Study II, which this chapter summarizes, examines the effectiveness of international monitoring (early warning) and management of the Rwanda conflict. Using the techniques of critical policy analysis, the study weighs the formulation and execution of policies against their stated objectives as well as accepted international norms for the behaviour of states and organizations. The research is based on a number of sub-studies, secondary sources (books and articles by academics and journalists, media studies, reports, etc.) as well as considerable primary data collected through interviews and document searches in the UN system (New York and Geneva), the NGO community, and visits to national capitals in Europe and North America (Paris, Rome, Brussels, London, Washington and Ottawa), and in Africa (Nairobi, Kigali, Kampala, and Dar-es-Salaam).
The study begins with the refugee problem prior to 1990, follows the civil war, then covers the build-up to the coup on 6 April 1994. The following 10 weeks are traced to understand the tardy international response to the genocide of more than a half a million persons mainly belonging to the minority Tutsi community, but including moderate Hutu political opponents of the regime. The concluding historical analysis reviews the security issues of the refugee camps in Zaire and the displaced persons camps in the south-west corner of Rwanda.
Actions and reactions in the developing conflict
By failing to deal with the festering refugee problem prior to 1990, both the Rwandese and the Ugandan governments set the stage for future conflict. Although the issue of Rwandese refugees in Uganda defied easy solutions, opportunities that existed remained unexplored or were not aggressively pursued. With the exception of Tanzania, the regional states were either indifferent or part of the problem. States further afield showed little interest. UNHCR was overburdened, understaffed, and lacked political or economic leverage to develop the requisite pressure to help resolve the issue, which, at that time, seemed minor in the global scale of refugee problems.
However, the refugee problem was becoming explosive. The build-up of tension leading to the 1990 invasion by the Rwandese Patriotic Front (RPF) forces was accompanied by many tell-tale signs, but was inadequately monitored. When the invasion was a fait accompli, however, it caused considerable international concern and reaction, both in the region and in Europe. France and Zaire came to the aid of the Rwandese government. Other actors, including Belgium, the OAU, and key regional states initiated diplomatic efforts to defuse the conflict.
One source of concern related to the principles at stake. The RPF forces who attacked across the border from Uganda consisted not only of refugees invoking their right of return, but constituted a significant segment of the army of Uganda. The invasion violated basic norms designed to ensure stability in relations among states; these are particularly well-developed in African regional international law. Moreover, those who undertook the early diplomatic rounds recognized that the ethno-political situation in the Great Lakes Region was delicately balanced, had recurringly exploded in Rwanda and Burundi, and could do so again.
The initial diplomatic efforts eventually led to the Arusha peace talks, initiated and led by the OAU and Tanzania. The process received considerable international attention and support and resulted in a comprehensive settlement. The United Nations assumed formal responsibility for overseeing the implementation of the Arusha Accords, but failed, however, to make adequate use of the OAU and local African states in this regard. As a consequence, there was a disjuncture between the mediation and implementation phases that contributed to undermining the Accords.
From 1990 onwards, civil violence against the minority Tutsi community and regime critics gradually escalated. Observers commonly linked the violence to the civil war, either as retaliatory measures or as warnings to the advancing RPF forces. However, two authoritative reports - one by an independent International Commission of Inquiry, and another prepared for the UN Commission on Human Rights Ð suggested a more radical and comprehensive design that foreshadowed events to come. Both reports determined that the killings were genocidal in nature and that existing authorities were substantially involved. As the war continued and the outlines of a peace formula took shape, additional reports indicated that Hutu extremists were organizing and arming themselves to derail the peace process and massacre "internal enemies". The creation of documented structures of violence (death squads, death lists, and, later, hate propaganda inciting violence) provided warnings of a potential genocide.
The UN Commission on Human Rights took little notice of its own report. Except for pointed diplomatic protests by Belgium, the findings of the international human rights inquiry were mostly filed away in national ministries and the UN system. With the partial exception of Canada, no state observed the recommendation to impose strict human rights conditionality on aid transfers. Some states were indifferent, others were concerned but concluded that the ongoing democratization process and the peace talks required their continued economic and political support, particularly since the peace agreement under negotiation would produce a new governmental structure that it was hoped would address human rights abuses. In this way, donors became hostage to their own policies.
