|
|
|
|
Early Warning and Early Response, by Susanne Schmeidl and Howard Adelman (eds.)
Bruce Jones
London School of Economics
Janice Gross Stein
University of Toronto
The authors analyze the role of NGOs in providing early warning of the genocide in Rwanda in order to shed light on the strengths and weaknesses of NGOs as ringers of warning bells generally. The authors begin by arguing that on the basis of obvious structural virtues as organizations, NGOs should have some specific advantages in the capacity to provide early warning. However, the authors conclude that the Rwanda experience does not indicate that such capacities are in fact being employed. In order to capitalize on crucial assets NGOs possess in general and, and in fact did possess in Rwanda, the NGO community would have had to: 1) maintain and deepen its connection with local communities throughout an emergency which sapped NGOs energies and preoccupied its leadership; and 2) understand at both policy and operating levels the high value of the assets it had and the importance of applying them to international processes of early warning.
I. Introduction
The frequency and ferocity of civil wars are deeply troubling for the international community and its international institutions, but particularly troubling for non-governmental organizations. Mandated to promote economic and social development, human rights and democratic development, and conflict resolution, NGOs are in the field early and actively. The stylized versions of NGOs and large international institutions are in sharp contrast: the first, lithe, flexible, quick, with a strong presence on the ground in comparison to headquarters, capable of rapid adjustment as conditions change, and with deep connections to local society; the second, large, slow, with cumbersome bureaucracies, requiring elaborately crafted mandates before any action is possible, and with complex procedures for deployment and action.
These stylized versions, admittedly abstractions which mask exaggerate differences, nevertheless suggest very different implications for the capacity of each of these organizational types to provide early warning.
Non-governmental organizations seem to have the virtues which are vices in large international organizations.
In this paper, we examine the role of NGOs in Rwanda in providing early warning of the genocide which ultimately occurred. Rwanda is in many ways a critical case to examine: as we shall show, NGOs were present in considerable strength for years before the genocide began. It can be argued, however, that because the genocide was a planned and organized conspiracy, no organization of any kind could have penetrated the conspirators and gained access to information that would have warned of an impending genocide. We shall show as well that this is an overly pessimistic judgment. Nevertheless, to be fair, we use the criterion of a warning of a likelihood of a major escalation of violence, rather than a specific warning of genocide, which may have beyond the capacity of any organization to imagine.
Analysis of the role of NGOs in Rwanda in providing early warning should shed light on the strengths and weaknesses of the international NGO communities as ringers of the clarion bell. We examine not only whether NGOs succeeded or failed in providing early warning, but if they failed, why they failed, and how their capacity can be improved.
We also review the record of NGO engagement in conflict prevention when it became apparent that the Arusha Accords were at risk. Again, the expectation is that the more flexible and responsive NGO community should be better positioned and more capable of engaging in attempts at conflict resolution than more highly structured, more distant, and more politicized international institutions. .Was the international NGO community, working with their local counterparts, positioned to engage in preventive diplomacy? Did they have the capacity to do so? Did they engage? If not, why not?
II. The Prelude to a Genocide: A Brief History of the Rwandan Civil Crisis
1. The Civil War
A history of the Rwandan emergency might begin with 1 October 1990, when an armed refugee movement, the Tutsi-dominated Rwandese Patriotic Front (RPF), invaded Rwanda from southern Uganda. The invasion was itself the culmination of previous strife of a type frequently seen in Africa, and elsewhere: victims or losers in a conflict seek refuge in a neighboring country, which then becomes a base for invading their homeland. In this case, large numbers of Rwanda’s Tutsi minority, who had been the losers in a violent struggle for political control in Rwanda during the decolonization process, fled to surrounding countries. The largest concentration of Tutsi refugees was in southern Uganda. Repeated efforts by the refugees to return had been of no avail until a generation of Rwandans born in exile launched a new invasion in 1990. The objective was to permit full and free settlement in Rwanda — a demand consistently denied by Rwanda’s President Habyarimana — and to force the regime to accept power sharing arrangements that would give Tutsis significant political representation in the government. The attack was propitiously timed to take advantage of support from Uganda’s President Museveni and a decline in the political and economic fortunes of the Habyarimana regime.
Invasion was quickly met with intervention. On October 5, France sent roughly 150 paratroopers from bases in nearby Central African Republic to bolster the Habyarimana regime. These troops did not engage the RPF, as some believe, but rather backstopped the Forces Armées Rwandaise (FAR) in Kigali, securing the airport and other major sites. Zaire also sent troops to Rwanda, and these did engage the RPF, notably in Gabiro where their presence was insufficient to stop the in preventive diplomacy? Did they have first major RPF victory of the war. Zaire’s troops were recalled shortly after that defeat.
The fortunes of the RPF underwent a sharp but temporary decline after the capture of Gabiro. Their commander, Major-General Fred Rwigyema, had been killed on the first day of fighting. Shortly after Gabiro was captured, two more senior commanders were killed in a FAR ambush. Disoriented by the loss of leadership, the RPF retreated, splitting into two groups, one which melted into the forests and swamps of the Akagera National Park in the east, and one which stole along the Rwanda-Uganda border to the Virunga National Park in the northwest.
