Journal of Humanitarian Assistance
The RPF invasion on the night of 1 October 1990 set off a series of responses, including regional and international military intervention, major shifts in the internal politics of Rwanda, and high-level diplomacy. During this phase, diplomacy vied with military assistance, human rights concerns and development aid in a disconcerting tension that allowed opponents of the peace process to develop strength even as the latter moved forward. The war resumed, and so did reprisal killings of civilians in an early indication of what was to come. This "second track" of violence - civil violence as distinct from civil war - ran parallel to the peace negotiations as the extremists organized and consolidated.
2.1. Track I: civil war and the response
2.1.1. From invasion to negotiations
Both France and Zaire sent troops to aid the government of Rwanda at the time of the invasion. Belgian units were also flown in, but only to protect their nationals, and they were soon withdrawn. With the French backstopping the Rwandese army in Kigali and "advising" on the handling of heavy artillery, and with Zairian paratroopers fighting alongside Rwandese troops, the RPF invasion was soon halted. The RPF regrouped in preparation for a long struggle.
The invasion triggered an extraordinary diplomatic activity that demonstrated the vitality of international mediatory structures (Jones 1995). Just two weeks after the invasion, Tanzania called a regional meeting of Heads of State of Rwanda, Uganda and Tanzania to discuss the situation. Realizing that a renewed conflict in Rwanda would at the very least cause a large inflow of refugees, Tanzania remained actively involved and became host as well as "facilitator" for the consequent peace talks. The OAU Secretariat was also active in recognition of the organization's principle that African states had a primary responsibility to solve regional conflicts. As a Tanzanian, the Secretary-General, Salim A. Salim, coordinated closely with Tanzanian mediation efforts. Soon other levels were informed or engaged - the (informal) Summit of Great Lakes Region Heads of State, the European Union and more peripherally the UN. Also the governments of Belgium, France and the United States at various times helped to move the process forward. The Belgian government was visibly involved within days of the invasion, when a troika of top ministers - the Prime Minister, the Foreign Minister and the Minister of Defense - toured the region to promote reconciliation and presented the case to the European Union for support.
A positive communique was issued from the first summit in Tanzania, marking an exemplary transition from war to diplomacy only two weeks after the invasion. Yet the initial success of regional diplomacy was short-lived. A formal cease-fire signed at N'Sele on 29 March 1991 lasted only to mid-April when fighting resumed. The first deployment of an OAU-sponsored Military Observer Team (MOT), as had been agreed to in October-November 1990, was a failure. The only states to contribute observers were Zaire, Burundi, and Uganda. Since none of them was considered neutral by one or the other Rwandese party, the force became inoperative. The Rwandese government insisted on prior clearance for every movement of the military observers, a demand that greatly diminished their ability to monitor cease-fire violations. A second cease-fire agreement (26 October 1991 at Gbadolite, Zaire) called for a revised monitoring force, to consist of military observers from neutral countries. However, the African Neutral Military Observers Group (NMOG) did not become operational until mid-1992.[18]
The limits of regional diplomacy were revealed when also the second cease-fire broke down in early 1992. It required a push from major powers - The United States and France - to move the conflicting parties to the negotiation table. Also the European Union, Canada, Switzerland, the Catholic Church and others weighed in. Eventually, this resulted, in the summer of 1992, in the launching of the Arusha talks towards a comprehensive peace settlement.
Both Uganda and later France pursued a dual policy of supporting their respective Rwandese partners, yet encouraging negotiations. In keeping with international conventions, Uganda formally denied it was supporting a rebel army - a claim maintained even after the RPF seized power in 1994. In fact, Uganda served as a rear base for the invasion, enabling the Front to regroup, recruit and mobilize among the refugee community, and evidently to funnel weapons and supplies to the RPF-controlled area inside Rwanda. At the same time, once the invasion was a fact but did not bring RPF a speedy victory, Uganda cooperated to get negotiations started. The fact that Uganda held the chair of the OAU at the time also suggested cooperative diplomacy. President Museveni agreed at the regional summit in mid-October 1990 to work with his Tanzanian counterpart, Ali Hassan Mwinyi, to persuade the RPF to accept a cease-fire. He further promised to explore the possibility for an African force to be deployed between the combatants once the cease-fire took hold.