Human rights organizations and states were also at odds on the issue of arms supplies. In 1992 and 1993, the former recommended that states (France and Uganda were obvious targets) cut off all arms supplies to the parties in the conflict. France openly defended its role and the right of a sovereign state to support a beleaguered friendly government. Uganda denied any involvement in helping the rebel army, yet its territory constituted the rear base for the RPF forces.
Would aid conditionality and an international arms embargo in the 1990Ð93 period have defused the conflict and prevented the genocide? Effective use of human rights conditionality is difficult, requiring fine-tuned and timely intervention. Arguably, there were windows of opportunity, particularly in mid-1992, when more pressure could have been put on the Habyarimana regime to deal with the extremist forces as well as the critical issue of impunity. Also military assistance (direct and indirect) to the Rwandese protagonists could have been calibrated better with the continuing peace process, particularly in dealing with the central issue of extremist forces who opposed the Arusha process and resultant Accords.
Though such speculations are debatable, firmer conclusions can be drawn about what did happen. By not standing firm on human rights conditionality, donors collectively sent the message that their priorities lay elsewhere. By permitting arms to reach the Rwandese protagonists, the possibilities for demilitarizing the conflict were reduced. Arms supplies reinforced the determination of both parties to seek a military and forceful solution to a political conflict. They strengthened the RPF's ability to advance militarily. They permitted the government to equip and expand its armed forces as well as para-military units, both of which became involved in the genocide.
When, as a result of the Arusha process, the Hutu extremists were excluded from the key instruments of the Broad Based Transitional Government (BBTG) and marginalized in the political process, alternative strategies were not developed to defang those extremists. The UN force (UNAMIR) sent to oversee the implementation of the peace agreement was given a mandate tailored to a classic, minimalist peacekeeping operation. Yet the force faced a situation considered by many - including some of those who planned the operation - as dangerously unstable. As the architects of the Arusha Accords had foreseen, conditions in Rwanda suggested a mandate with broader powers to protect civilians and seize arms caches. Further, the UN Security Council established a force that was structured and financed to satisfy a cost-conscious United States, increasingly unwilling to support UN peacekeeping, rather than to meet the needs on the ground. The force was inadequately supported and slowly deployed relative to the need for speed - considered essential to maintain the peace process - though relatively rapidly given the normally cumbersome UN procedures. The operation had no flexibility to respond to changing circumstances, in particular those caused by the crisis in neighbouring Burundi in October 1993.
In the months immediately preceding the genocide, many additional signs indicated that the implementation of the Arusha Accords was faltering and that massive violence was being planned. The air was full of extremist rhetoric on radio, in public rallies and at official cocktail parties. There were assassinations and organized violence. Detailed intelligence reports were passed to New York and the Belgian military authorities by the unofficial UNAMIR intelligence unit documenting the military training of militias, hidden arms caches, and plans for violent action. Unequivocal warnings reached the UN Secretariat in January regarding a planned coup, an assault on the UN forces to drive them out, provocations to resume the civil war, and even detailed plans for carrying out genocidal killings in the capital. The cable was placed in a separate Black File, designed to draw attention to its content, and circulated to several departments in the UN Secretariat. However, senior officials in the Secretariat questioned the validity of the information and made no contingency plans for worst-case scenarios. Similar intelligence failures were evident on the state level, particularly in France and Belgium, both of which had a considerable capacity for overt and covert information gathering in Rwanda at the time.
Thus pieces of information were available that, if put together and analyzed, would have permitted policy-makers to draw the conclusion that both political assassinations and genocide might occur, and that the scale would be different from past patterns (1959-1963; 1991-1993) of "just" hundreds or thousands of victims. Yet this analysis was not done. Although some had available fragments of prescient and significant information, the enormity of the genocide took virtually all by surprise. The failure to anticipate planned and targeted mass murder was particularly significant given the political commitment and actual involvement of the UN in Rwanda, the legal right and moral obligation to act to prevent genocide according to the Genocide Convention, and the enormous cost of a miscalculation.