In Virunga, the RPF regrouped under Major Paul Kagame, a charismatic Tutsi in his thirties who had earned his leadership position through years of tough fighting: first with the Tanzanians against Idi Amin, then with Museveni against Milton Obote, and finally as Museveni’s Deputy Chief of Military Intelligence. Recalled from a training program at Fort Levenworth, Kagame provided the RPF with the necessary leadership to sustain the rigours of life in the high altitudes and cold weather of the Virunga mountains. From this vantage point they launched a series of guerrilla attacks in northern Rwanda which succeeded, by mid-1991, in making the northern part of the country a region where the FAR could not travel except at high risk. In November 1991, the FAR launched a strong attack on the RPF’s position in the Virungas. When that attack failed, the RPF consolidated its position in the north, creating a de facto RPF-held zone extending along almost the whole Uganda-Rwanda border by early 1992.
2. The Negotiations
While the fighters fought, the talkers talked. Notwithstanding the somewhat peripheral nature of Rwanda to major western powers, the October 1990 invasion quickly triggered a series of regional and international peacemaking efforts. Just two weeks after the invasion, Tanzania called a regional meeting of the Heads of State of Rwanda, Uganda and Zaire to discuss the situation, and, fearing further refugee flows, remained actively involved and became host as well as “facilitator” for the subsequent peace talks. The Organization of African Unity (OAU) Secretariat was also active in recognition of the organization’s principle that African states have a primary responsibility to address regional conflicts. Other actors were soon informed or engaged — the (informal) Economic Community of the Great Lakes Region (CEPGL), the European Union and, more peripherally, the UN. Additionally the governments of Belgium, France and the United States at various times helped to move the process forward. The Belgian government became actively involved within days of the invasion, pushing forward a regional mediation process which achieved a cease-fire within four months.
At the organizational and the state level, warning of an escalation of violence and an increase in refugee flows did trigger a response. That the response was ineffective and insufficient should not mask the evidence that the fighting did serve as a warning and the warning did trigger a response by African leaders. All this activity was ineffective; the initial success of regional diplomacy was short-lived. A formal cease-fire signed at N’Sele, Zaire on 29 March 1991 lasted only to mid-April, when fighting resumed. The limits of regional diplomacy were revealed when a second cease-fire broke down in early 1992. It required a push from France — supported by more limited but parallel diplomatic suasion by the US Under Secretary of State, Herman Cohen — to get the conflicting parties back to the negotiation table. The European Union, Canada, Switzerland, the Vatican and others also counseled peace talks. The result was the Arusha peace process, launched in the summer of 1992, which concluded in a comprehensive settlement signed in August 1993.
3. The Arusha Period (June 92 – August 93)
The Arusha process brought together the RPF, the ruling party, and Rwandese opposition parties which had grown up during the civil war. Supporting the negotiations were international organizations which had a stake or a role in Rwanda, including the OAU, western donor countries, and the UN. The Tanzanian government formally acted as a “facilitator”, and undertook a sustained and skillful mediating effort that was critical in bringing about an agreement.
From the start of the Arusha process in June of 1992 until February 1993, an uneasy cease-fire curtailed RPF activities. The Arusha negotiations began to outline the course of a peaceful transition of power from the Habyarimana regime to a broad based regime. However, by January 1993 the Arusha negotiations were stuck over the details of power-sharing arrangements. A government massacre of Tutsis in the north of the country at the end of January signaled the unwillingness of Kigali hard-liners to agree to proposed compromises. Frustrated by the lack of progress and enraged by the killings, the RPF launched an offensive in February 1993.
The RPF offensive was a major success. Thousands of the FAR’s troops fled in the face of the advancing enemy, many of them deserting the army altogether. The offensive shocked Kigali and threw the government forces into disarray. With the FAR scattering in front of them, the RPF fought to within 23 miles of Kigali, stopping only in the face of French reinforcements and international pressure. RPF, French and Tanzanian military sources agree that had the RPF chosen at this moment to continue to fight, the French reinforcements would have been insufficient to prevent total FAR defeat.
In New York, the UN system responded to the disruption of the peace process. In March, the Secretary-General sent a team to Rwanda which helped bring the parties back to the negotiating table. The Security Council also approved a military observer mission to monitor the Uganda–Rwanda border, designed to stop Ugandan supplies flowing to the RPF. Again, the largest international organization, acting on warnings of further escalation, did respond. Again, the response was inadequate to the scope of the challenge.
As the negotiations neared completion, the role of the peacekeepers was extended. The Accords called for the deployment of a Neutral International Force to oversee the implementation of the agreement, the principal feature of which was the creation of a transitional broad based regime leading to democratic elections scheduled to be held 22 months after the signing. A new national army composed of units from both sides in the civil war was to be formed; the rest would be demobilized. Invoking Chapter VI of the UN Charter, the UN Security Council voted to deploy a UN Assistance Mission in Rwanda (UNAMIR) to oversee the installation of the transitional government and other parts of the peace agreement, in particular, to assist in disarming and demobilizing the two armies.