2.1.2. France's dual policy
During the invasion and in its immediate aftermath, French policy was decided at the level of the President. The first response was to send "a few boys to help old man Habyarimana", as the head of Africa Unit in the Presidential palace said at the time (Prunier 1995). France had previously done the same for other allies in Francophone Africa and had paratroopers stationed in the region for that purpose. In Rwanda's case, the determination to assist was sharpened by the fact that the rebels had come from and in some respects were part of Anglophone Africa. The social origins and platform of the RPF also made the Front leaders anathema in some French conservative and army circles where the pejorative term "Khmer Noir" was used, a label projecting the horrors of Cambodia's Khmer Rouge onto an African context.
As the military weakness of the government was revealed, however, the Africa Unit in the Elysée Palace and the Military Assistance Office in the Ministry of Cooperation embarked on a policy to rapidly build up the Forces Armées Rwandaise (FAR).[19] With French technical and military assistance, including credits for arms purchases, Rwanda's tiny army tripled from 1990 to 1991 and by the time of the peace agreement (mid-1993) had at least 30,000 men. The French military assistance unit in Kigali (DAMI) provided training and operational guidance under the 1975 military assistance agreement, as amended in August 1992. Numerically, the French were not a major presence. Some 40 coopérants militaires were gradually increased to 100. The size of the paratroop detachment, Force Noroît, first sent in Ocotober 1990, varied according to the rythm of the war. After the first cease-fire in 1991, the force was reduced to about half (168 men), but increased after the RPF offensives in 1992 and 1993 to reach a peak of some 700. According to French authorities, Force Noroît had a strictly defensive mission, including defense of expatriates and securing the airport, and was deployed in and around Kigali. French government spokesmen have consistently and categorically denied reports to the effect that the paratroopers occasionally participated in the war.[20] The French Ministry of Foreign Affairs, moreover, has taken pains to point out that the force also had humanitarian functions in addition to its military duties, including vaccinating civilians and distributing food to displaced persons.[21] Granted that the French military presence was both small and had a restricted mission, the effects nevertheless had strategic and political significance. By taking on a major role in training the rapidly growing Rwandese army, and by sending in paratroopers every time the RPF launched an offensive, the French government clearly signalled that it stood by the Habyarimana regime. This was the interpretation within Rwanda as well, which contrasted with the hostile attitude shown by many regime supporters towards Belgium, precisely because Brussels decided to abort its 1990 intervention rather than turn it into a long-term military presence on the French pattern.
The role played by the Elysée Palace and the Ministry of Cooperation was increasingly questioned by the Ministry of Foreign Affairs, which around mid-1991 recognized a compelling logic of negotiation in a situation it characterized as follows: the RPF might win militarily but not politically, while the government could not win militarily, though it might command the numbers to win politically (Callamard 1995). According to this logic, France could best salvage its interests in Rwanda by promoting a negotiated settlement on terms favorable to the Habyarimana regime. In October 1991, the director of the Africa bureau in the Ministry of Foreign Affairs attempted to bring representatives from the Rwandese government and the Front together in Paris.[22] The efforts did not bear fruit until late spring 1992, when the two parties agreed in Paris on 8 June to open formal peace talks.
By giving military assistance to the Habyarimana regime in time of war, France followed customary alliance politics, rather than the neutral international practice of not supplying arms to belligerents during war, as observed in this case by Belgium. At the time there was no international arms embargo on Rwanda. The three formal cease-fire agreements signed in 1991-92 prohibited the "infiltration of ... war material to the area occupied by each party"[23], but all had been violated by one or the other Rwandese party and hence lost much of their contractual force. Until the signing of the Arusha agreement, therefore, France could argue - as indeed it did - that it exercised the common right of sovereign states to give military aid to a friendly government facing a rebel force. French authorities did not deny selling arms or giving credits to the Habyarimana regime for purchasing weapons elsewhere, but refused to comment on reports of particular transactions.[24]
From this point onwards, and especially as the situation deteriorated in 1994, French policy was shaped by competitive interplay among domestic institutional actors, which had different interests and perspectives. The result was a dual policy that supported negotiations but simultaneously built up the Rwandese armed forces and embraced the regime politically. The policy helped to move negotiations forward at the inter-state level, while at the same time provided political space and resources for the regime to consolidate, and - indirectly - for the Hutu extremist fringe to develop.