While mandated to help implement the peace agreement, the UN made no preparations to deal with a breakdown of the Accords, except to withdraw. Nor were there contingency preparations to deal with the plans to scuttle the Accords or the massive violence plotted by the extremists. Generally, the UN Secretariat interpreted UNAMIR's mandate and terms of engagement narrowly, and on several occasions denied the Force Commander permission to search for and seize arms caches. When developments in early 1994 further eroded the peace accords, the Secretary-General and the Security Council threatened to withdraw the UN force, hence strengthening the hands of the extremists. No member of the Security Council came forward to suggest a different course of action. On the contrary, the Council kept UNAMIR on a tight leash with only a three months' authorization, accompanied by admonitions of caution and cost-cutting.
Crisis and response
In the months before the crisis struck, UNAMIRÕs presence contributed to a false sense of security in Rwanda. When events came to a head on 6 April, the UN collectively failed. There was an absence of leadership at UN headquarters in New York. The Secretary-General, travelling at a brisk pace through Europe, misread the nature of the conflict. The understaffed and overstretched Department of Peace-Keeping Operations seemed paralyzed. In the Security Council, the killing of 10 Belgian peacekeepers created a political surge to withdraw, although this was not recommended by UNAMIR's Force Commander nor African countries contributing troops. Information on the genocide under way was already available when the final decision was made to reduce the force drastically.
Once the direction and magnitude of the genocide became undeniable, the UN reversed itself and accepted an obligation to protect civilians. However, the realization of this peacekeeping mission (UNAMIR II) was hampered by the unwillingness of key members to pay for or provide troops, and to match troops with equipment in an expeditious manner. The force was deployed only after the genocide and the civil war had ended.
France's role in Rwanda was significant but marked by multiple contradictions. While warning the Security Council in early 1993 that massacres were a real possibility, France supported a regime that was deeply compromised by human rights violations. France urged the UN, rather than the OAU, to take the lead in monitoring and implementing the peace agreement, but subsequently did little to support UNAMIR I. Nor did France pledge support for UNAMIR II, even though the French Foreign Minister was the first cabinet member of a government holding a permanent seat on the Security Council to identify the massacres as genocide (16 May 1994). With the aid of some of its African ex-colonies, France subsequently undertook a unilateral intervention, Operation Turquoise, endorsed by a Chapter VII Security Council resolution. The action saved many lives and undoubtedly prevented an additional mass outflow of refugees from the south-west of Rwanda, but came very late - two and a half months after the genocide commenced and when the civil war was almost over. Further, the intervention was open to misinterpretation, and did not serve to disarm the extremists or prevent suspected organizers of the genocide from escaping.
After massive numbers of refugees, retreating government forces, and the assumed perpetrators of the genocide crossed into Zaire and Tanzania in April-July 1994, UNHCR warned the UN in New York about the attendant security problems in the refugee camps. The Secretariat took the unprecedented step of examining the issue in a peacekeeping context, but the Security Council proved unsupportive. After significant delays, the problem fell back into the hands of UNHCR, which resorted to a novel and reasonably effective solution to police the refugee camps. The arrangement did not and could not deal with the broader security threats posed by the existence of militarized communities in exile, and this problem was left to fester.
The large concentrations of internally displaced persons in south-west Rwanda presented a domestic version of similar problems. These came to a head when the Kibeho camp was closed in April 1995. The operation resulted in the killing of large numbers of men, women and children, mostly by Rwandese government forces firing on IDPs, but also by extremists within the camp. The disastrous outcome notwithstanding, the coordinated efforts that went into the planning of the Kibeho operation by UN agencies, NGOs and the new Rwandese government were steps in the right direction. Although the execution was faulty, the faults were not inherent in the decision-making model of coordination, which could be utilized in the future. Similarly, the arrangement for providing security for refugees in Zaire exemplifies a workable solution to a difficult problem.