4. Building and Subverting the Peace
The signing of a peace agreement in August 1993 and the deployment of UNAMIR two months later — i.e. the initial success of what may be characterized as preventive diplomacy — did nothing to halt the rapid deterioration of the political situation in Kigali in the fall of 1993 and spring of 1994. As peacemakers and peacekeepers attempted to put the Arusha structures in place, their opponents — hard-line forces in Kigali who lost out in the Arusha process — worked to knock them down. As the pro-Arusha forces attempted to establish the institutions and mechanisms of power-sharing, the extremists laid the ground work and then set in motion a radical alternative to the Arusha power-sharing plan: a mass genocide against the Tutsi population and a return to war against the RPF.
Their efforts were bolstered by the assassination of President Ndadaye in Burundi in October 1993, and the mass killings which followed. In that month, a five month old experiment in democratic power-sharing in Burundi collapsed when the Tutsi-dominated army launched what turned out to be an abortive coup which nevertheless undermined the nascent democratic system. This in turn generated a week of mass killings, Hutus forces took their revenge on Tutsi civilians and the Burundian army retaliated with bloody force. Later estimates would place the dead at between 35,000 and 50,000 split equally between the two ethnic groups, but at the time in the region the killing spree was widely reported to be far larger — as high as 150,000 to 200,000. The reality was bad enough, and sent tens of thousands of Burundians fleeing from the country.
The killings in Burundi helped push Rwanda faster down the slope into collapse and genocide. Since February 1993, it had been moving steadily down that slope, but until October 1993 there was still a reasonable hope that the course could be reversed. After the killings in Burundi, the course was firmly set. Among other things, the October 1993 killings in Burundi sharply increased the fears of the Rwandan population. To western ears, the Habyarimana regime’s propaganda that the RPF were returning to Rwanda to re-impose slavery and other evils on the Hutu, rang shallow and cynical; yet across the border in Rwanda’s sister country, a Tutsi military was executing Hutus in their masses (the stories of Hutus massacring Tutsis did not make into the extremist propaganda.) Moreover, the killings had come only months after Burundi had attempted a democratic experiment similar to that now being proposed for Rwanda.
What is more, the UN peacekeeping force being sent to Rwanda to protect the populations from such a breakdown (this was not UNAMIR’s mandate, but it was very much how it was perceived in Rwanda), was late arriving: UNAMIR Force Commander Dallaire and his advance mission had the misfortune to arrive in Rwanda on the same day as the assassination in Burundi. Not only was UNAMIR immediately required to divert some of its (not over large) force to the Burundi border region, but the stark contrast between the late, partial arrival of UNAMIR and the killings in Burundi meant that UNAMIR started life in Rwanda under a cloud of Rwandan skepticism (which turned out to be well-justified.)
5. Descent into Genocide
The deployment of UNAMIR in Rwanda did little or nothing to slow Rwanda’s disintegration. By February 1994, the program of assassination and disruption reached a fevered pitch with the killing of leading opposition member, and key moderate, Felicien Gatabazi. On February 23, UN High Commissioner for Refugees (UNHCR) special representative Michel Moussali warned that Rwanda faced “a bloodbath of unparalleled proportions” unless action was taken to restore stability. In the end, it was precisely when action was taken to restore stability — when Tanzania and the OAU convened an emergency session of the Arusha principals on April 5 — that the bloodbath was unleashed. The plane carrying Presidents Habyarimana and Ntaryamira from that meeting was shot down on April 6, and genocide begun in Rwanda.
The genocide was planned and to a large extent controlled by a tightly organized group of extremists from within the Habyarimana power structure: members of the ruling Mouvement Révolutionnaire National pour le Développement (MRND) party, leaders of the Presidential Guard, the interahamwe and impuzamugambi militias, and members of the hard-line political grouping, the Coalition pour le défense de la République (CDR). In the first days of the killings, this group massacred the Tutsi population of Kigali and wiped out the ranks of moderate politicians and civil society leaders, most of them Hutu. Over the next three months, unchecked by any international force, the extremists systematically slaughtered Tutsi populations across the country, killing hundreds of thousands of people before the RPF’s victory on 17 July 1994 drove them into final retreat.
The central motive for the genocide was to retain political power and the economic rewards that went with it. Given the history and ethnic structure of Rwanda, the political contest over who would control the state machinery had developed along a deepening majority-minority divide. Members of the Hutu majority community which planned, organized, and directed the genocide stood to lose power as a result of the power-sharing arrangements negotiated in Arusha. Additionally, some feared that the RPF would use its legitimized entry into national politics and foothold in the new national army to engineer a coup. In either case, the Tutsi were perceived as the winners. The fear which this prospect generated among the Hutu population — and which took extreme forms among some of the power-holders and their followers — must be understood against the historical memory of Tutsi overlordship before and during the colonial period, and the practice of treating power transfer as totalistic, within a context of eliminationist psychology. As the direction of civil war became clear, and power-sharing was on the horizon, Hutu extremist ideologues deliberately exploited this history to whip up ethnic fears to create space for their program of mass killing. They were aided in this by the repeated displacement many Hutus experienced as a result of the civil war, and by events in Burundi, which lent credence to their radical portrayal of the ‘Tutsi devils’ of the RPF.
It is critical to be clear that the genocide was not spontaneous, not an eruption of ancient tribal hatreds, as it was quickly portrayed by the western media. Rather, this was a planned, coordinated, directed, controlled attack by a small core, with the support of the senior elements of the state machinery, and arguably as many as 100,000 but possibly as few as 25,000 ‘henchmen’ — a large number of people, especially at the larger estimate, but virtually all under the direct control of central authorities.