2.1.3. The beginning of democratization
The RPF invasion had a transformative impact on Rwandese politics by giving significant momentum to a process of democratization that had started previously, but so cautiously it was hardly noticed. Responding to demands for democratization expressed at the Franco-African summit in La Baule in 1990 and elsewhere, Habyarimana had started to liberalize the political system in 1989-90. The RPF brought the process decisively forward by making radical demands for power-sharing, backed by the force of an invading army. Foreign donors added to the pressures by emphasizing that democratization with a multi-party system was necessary to end the war, and also desirable in itself. Noticing the trend and sensing that the regime was yielding, the domestic opposition became more insistent. Habyarimana responded first by forming a token coalition government. As pressures visibly mounted throughout the spring of 1992, he agreed to a second coalition that had significant representation from the newly formed opposition parties.
The formation of the second coalition government in April 1992 had significant and contradictory consequences. The timing was in some respects not auspicious for democratization - the country was at war, the economy was deteriorating, and ethnicity was re-emerging as a polarizing dimension. In part, multipartisme became a cloak for particularist interests, encouraged ethnic mobilization and fed political fiefdoms that usurped the civil administration, as a disillusioned former minister wrote from exile (Gasana 1995:231-2). Ultimately more significant was the effect on Hutu extremism. The combined pressures for peace negotiations and democratization provided a double-bored attack on the existing power-holders. When political forces later polarized, the extreme Hutu fringe translated this into an absolute struggle with only final solutions. In 1991 and early 1992, however, the most visible political forces were: a "presidential tendency" that was conservative, had a regional base in the north-west, was inclined to defend Hutu supremacy, and relied on a strong party, the militia, the army, and Habyarimana to preserve its power; a second "moderate tendency" composed of various newly formed political parties that sought to replace the Second Republic by non-violent means and institute a vaguely formulated plural political system; and a third, radical force represented by the RPF that aimed to establish a new order, ideologically defined as political pluralism based on a social structure that denied the legitimacy of ethnicity in political mobilization, but suspected by its enemies to entail "une inversion hégémonique en faveur de l'ethnie minoritaire tutsi" (Gasana 1995:224).
The formation of a coalition government that included the "moderate tendency" had a salutary effect on the peace negotiations. The first act of the Foreign Minister of the new coalition - a member of the opposition party, Mouvement démocratique républicain - was to call for political talks with the RPF. This call was picked up by inter-state diplomatic mechanisms and, after intercession by the United States, France and the Vatican, led to the Arusha talks in the summer of 1992. Having included his domestic opponents in the government, Habyarimana permitted them to lead the negotiations at Arusha.
2.1.4. The Arusha period
The Arusha process - as it came to be called after the resort area in Tanzania where the peace talks were formally launched on 12 July 1992 - was pursued at a very high level in the region, drawing in heads of state and foreign ministers. The core negotiations on a future peace agreement had participant-observer delegations from five African states (Burundi, Zaire, Senegal, Uganda and Tanzania with its "facilitation" team), four Western countries (France, Belgium, Germany and the United States), and the OAU. The UN was brought in at the invitation of the OAU, and UNHCR attended as observer. Britain, Canada, the Netherlands and the European Union - important members of the Western donor community in the region - closely monitored the process from their local embassies. Nigeria was represented at the Arusha-linked Joint Political Military Committee (see below).
Formally acting as a "facilitator", the Tanzanian government undertook a sustained and skilful mediating effort that was critical in bringing about the final agreement.[25] The OAU had earlier organized a military observer group to monitor successive cease-fires, and maintained a high-level presence at the negotiations, with either a senior political officer or Secretary-General Salim himself representing the Organization. For Salim, the Arusha process was a challenge to demonstrate that the OAU could make a critical contribution to conflict resolution in Africa. To this end, Salim had just established a new OAU Conflict Resolution Mechanism that would deal with internal as well as inter-state conflict. Formal approval was expected at the Cairo OAU Summit meeting in 1993. Given the tradititional OAU sensitivities on internal affairs, it would be gratifying to have a freshly mediated peace agreement to put before the Summit.
The Arusha process represented a multi-pronged strategy of conflict resolution. A preliminary phase was designed to obtain a cease-fire. The earlier N'Sele and Gbadolite cease-fire agreements were amended to form the basis of the Arusha cease-fire, announced on 12 July 1992. All cease-fires were violated, but the process permitted continuous communication and led to the establishment of a monitoring machinery. In particular, the Arusha process implemented a mechanism agreed to in earlier negotiations: a Joint Political Military Committee (JPMC) ran parallel to the Arusha talks and provided an escape valve that allowed all parties to discuss violations of the cease-fire without derailing the broader negotiations. The negotiations on substantive issues of the conflict, which started in mid-1992, covered the rule of law, the establishment of transitional institutions, the repatriation of refugees and resettlement of the internally displaced, and the integration of the two armed forces into one integrated national army.