During the whole conflict, but especially after the coup on 6 April, the overall failure of the media to report accurately and adequately on a crime against humanity significantly contributed to international disinterest in the genocide and the consequent inadequate response.
Early warning
Whatever the failures in media coverage, prime responsibility for the failure to read the signals and to respond adequately cannot be placed on the media. Why were the signals that were sent ignored? Why were they not translated into effective conflict management? Failures of early warning are attributable to many factors. The UN was poorly organized to collect and flag information about human rights violations and certainly genocide. There was a failure in both the UN system and the NGO community to link human rights reports to dynamic analyses of social conflict so as to provide strategic policy choices. There existed an internal predisposition on the part of a number of the key actors to deny the possibility of genocide because facing the consequences might have required them to alter their course of action. The mesmerization with the success of Arusha and the failure of Somalia together cast long shadows and distorted an objective analysis of Rwanda. The vast quantity of noise from other crises preoccupied world leaders. The confusion between genocide as a legal term, referring primarily to an intent, and the popular association of genocide with massive murder in the order of hundreds of thousands, created confusion. Finally, a general desensitization developed with respect to mass slaughters, and the possibility of a massive genocide actually occurring seemed beyond belief.
Major states with the capacity to monitor and anticipate the crisis were either not interested, or, if interested, were unwilling to undermine a friendly government. In such a situation, international organizations exist in part to pick up the slack, but neither the UN nor the OAU did so in the Rwanda case. The UN had poorly-developed structures for systematically collecting and analyzing information in a manner relevant to preventive diplomacy and conflict management. The newlyÐformed inter-agency arrangement for early warning (HEWS) was oriented mainly towards humanitarian operations; it was not equipped to detect or analyze political and military warning signals. Within the Secretariat, information collection and policy analysis was divided among the DPA, DPKO and DHA. There was also a disjuncture between information collection, analysis, and the development of strategic policy options. Thus DPA was assigned the responsibility for monitoring events in the region, but not for developing related strategic policy options. One of the most significant sources for early warning, the UN human rights monitoring system, was not part of the information-gathering structure in the Secretariat and, arguably, became isolated from the decision-making process. In the field, the UN had no formal capacity for collecting intelligence; never- theless, UNAMIR, through the initiatives of both the Canadian Force Commander and the Belgian Kigali-sector Commander, succeeded in running minimalist, if irregular, intelligence operations. The other main organization concerned - the OAU - had virtually no capacity at all for early warning data collection and policy analysis.
The shortcomings of early warning in the Rwanda case go further. The issue is not better quantitative data or formal modelling. More simply, the UN lacks a system for drawing on existing information sources, in the region and outside, from specialists in state agencies, academic institutions, rights monitoring agencies, and the various agencies of the UN itself. The UN lacks a specialized unit, without operational responsibilities, for analyzing such information and translating that analysis into evolving strategic options that can be channelled directly to the Secretary-General. Both the UN and NGOs failed to relate human rights monitoring to analysis of the development of social conflict and, hence, to assess the direction of events. When the UN became involved in a peacekeeping mission, the monitoring of political developments was not linked with contingency preparedness. Without contingency planning, the UN was left with a short time-frame and few resources to respond to sudden changes in the situation. This point is critical for two reasons: even under the best of circumstances, it is impossible to pinpoint specific future outcomes of complex social conflict; secondly, the absence of contingency planning limits both what the decision-makers will hear and the options they are willing to consider.
Despite the shortcomings of early warning, at the critical stage the relevant actors dealing with Rwanda knew that the situation was unstable and dangerous. Yet the sustained and careful attention so necessary to successful conflict management was lacking. In part, early action is problematic and preventive diplomacy is inherently difficult because outcomes are uncertain, reflecting the typical complexity of cause-and-effect relations in social conflict. Moreover, policy-makers who are continuously faced with actual crises are disinclined to pay attention to hypothetical ones, even though experience tells us that "prevention is better than cure". The lack of international investment in early conflict regulation signified a more fundamental disinterest in Rwanda. The UN Security Council authorized only a minimalist peacekeeping force, and the Secretariat insisted that UNAMIR maintain a low profile. When the crisis struck, and it became clear that massive genocide was under way, there was still no effective international action.