All sources estimate deaths in the Rwandan genocide between 500,000 and 1 million, with the bulk of estimates falling at the larger end of the spectrum. Analysis of pre-genocide census data, counts of refugee populations, and estimates of deaths, lead to an informed estimate of 800–900,000 deaths, the figure chosen by Studies II and III of the Joint Evaluation, as well as by Gérard Prunier, in his The Rwanda Crisis (1995). The rate of killing in Rwanda, roughly 75,000 victims per week sustained over a twelve week period, exceeds that of Cambodia, and equals the industrialised Nazi death machine at its most active point.
III. Prelude to an Emergency: NGO Involvement in Rwanda
1. NGOs before the Genocide
Rwanda, politically marginal in the international community, had long been the site of significant NGO involvement. A number of international NGOs had a long-standing presence in Rwanda: OXFAM and Catholic Relief Services (CRS) since the late 1960s, CARE and others since the early 1980s. NGO work in the country was broad but fairly conventional. Seed distribution, provision of agricultural tools and implements, irrigation: these were the staples of NGO efforts to alleviate poverty in what was then (and is still) one of the world’s poorest countries. OXFAM, CARE, and MSF were the three largest international NGOs. Other major humanitarian agencies included Belgian Red Cross, the International Committee of the Red Cross, and the UN humanitarian agencies.
The onset of the civil war in 1990 was not, for most NGOs, a source of major disruption. Some NGOs concentrated their work in the southwest of the country where average land holdings were smallest (ca. 0.25 hectares, compared to ca. 2.0 hectares in the northeast, the breadbasket of the country.) The southwest was largely unaffected by the civil war, which was confined until 1993 to a zone along the northern border with Uganda and a zone along the northeastern border with Tanzania. Other NGOs, including CARE, which had programs in Byumba province in the northeast, experienced some disruption when the initial invasion occurred in that region. However, FAR reverses of RPF advances soon saw the fighting disperse towards Virunga National Forest in the northwest and the Akagera National Park in the east, neither of which were areas with particularly intensive NGO activity. For the most part, from the first invasion in October 1990 until the February 1993 offensive, NGOs continued to work in Rwanda in traditional development modes of operation.
2. Efforts to Forestall an Escalation of the Crisis
In the years since the Rwanda crisis, many have asked whether development and relief NGOs could not play a more substantial role in preventing humanitarian emergencies, rather than simply responding to them. For the most part, international NGOs in Rwanda from 1990 to early 1993 did not play any active role in conflict resolution or in attempting to prevent potential escalation. There are two important and one minor exceptions.
First, Catholic Relief Services apparently used its position in Kigali, a quite influential position which came through its strong connection to the Catholic church, to support broader church efforts to deal with the question of ethnicity and to promote acceptance of the idea of sharing power within the Habyarimana regime. Few details of this process have been made public, and CRS’ efforts were for the most part bound up in efforts by such figures as the Papal Nuncio to have the church play a constructive role in the Rwandan peace negotiations. That these broader efforts were ultimately a failure does not necessarily mean that CRS’s efforts were wasted or misguided. What little evidence of these efforts exists suggests that CRS’s work was a small part of a broader process and had little either positive or negative impact. In any case, CRS is something of a case apart for NGOs, given its strong ties to the Catholic church, which gives it political weight, but may deny it impartiality.
The lessons of the second exception are more widely applicable to international development and relief NGOs. In 1992, OXFAM launched a program to tackle the violence and conflict in Rwandan society. OXFAM was, according to diplomatic sources, one of the few agencies in Rwanda with “power in Kigali;” the others were Caritas/CRS, and the ICRC. It’s influence came not through its high level connections among official circles in Kigali — indeed, OXFAM staff may be surprised to learn that they were considered to have power in Kigali — but through its important role as a the major funding agency for ‘civil society’ groups. OXFAM used its position to attempt to diminish the recourse to violence in Rwanda society. Most visible among their efforts was a program called Education for Non-Violence and Democracy (ENVD), run under the broader umbrella of a Catholic Church program called ‘Justice et Paix’. 2 The experience illustrates many of the challenges that face NGO efforts to engage in this sort of political programming.
The ENVD program was advertised through local parishes and brought people together to raise their awareness of issues such as ethnic relations, which were rarely discussed in Rwandan society, democratic process, and human rights. Its strength was that the issues addressed were those raised by participants. The program had some success, at least in that OXFAM received calls from participants for more opportunities to engage in this type of dialogue. However, the program was limited in scope — never, for example, attempted in Kigali, the power centre of Rwandan society. In 1993, as the crisis in Rwanda began to deepen, the ENVD was in hiatus, in a period of adjustment as the lessons of the early round were absorbed.
In hindsight it is clear that the program suffered from major weaknesses; these weaknesses are clearly recognized by OXFAM staff who dealt with the program, either in Kigali or Oxford. Most important is the question of having run the program through the Catholic Church. Anne Macintosh, the country director responsible for the program, later argued that OXFAM might have done well to have involved the Protestant churches. OXFAM did attempt to involve Rwandan human rights organizations, but without success. 3 The Catholic Church itself was grappling with the thorny issue of ethnicity in its own ranks. 4 Indeed, as has now become clear, many of the church’s leading figures, up to and including the Catholic Bishop of Kigali, Msr. Vincent Nsengyumva, were deeply implicated in the planning and later the commission of the genocide. Moreover, several of the people who OXFAM had worked with and put through the ENVD, themselves turned out to be accomplices in the execution of the genocide. It is important to note, however, that other members of the Church were actively, and passionately involved in the search for non-violent solutions to Rwanda’s political crisis, including Jean-Pierre Goding, one the ENVD’s principal animators.