The central problem lay elsewhere. Given the starting point of an authoritarian regime, it was evident that any progress in the talks would require significant concessions from the government. The existing power-holders had already tasted loss when reluctantly agreeing to a coalition government. This was nothing, however, in comparison with the radical changes spelled out in the full Arusha accords. In effect, the Accords represented a veritable coup d'état for the RPF and the internal opposition, as a close observer later noted (Reyntjens 1994).
The Accords moved Rwanda from a presidential to a parliamentary system of politics. Most of the power of the Presidency would be transferred to a Council of Ministers. Habyarimana - who was to occupy the Presidency during the transition period prior to elections - was left with a ceremonial position. During the transition period, seats in the Council of Ministers of a Broad-based Transitional Government (BBTG) would be divided between the government and the internal opposition parties as well as the RPF. The ruling party (MRND) was left with only 5 out of 19 portfolios, and most were politically insignificant. The regime also would lose power in local administrative structures and the judicial branch. The distribution of portfolios agreed to at Arusha accentuated the dramatic nature of the change. Ideologically and politically, the BBTG represented a frontal attack on the power base erected by the Habyarimana government during twenty years of ruling the country _- a denial of authoritarian rule, of "Hutu power", and especially north-western-based Hutu power, which was the regional constituency and political backbone of the regime. The terms of the Agreement signalled a pluralist state and a civil society, and the idea of a people united by a common Rwandese nationality rather than divided along narrow lines into Hutu versus Tutsi.[26]
Significantly, the hardliners who surrounded Habyarimana - senior ministers in the MRND party, senior military figures in the armed forces and the Presidential Guard, and members of the newly formed Coalition pour la défence de la république (CDR) - were denied power in the transitional institutions. In spite of advice to the RPF from Western states that it would be a better tactic to coopt the extremists than exclude them, the CDR was not included in the power-sharing formula of the BBTG, nor allocated seats in the transitional Assembly. Habyarimana had demanded that the CDR be represented; the RPF, correctly noting that the CDR was both extreme and a non-party, was absolutely opposed and prevailed. The extremists responded by openly threatening to bring about "an apocalypse", as their spokesman, Col. Bagosora, declared during a session at Arusha when the power-sharing protocol was negotiated.
The losses incurred by the existing power-holders in the political sphere were paralleled, and thus sharpened, by equivalent losses in military matters. Hoping to secure at least continued control over the armed forces, the government opened the bid in negotiations on integration of the armed forces by offering the RPF a mere 20% share of the joint army. The Front rejected this outright, and the government was eventually forced to accept a much modified scheme that split positions in the joint command 50-50, not only at the high command level but all the way down to field commanders, and narrowly favored the government in a 60-40 distribution of troops. Observing the trend, a Western diplomat cabled home that the agreement had moved negotiations ahead of the consensus and risked repudiation by forces in Kigali. He was to be proved correct. Related military issues were also hard to accept for the army. Over 20,000 government troops would be demobilized, more than twice the figure for the RPF.[27] While the government soldiers had demonstrated little enthusiasm for fighting, especially during the 1993-RPF offensive, demobilization seemed even less attractive.
That the Arusha agreement produced such profound changes - which were registered as defeats by the government and especially by hardliners in Kigali - can be attributed largely to the balance of power between the Rwandese parties. The RPF negotiating team was extraordinarily strong, with discipline, preparedness and commitment in stark contrast to the fractionalized, ill-disciplined, and ineffective government team. Reflecting the divisions at home, the Kigali delegation was composed of some Habyarimana allies, but led by opposition party members, and obstructed by recognized members of extremist factions, which also were at Arusha. Equally, if not more important, negotiations on the military issues were conducted at a time when the military position of the government deteriorated sharply.
In February 1993, the RPF had launched a major offensive to break Habyarimana's opposition to the power-sharing formula of the BBTG protocol, and to protest the massacre of 300 Tutsi in Bugogwe. The offensive shocked Kigali and threw the government forces into disarray. Rebel troops fought to within 23 km of the capital, despite the fact that France sent a small contingent of paratroopers to reinforce government troops. The Front's victories clearly showed their military superiority; anticipation of success on the battlefiled was possibly a factor in the RPF's decision to undertake the offensive on the eve of the scheduled talks on military matters at Arusha.