Conflict management
Throughout, some individual and collective actors did the most with the least under difficult or adverse circumstances. Human rights NGOs monitored the situation. Tanzania struggled to turn the Arusha process into effective preventive diplomacy. UNAMIR I tried to function proactively despite tight reins prior to 6 April; many remaining units - along with the ICRC - bravely sought to save civilians once the killings started.
This could not compensate, however, for the overall failure of the international community to attempt to prevent or stop the genocide, or its very inadequate efforts to mitigate it. In one sense, the inaction can be seen as a result of the propensity of states to be guided by narrow self-interest rather than moral obligations to uphold international norms of justice. However, this propensity has historically varied over time and place; its prominence in the Rwanda case, therefore, requires additional explanation.
No state involved in the conflict happened at the time to have the optimal combination of interest, capacity and neutrality that could have generated appropriate early warnings and translated them into conflict-mitigation strategies. More fundamentally, the Rwanda conflict occurred in a period when the United Nations was acting in an expansive yet highly selective fashion, reflecting a structural mismatch between the responsibilities of international institutions and interests of states in the post-Cold War world.
Revitalized by the end of the Cold War, the UN in the 1990s rapidly expanded its peacekeeping operations throughout the world. Rwanda was added to the list in October 1993. However, the framework for peacekeeping was set by the distribution of power in the Security Council, which represented the world as it was half a century ago. Apart from France, the major powers on the Council were uninterested in a small Central African country that was marginal to their economic or political concerns, and peripheral to international strategic rivalries. By their power of veto and finances, the Permanent Five controlled the peacekeeping and enforcement operations of the UN. The only state with a demonstrated ability to energize the Council in a crisis - the United States - was haunted by memories of Somalia and determined not to get involved in another African conflict. It was also preoccupied with crises elsewhere, especially in Bosnia and Haiti. The lack of interest in Rwanda on the part of the major Western states left France to define a large part of the policy field; the result was to magnify the consequences - negative as well as positive - of unilateralism.
Within the UN system as a whole, there was no locus for assessing key policy questions. How, for instance, can the democratization process be promoted without exacerbating ethnic and regional tensions or creating excuses for human rights violations? How are extremists to be controlled? Moreover, there was too little effort at policy coordination when opportunities appeared.
The rationale for UN peacekeeping is that it provides a neutral force, independent of partisan interest. However, partisan interests can provide motivation and energy to be directed at a problem when a commitment to conflict resolution per se is lacking. This is the conundrum. Without either kind of interest, the UN as a collective actor was unable to mount an adequate peacekeeping force expeditiously and cut through the byzantine problems endemic to UN peacekeeping. The mixÐandÐmatch system of deployment was slow and inadequate. Lacking a powerful patron in the Security Council, the Rwandese operation was subject to cumbersome and bureaucratic procedures that involved delays and inflexibility, and gave insufficient autonomy to the leadership in the field.
The international community might have responded better had the early warning systems generated a clearer anticipation of forthcoming events. On the other hand, conflict management is a function of interest and capacity, not only to ensuring that information is collected and communicated, but to react. In this respect, regionalism appears as a critical and positive force that was not sufficiently recognized or utilized. Structures of conflict resolution and peacekeeping could have been strengthened by more involvement of regional and subregional actors - the OAU and the sub-regional grouping of the states in the Great Lakes area Ð in the decision and management structures. After all, these actors had definite interests in the conflict and a critical stake in the outcome. Strengthening regional mechanisms for conflict resolution and peacekeeping will require financial support from richer states since most of the world's conflicts occur in regions where the parties have the fewest resources to deal with them.
The consequence of these cumulative fault lines in the international system was an inability to stop or significantly mitigate a genocide of immense proportions.
Note:
1. This summary of Study II, Early Warning and Conflict Management, was prepared by Astri Suhrke and Howard Adelman.