Finally, one other international NGO attempted to play a role in the peace process. London staff of Christian Aid who had ties to Uganda’s President Museveni, held secret meetings in Bujumbura with the RPF and the government of Rwanda. However, no one — including the RPF — believed that the Christian Aid initiative had any relevance. Kigali-based diplomats of the period dismissed Christian Aid’s initiative as “trivial.”
3. Shifting to Emergency Programming
By early 1993, NGO’s development work in Rwanda was seriously jeopardized by the consequences of the February 1993 RPF offensive. In retrospect, this is a critical turning point both in the NGO’s experience and in the crisis itself, and is explored at some length below. The offensive produced large numbers of displaced persons, and forced NGOs onto an emergency footing.
This process was furthered in October 1993 when political violence in Burundi produced large numbers of refugees (JEval, III, 29). At least 375,000 refugees crossed into southern Rwanda, and another 250,000 fled to Tanzania. A further 300,000 were displaced inside Burundi and some tens of thousands also crossed into Zaire. Because the NGOs in Rwanda were already on an emergency footing, they were quickly able to divert some of their capacity to dealing with Burundian refugees in the southern province of Butare, where they were concentrated. By the time this process was fully in place, NGOs in Rwanda were completely on an emergency footing, and so remained until April 1994, when the genocide was launched.
In the first chaotic days after the assassination of Habyarimana and Ntaryamira, NGOs scrambled not to respond but to escape. This was not a panicked reaction or an irresponsible one, but a response to attacks on local NGO staff and serious threats to the security of all NGO personnel. NGO country directors (CDs) claim, with justification, that there was no time in the first week to attempt to focus on what was happening around them; getting their staff to safety was an absolute and consuming priority. Efforts to secure local staff were far from successful; every major international NGO lost local staff to the genocide. Local NGOs often fared far worse; many human rights organizations were wiped out in the first days. Having supported efforts to loosen the regime’s grip on power through promotion of human rights and democratic process was the equivalent of a death warrant.
What staff could, fled. Many international personnel were evacuated out of Kigali by air, while others crossed by land to neighbouring countries. Agency staff were dispersed throughout the countries of the region, and it was some time before they were able to regroup and consider their responses. Many local staff were unable to escape at this stage: many parts of the city were under fire; some international NGO staff being evacuated by Belgian, French, and UN military convoys were told not to bring local staff along, for fear that having Rwandans in evacuation convoys would endanger all in the convoy. Many international NGO staff now recall leaving their Rwandan colleagues behind as one of the most traumatic elements of the entire period. 5 Some international NGO staff were able to use their authority — and real bravery — to save the lives of local staff threatened by militias and organize their escape from Rwanda. Many more were killed than saved.
IV. Early Warning of the Genocide: A Missed Opportunity
The genocide took the international NGO community by surprise. Many had been tracking the political situation, and knew that the Arusha peace was unstable. But no NGO saw what was coming in either its scale, its swiftness, or its brutality.
This was not because there were no signals of what was brewing. With the benefit of hindsight, signals can be pieced together to form an alarming pattern: rampantly extremist propaganda in various media sources, notoriously Radio et Television Libre Milles Colines (RTLM), distribution of parish lists to central military committees, orders sent to local administrators to provide standby transport and logistical facilities for “national security purpose”, the open training of militias, the arming of civilian groups. In hindsight, these and other activities signaled a major escalation of violence, if not a genocide. At the time, however, nobody in the international community, NGOs included, accurately estimated the form and scale of the coming escalation of violence. How might this failure have been averted? What opportunities, if any, were missed?
If there was an opportunity, a period during which NGOs might have been able to hear and process the signals of the momentum to genocide, it was in the summer and fall of 1993. Growing extremism in Rwanda was most concretely evident in an increase in the number of ‘small’ massacres of Tutsis in various parts of the country — two or three hundred killed every time a concrete step towards power-sharing was taken. The pattern was there. It was precisely such a massacre, moreover, that prompted the February 1993 RPF offensive. The humanitarian consequences of that offensive produced a quantitative increase in NGO involvement in Rwanda. It also produced, however, a corresponding qualitative shift in the nature of that involvement, one which diminished NGO sensitivity and intelligence capacity precisely when these were needed most.
On February 8th, 1993, the RPF broke out from their front lines a few kilometers south of their headquarters at Mulindi. Mulindi is at the northern end of the highway which leads from Kigali, through the major northern city of Byumba, to the Ugandan border at Gatuna. The RPF advanced southward along this route, as well as striking southwest towards Ruhengiri, which they quickly captured. Within days the RPF had tripled the area under their control. As noted above, the FAR’s resistance was decidedly weak, and by February 22nd, the RPF had moved to within 23 kilometers of the capital, Kigali. France rushed two waves of reinforcements to the country.