The Arusha process has been commonly referred to as preventive diplomacy and judged against the catastrophic violence that followed. Held against this standard, it is clear that the process failed. But the main objective at Arusha was not to prevent a clearly perceived future calamity.[28] More conventionally, the aim was to end a war and construct a post-war peace agreement that reflected a situation short of total victory or absolute defeat. This required a formula for immediate power-sharing as well as agreement on a broad range of issues: the nature of the future political process, rules for demobilization and integration of the respective armies, and a system to facilitte the return of refugees and internally displaced persons. Similar agreements had recently been negotiated to end wars in Central America (El Salvador and Nicaragua), in Southern Africa (Namibia and Mozambique) and in South-east Asia (Cambodia). These, if any, were the models. The closer the Arusha process came to end the war and to define the protocols that would structure future politics, the closer - most participants believed - they had come to achieve peace. In this calculus, the primary criterion for success was to obtain the signatures of both parties on a peace agreement.
The final agreement dealt comprehensively with all the issues related to ending the war and opening up the political system. As such, it was a "perfect" agreement, as some observers claimed at the time. Yet the accords evidently did not rest on a new-found political consensus created by defeat, exhaustion or the emergence of new political forces. At the time, some Arusha observers suspected that the RPF was using the accords merely as a convenient way-station while preparing for a coup. As for the government, Habyarimana had repeatedly announced his unwillingness to sign, and disinclination to uphold, this "piece of paper", as he disparagingly referred to the accords in November 1992. Instead, the accords formalized what André Guichaoua has called une paix militaire.[29] The Accords were a victor's agreement. The RPF had extracted major concessions at the negotiation table by means of military strength. However, the Front had demonstrated its potential military superiority rather than inflicting an actual defeat on the government forces. The latter fell back but were not beaten. The coercive apparatus of the regime not only remained intact, but - in the logic of an armed peace - was strengthened. The difficult side-negotations on the modalities of providing security for the RPF members in the transitional government further underlined the point.
In this situation, one aspect of the peace agreement appears as a critical weakness. The Arusha process did not deal with the losers to the agreement. The extremists were not co-opted into the political process, nor neutralized by other means. They remained in a position to wreck the entire edifice.
The exclusion of the extremist CDR from the transitional government has become a major issue in subsequent evaluations of the strength and weaknesses of the Arusha Accords. Clearly, the mantra of the Namibia negotiators in 1988 - "in this negotiation, there will be no losers" - was not invoked at Arusha. Whether the alternative South Africa-style solution would have worked in the end remains conjectural, and opinions were divided at the time. Most observers at Arusha concluded it was unwise to exclude a major force in Rwandese politics, as this faction would have no stake in the successful implementation of the Accords. Other diplomats in Kigali shared the RPF assessment that the extremists were not interested in sharing power but bent on destroying the new political order that was drawn up. In this view, a speedy implementation of the agreement seemed the best means to control its violent opponents.
The party with most immediate leverage on the RPF was Uganda's President Museveni. Apparently he tried but failed to sway his erstwhile allies on both the CDR issue and the military protocol. During the course of the civil war, Museveni evidently had lost some of his influence over the RPF. The Front's repeated successes on the battlefield gave it a stronger basis for autonomous action. A second factor was Museveni's decision to distance himself formally from any involvement in the civil war, and, in particular, deny that arms were crossing the border in violation of international norms. During the Front's February 1992 offensive, Museveni wrote to the Security Council inviting the UN to deploy observers along the Uganda-Rwanda border to verify that Uganda did not aid the offensive [30]. This request, which resulted in the establishment of the UN Observer Mission Uganda-Rwanda (UNOMUR)(Res.846(1993)), was seen as a betrayal by some factions in the RPF.
2.1.5. The UN and the OAU: structural disjuncture
Since early 1993, OAU Secretary-General Salim A. Salim had pursued the idea of strenghtening the small group of military observers fielded by African states in Rwanda through the OAU (NMOG). The RPF offensive and resultant cease-fire in March 1993 had expanded the buffer zone to be monitored. The OAU Secretariat felt that a rapid response was necessary to stabilize the cease-fire and pave the way for a larger international presence as the prospects for a full peace agreement drew near(Tekle 1995). The present NMOG of only 50 men could not effectively monitor even the smaller buffer zone, but OAU was unable to raise funds for a larger force.[31] Turning to the UN for assistance in March 1993, Salim twice pursued the matter in letters to the UN Secretary-General, personally presented the case to Security Council members and presented a detailed proposal to a special envoy of Boutros-Ghali who visited Addis Ababa. His goals were modest: the maximum option was for an OAU force of 500 men at an estimated cost of US$2.5 million per month. After preliminary discussions with the Security Council, the response of Boutros-Ghali was that the request for UN logistics and financing "could only be entertained if the operations were under UN command and control."[32] This went against the OAU's wishes to develop a field presence so as to complement its mediatory efforts in Arusha, and demonstrate that African states indeed took an active part in solving African problems.
