their rebel forces and support base.

Hence local dynamics existed independently from national dynamics, and in fact fueled national cleavages. National elite leaders acted as ethnic entrepreneurs, and employed top-down strategies of mobilization. In attempting to understand why followers followed, analyses at the individual level revealed that desires for making a difference and yearnings for camaraderie drove individuals to believe these “activist” strategies. Furthermore, interactions between the different levels complicated the conflict and prolonged violence. This illustrates how bottom up and top down forces worked together to perpetuate violence in Burundi’s civil war.

One strength of the micro-level approach is its ability to explain causal mechanisms. While the macro-level CH combined model simply notices correlations between the onset of violence and factors such as ethnic dominance, low costs of rebellion, or social cohesion, micro-level lenses show the steps through which an observed cause leads to the observed outcome.

A second strength of the micro-level approach is its production of more accurate data and in-depth analyses. Macro-level analyses through the CH model used national statistics of poverty such as GDP, which overlooked the inter-village differences in income. In fact, the CH model concluded that “poverty affects Hutus and Tutsis equally” since the distribution of land was “not significantly different between Hutu and Tutsi landowners,” according to Ngaruko and Nkurunziza’s 2000 Journal of African Economies article.

However, a closer examination reveals that there indeed were land differences, based on rainfall patterns and varying coffee prices. The CH model’s use of national statistics to assess poverty overlooked variations between villages, and thus did not capture the fact that certain villages were significantly poorer than others, and thus were more likely to fight. In contrast, the micro-level approach provides significantly more accurate information. Individual interviews are the most granular and direct sources of information. Furthermore, interviews and surveys with individuals not only uncovered a myriad of personal motivating factors, but also supplied stories and specific examples to accompany the data.

A third strength of the micro-level approach is that it often dispels primordialist assumptions. The CH model for Burundi suggests a primordialist reading of violence, by claiming that Hutu-Tutsi fighting in neighboring Rwanda somehow sparked ethnic hatred in Burundi. This implies, without providing due evidence, that, since two groups fought with each other in one country, these groups naturally fought with each other in other countries.

Micro-level analyses counter the primordialist reading by showing that identities are flexible and can vary by context and situation. These analyses utilize more constructivist understandings that identities are malleable, can change over time, and that different aspects of one’s identity can be emphasized depending on surrounding environmental factors. For instance, if individuals yearn for a sense of belonging, they are often willing to change groups to fulfill that desire. Micro-level lenses counter the danger of an incorrect primordialist reading.

A weakness in the micro-level approach is its difficulties with data collection. It requires immense human resources: language training for interviewers to converse more naturally with civilians, financial resources to distribute materials to multiple provinces, households, and individuals, and time to aggregate complex data and analyze information. A second weakness in the micro-level approach is its challenging policy implications. In high-pressure situations, legislators and decision-makers don’t have the time or energy to fully understand the complexity of local dynamics and smaller-scale conflicts. Instead, decision-makers oftentimes prefer top-down methods and national lenses that present clear, if oversimplified, conceptualizations of conflict. While a weakness of micro-level approaches is its challenging application to policy, micro-level understandings are critical for future policy reforms to encourage more sustainable peace. For instance, Séverine Autesserre argues, in her 2010 book The Trouble with the Congo, that the bureaucratic culture of key international organizations push decision-makers to prefer national-level solutions, and that these top-down initiatives for post-conflict peacebuilding often fail to address the root causes of conflict. Therefore, though the applicability to policy of micro-level lenses is often complex or unclear, it is crucial that policymakers to embrace this challenge to build sustainable peace.

A third weakness is that micro-level explanations may be confusing for the lay reader. Persons unfamiliar with Burundi’s civil war may be “bogged down” by such small-scale and minute details, and may erroneously assume that all instances of local violence are absolutely unrelated and distinct from each other since they are driven by only local factors. Therefore, over-exaggerating the micro-level approach would confuse lay readers by ignoring overarching patterns that connect seemingly separate acts of violence.

To understand the onset and perpetuation of violence in Burundi’s 12-year civil war, both macro- and micro-level explanations must be used. The CH model’s top-down conceptualization of the main cleavage between Hutus and Tutsis must be complemented with micro-level explanations, which reveal that economic factors were crucial in fueling violence at the village level, while grievance-based explanations were more prevalent at the individual level. Local acts of violence were not mere extensions of the major cleavage, but interacted with top-down dynamics and independently fueled conflict. Future research should examine the weight of micro-level vs. macro-level lenses. The answers may have important implications for creating peace.

The cases of Burundi and Côte d’Ivoire reveal how political analyses and subsequent conclusions are contingent upon the level of examination one employs. Top-down analyses that first investigate national-level tensions may be sufficient in some cases. However, many others require bottom-up, grassroots-level explorations. Any understanding of Burundi’s civil war, for instance, would be inadequate without this second direction of analysis. This research reveals that academics and policy-makers must be flexible, employing different analytical tools to fit different occasions and understanding different directions of causes of violence, whether bottom-up, top-down, or both. Only then is it possible to craft efficient solutions that truly address the sources of civil violence. •