The RPF offensive produced a wave of roughly 400,000 internally displaced persons (IDPs). Almost 500,000 people had previously been displaced from northern Rwanda by fighting during the civil war. They had been temporarily housed in camps in Byumba Prefecture and elsewhere, and although some had returned to their homes with the signing of the July 1992 cease-fire, many remained in these camps. The total IDP population in Rwanda now swelled to almost 900,000. As Anne Macintosh, OXFAM U.K.’s country representative, later noted, this was the first of a series of disaster records that Rwanda was to set: the flood of IDPs in February 1993 was the largest, fastest displacement the humanitarian agencies had ever seen.
The efforts of international NGOs to support sustainable economic and social development in Rwanda — as well as those efforts, such as they were, to build up the capacity of Rwandan society to resist an escalation of violence — were then put on hold to meet the desperate needs of the hundreds of thousands of displaced persons. All the international agencies scrambled to respond to these needs.
The international community’s response to the IDP crisis was somewhat slow to get off the ground, but once moving was coordinated and effective. MSF was characteristically first on the scene, providing emergency health and sanitation services to several camps just north of Kigali. MSF asked OXFAM to provide water in three camps which held 95,000 people in total, not 15 kilometers from Kigali. OXFAM fulfilled its specialty niche: John Howard, a water engineer, flew out to Kigali to provide portable water systems and get them running; once operational, they were handed over to MSF and CARE. During the process of setting up water systems in these three main camps, the NGOs were informed of the existence of other camps near Murambi. In the end, NGOs collaborated on servicing a total of eleven camps. 6
From an operational perspective, the international NGO community successfully moved from development to emergency relief modes of operation. Response time was perhaps not what it might have been, but it was sufficient to the task. The needs of the IDPs were ably met, especially when the numbers eased somewhat at the end of March, with the signing of the Kinihara Cease-fire and the creation of a demilitarized zone (DMZ), which covered the RPF’s territorial advances. The cease-fire and establishment of a DMZ allowed many IDPs to return home. This did not entirely end their reliance on humanitarian assistance, however, and CARE and other NGOs began work on providing the returnees with seeds, agricultural implements, and other necessary materials to regenerate economic activity in the area. In general terms, the NGO response to the IDPs was considered technically highly successful.
Why Did Early Warning Fail?
The technical excellence of the international NGO community in responding to the humanitarian emergency ironically diminished its focus on the signals which now began to grow about the plan for a genocide. The ‘quantity’ of NGO engagement with Rwandan society grew during this period, but — from a political perspective — the ‘quality’ declined.
The impact of the IDP crisis was to shift attention and resources away from traditional development activities. More innovative programs tackling issues that were likely to escalate the conflict — such as OXFAM’s ENVD — were largely displaced by relief activities. Development programs did not stop, but no longer did these have the full attention of the NGO’s staff, whose energies, especially at the country director level, were increasingly geared to the management of the emergency operations. Although most agencies ran separate emergency teams, these nevertheless in most cases reported to the same country representative. These were among the most experienced people, whose attention might otherwise have been focused on the worsening political climate.
This diversion of energy and focus to emergency operations had a paradoxical impact. Most agencies began to loose track of the political situation in Rwanda at this critical time. This shift in attention away from the political context was tragic and paradoxical, paradoxical because emergency operations brought three things which should have enhanced NGOs’ capacity to track and analyze political events — money, access to information, and a wider scope of movement within Rwanda — though that movement was sometimes curtailed when the regime imposed travel restrictions in areas where political violence was manifest. Instead, the emergency shifted NGOs into a mode of operations which, while it did not prevent them from ‘hearing’ what was going on in the populations with whom they were working with, focused their energies on other elements of their work, particularly the delivery of large volumes of relief supplies.
Before the IDP crisis, most international NGOs in Rwanda had fairly limited movement in Rwanda. This was not a function of insecurity, although in the north of the country insecurity was something of an issue. For the most part, NGOs restricted their own movements in the sense that few traveled outside the areas where they had specific programs. Their familiarity with conditions in the country was generally dominated by the settings of their development projects. Caritas was exceptional, with projects across the country.
The sudden onset of emergency conditions in February 1993 began to change these conditions. Suddenly, NGOs who had projects in just one region of the country were now responding in a much wider territory, and traversed the entire country in fulfillment of operational demands. Their access to RPF held territory was also at this time considerably enhanced, with reseeding and other programs getting underway in the zone. Moreover, NGOs suddenly found themselves with large budgets and good access to UN and diplomatic sources of information. Whereas in 1992, money for basic equipment was scarce, by March 1993 funding for even such expensive items as cars and trucks was no longer a problem. Even more important, from a mapping and intelligence perspective, NGO country directors (CDs) began to find themselves included in the first circles of information sharing. One CD who had previously had access, only occasionally, to second-secretary level diplomats and low level UN staff, now had access to the UN heads of agencies and the first-secretaries of the various chanceries, both key political analysts and information conduits.
Yet the combination of money, information, and physical access to the country did not result in an enhancement of NGOs’ political analysis or monitoring of the likelihood of escalation. First, NGO personnel simply lacked time and energy. The management of emergency programs sapped time and focus away from both development programming and political analysis not only for those directly involved in delivering relief, but equally for the rest of the development community. According to CDs in Kigali at the time, and the managers to whom they reported in headquarters in the west, NGOs’ knowledge of pre-1993 Rwanda was ‘displaced’ in the race to respond quickly to crisis conditions.
The second reason for this loss of ‘intelligence’ was that in virtually all NGOs, emergency operations are run by expatriates flown in on short notice to meet relief needs. Expatriates coordinate with other expatriates to provide specialist services in water supply, distribution, health, and social services. These expatriates certainly hire large numbers of local staff, often from within the communities they are servicing. In contrast to development programs, however, these local staff are for the most part used in technical capacities, implementing programs designed by international experts. Local staff become less central in the development of programming. Little in the style of emergency response by most NGOs promotes sustained dialogue with the local community whose needs are being met. Relief programmes are usually pre-conceived and pre-packaged, often literally so. These standard operating procedures allow relief agencies to respond quickly and efficiently to crisis which break at undetermined times in unspecified places, and clearly help save many lives. They do, however, tend to distance the NGOs from the lives and politics of the communities in which they are operating. 1993 saw a substantial growth in the expatriate population of Kigali and a commensurate decline in the meaningful contact between NGOs and local partners and peoples. This phenomenon was less pronounced for NGOs which used their development staff in country to run their emergency operations, CARE among them, and most strikingly pronounced for those NGOs which relied on separate emergency teams flown in for the purpose.
This shift from development programming to emergency relief, with its important but unintended consequences, started with the IDP crisis and escalated when refugees from Burundi fled mass killings in that country in October 1993. Many of these went to southern Rwanda, and the shift to emergency footing in Rwanda was complete. Not only did the NGOs already present in Rwanda augment their emergency operations to deal with the influx, more NGOs arrived in the country to lend a hand, most with no experience in the region, let alone Rwanda. The transformation of the NGO presence in Rwanda from one largely doing development work, with substantial contact with the local population, to one dominated by large international relief programmes run by expatriates, was now complete. According to many Rwandans, and a number of seasoned international observers, this shift had the character of “an invasion of the kids” — a reference to the youth and inexperience of many emergency staff.
The loss of sensitivity to the local situation that the shift to emergency mode occasioned could not have come at a worse time. The months following the IDP crisis (March through October of 1993) were arguably the decisive moment in the struggle between those forces in Rwanda who sought to contain violence through negotiation and institution building, and those who sought to undermine the institutions and processes of peace with a massive escalation of violence. It was at this stage that extremist forces in Kigali began to realize that they were on a losing course, both on the battlefield and around the negotiating table, and started to pay serious attention to developing a radical alternative. The groundwork for a genocide was laid at this time, right under the noses of NGOs and other international actors.
Ironically, many of the young men who later joined the forces of genocide — recruited into extremist youth militias — came from the camps which NGOs were providing with emergency relief . The experience of repeated displacement (aid workers estimated that some Rwandans had been displaced 4 or 5 times before April 1994) created deep reserves of fear and ill will among this population, as did revenge atrocities which the RPF committed at this time (Ndiaye 1993, Prunier 1995). Looking back on the experience of responding to the IDPs, a senior ICRC official lamented the lost opportunity to see into this window on what was happening in Rwanda. One can only speculate on what might have been done differently had a more sensitive ear been attuned to this population in 1993.
V. Conclusion
1. The Failure of Early Warning
The story of early warning of the Rwandan genocide from international development and relief NGOs is easy to tell: there was none. This is not to say that NGOs were unaware of a generalized decline in order and security in Kigali, or were unaware of escalating atrocities in the countryside during 1992 and 1993. It is to say, however, that no NGO, international or local, development or human rights, foresaw the escalation of what was in fact a low-level civil war into one of the most intense killing episodes of the century.
NGOs were not alone in their failure to anticipate the genocide. No institution of the international community accurately predicted the unfolding of events in Rwanda. Some particularly prescient individual analysts, whose full-time job it was to interpret political developments, did better: Steve Browning of the U.S. Embassy in Tanzania correctly argued as early as 1993 that the Arusha Accords would meet large scale violent resistance in Kigali; Michel Moussali of the UNHCR predicted a bloodbath; and a CIA officer produced a desk-analysis which drew a worst case scenario of over half a million dead. Most importantly, General Romeo Dallaire of UNAMIR relayed evidence of plans for a genocide to the UN in New York in January 1994. Unfortunately, this cable had little impact in the office of the Secretary General. But as institutions whose function it is to assess the likelihood of political change and escalation to violence, neither the CIA, nor the State Department, nor any branch of the UN anticipated what was to come.
We began by arguing that NGOs should have some specific advantages in the capacity to provide early warning. Whereas the UN and diplomatic communities are remote from the situation on the ground, girded in capital cities and talking only to government sources, NGOs are in the field and talking to local actors, interacting with the communities most directly affected by the likelihood of violence. Most important, NGOs can communicate with local actors in their own language, and to a certain extent from within their own culture, through links to local actors in civil society. These structural features, it is argued, should allow NGOs to ‘hear’ early warning signals which would not reach the ears of diplomats or UN agencies.
The Rwanda experience does not support this argument. In Rwanda, a number of local actors and community groups spoke ‘good words’ to NGOs in French, and simultaneously participated in the planning of the genocide in Kinyarwanda. 7 Despite the fact that they were in touch and collaborating with Rwandans throughout the society, NGOs gleaned no insights which gave them any form of early warning of potential or actual escalation. Moreover, NGOs delivered relief supplies to tens of thousands of men who would later participate in the genocide militias, without learning anything about the development of that movement.
We argued that the failure is explained in large part by the shift to emergency operations, with its attendant increased role for expatriates and its ‘off-the-shelf’ nature. If this shift is one of the explanatory variables, it has important consequences. Crises rarely arise without prior tension and violence; they build up in stages and as these stages escalate, NGOs increasingly shift into emergency operations, reducing their sensitivity to local political conditions. The NGO experience in Rwanda suggests that as the in-country situation becomes more intense, the likelihood of NGOs being able to capture early warnings diminishes, just when this capacity is most needed.
2. The Counterfactual Arguments
We have argued that the shift from development programming to emergency relief helped to blind those seasoned NGO personnel and reduced the likelihood of early warning of escalation in violence. This argument runs up against two important caveats. First, if this argument is correct, it is useful for policy only if the NGOs in Rwanda could have refrained from switching their focus to emergency relief once the number of displaced persons began to climb, or if they could have retained at least some focus on the political context in which emergency relief was being delivered. The first is highly unlikely; NGOs would not refuse to provide emergency assistance to hundreds of thousands of displaced persons.
The second is more likely, but would require explicit changes in standard operating procedures. NGOs, deeply engaged in emergency relief, working under intense time pressure, would nevertheless have to commit to monitor closely the local political context, share that information and analysis with other NGOs in the field, and communicate their analyses back to headquarters. Despite the real difficulties in dividing attention — and resources — this way, it is overwhelmingly in the interest of the NGO community to do so. NGOs in Rwanda were affected quickly, adversely, and dramatically by the escalation of mass violence and by the genocide. It is essential that NGOs engaged in emergency relief retain some of their connection to the local communities they have served in more placid times. It is that connection to local communities, familiarity with local culture, and proficiency in local languages that is the most important asset of the NGO community for all the different kinds of work it does. This asset is also a vital resource for effective early warning.
Second, our argument implies that had the NGOs continued to do largely development work, they would have enhanced their capacity to provide early warning of the likelihood of mass violence. While this counterfactual is in and of itself not difficult to sustain, the critical question is whether this enhanced capacity would have been sufficient? On this point, the evidence is at best ambiguous; indeed, one can be deeply skeptical of this counterfactual. The contact development workers had had with the local population had not bridged the communication gap which had existed before the RPF invasion. Not enough NGO staff, as we have seen, were proficient in local languages, and not enough were sensitive to local cultures and the importance of oral traditions. To be effective as early warners, NGOs in Rwanda would have had to be truer to the stylized version of an NGO we presented at the beginning: rooted in local cultures, connected to local communities, and sensitive to traditions.
The evidence Rwanda allows only a tentative and speculative argument about the capacity of NGOs to provide early warning. The NGO community would have had to:
Had all these conditions held, paradoxically, the greater resources that poured in in 1993 for emergency relief could have been used in part to monitor the changing political context at least as intensively as in 1992. In other words, the switch to emergency relief need not necessarily have involved a reduction in attention to the political context which, after all, would be critical to the work of the NGO community. Under these conditions, it is possible, but far from certain, that more of what was happening in Rwandan society may have filtered through to NGOs; a partially deaf ear tuned to a signal is better than no ear at all. The large international institutions had almost no ear at all.
Endotes
Note 1: Alexander L. George and Jane Hall, “The Warning-Response Nexus,” the Carnegie Commission on the Prevention of Deadly Violence. Back.
Note 2: Other activities included channelling information to Amnesty International and other human rights agencies. Back.
Note 3: As many observers of the Rwandan human rights scene have noted, Rwandan human rights NGOs by 1991–92, were engaged in intensive infighting, and many were increasingly associated with particular political parties. Back.
Note 4: The Papal Nuncio in Kigali at the time had been sent out specifically to help the local church come to grips with the divisive ethnic relations within its own ranks. He had begun to register what appeared to be some small successes when the genocide was launched — with the connivance and participation of most of the Church’s leading figures, including Msr. Nsengyumva. Back.
Note 5: In the aftermath of Rwanda, some agencies — notably Action Aid — have made commitments to their local staff that the agency will support them even if they have to withdraw from a country of operation. How this might be implemented in emergency circumstances is unclear. Back.
Note 6: The breakdown of responsibility was essentially as follows. CRS, Caritas, and the ICRC were the main implementing partners of the World Food Program (WFP) in food distribution; OXFAM and CARE led on water and sanitation, supported by UN Childrens’ Fund (UNICEF) and Caritas; MSF (H) and MSF (B) cooperated on the provision of health services, with the assistance of the Belgian Red Cross (BRC), MdM, and AICF; social services were the responsibility of OXFAM–Quebec, Caritas, and Terre des Hommes. These efforts were, by all accounts, ably coordinated by two bodies, the European Community Humanitarian Office (ECHO) and the UN’s Department of Humanitarian Affairs (DHA). Bimonthly coordination meetings allowed for a reasonably smooth division of responsibilities and a fair degree of collaboration. Back.
Note 7: Kinyarwanda speakers among the NGO community were (and are) few and far between. Back.
Email this citation