Finances were not the only issue. The larger question concerned the role of the respective organizations in an eventual peacee-keeping force in Rwanda. Since late 1992 the French Ambassador to the UN had started to lobby the Security Council for a post-war UN military presence in Rwanda. In early 1993 - and especially when the RPF offensive in Feburary pushed Rwandese government forces back - French efforts had become so insistent that it was "a standing joke in the Council", according to one ambassador present. From a French perspective, such a force would at best check the RPF's advance and provide a breathing space for the government both by positioning itself between the two belligerents and by monitoring the border between Uganda and Rwanda. At the same time, a UN force would be an answer to increasingly insistent demands from the RPF that French forces withdraw, as had been stipulated already in the 1991 N'Sele agreement. With a UN force in Rwanda and a French veto on UN peacekeeping operations in the Security Council, Paris could still exercise some influence after its own troops were gone.
The government of Rwanda, which had been elected to the Security Council, was generally ineffective, but on this issue vigorously supported France. The OAU was, with some reason, regarded as partisan towards the RPF and at any rate unable effectively to monitor or maintain the military status quo. By the same logic, and because France was promoting a UN force, the RPF wanted an OAU command.
Travelling to Dar-es-Salaam for a summit meeting on 5-7 March 1993, Salim nevertheless obtained approval from both the Rwandese parties to move ahead with an expanded OAU monitoring force. Two days later, the initiative was effectively upstaged in the UN Security Council. France introduced a resolution proposing a UN peacekeeping force for Rwanda, operating "in conjuction with" the OAU. (S/25400(1993)). Non-aligned members of the Security Council and some European states cautioned that French efforts might be viewed as an attempt by Paris to salvage its influence in Rwanda and warned against sidelining the OAU. While the languge of the resolution subsequently was softened, the signals given to OAU became clearer as every effort by Salim to increase even the monitoring function of NMOG slightly, was rebuffed.33 Boutros-Ghali sided with his own organization, despite formal deference to the importance of regional initiatives in conflict resolution. Apart from insisting on UN command and control, he added that any discussion of UN assistance would have to await the outcome of the Arusha talks. In the meantime, plans for a UN peacekeeping force went forward.[34]
In face of these discouraging messages, OAU abandoned efforts to play a significant role in the monitoring and implementation of the Arusha peace agreement. The development raises important questions concerning the appropriate working relationship between the UN and the OAU and, in this case, its impact on subsequent events in Rwanda. OAU admittedly had limited experience in peacekeeping and had problems fielding even a small military observer group. It was wholly dependent upon the UN to provide logistics and finance. Yet it was effectively discouraged by the UN and key Security Council members from taking a larger role. This is particularly ironic given the demonstrted ineffiency of the UN to deploy and develop the peacekeeping operation it had so aggresively sought (see chapter 3).[35] When the crisis later erupted, on 6 April 1994, it was the European battalion of UNAMIR that withdrew and crippled the future options of the force; the African battalion stayed. The political instinct among the veto-wielding members of the Security Council was to abandon the mission. Prima facie, there is no reason to think that the OAU would have performed worse.
Figure I: A pattern of violence
2.2. Track II: civil violence and human rights conditionality
The extremists in Kigali had a different vision of the peace process. As the Arusha process laid the groundwork for a virtual political revolution in Rwanda - the framework of a "new order", as it was said in Arusha - the Hutu extremists planned their own bloody, reactionary vision of change. In retrospect, that much has become clear. But at the time, what did donor countries and the international community more generally know about the extremists?
2.2.1. Early warnings
The political scene in Kigali in 1992-93 was admittedly complex, and a flurry of rumors and propaganda further complicated the picture. Key questions remain unclear even today, including the relationship between Habyarimana and the extremists. Was he a master or a victim of the hardliners, and if he moved from one role to another, when did this happen? Outside analysts also had to decipher the meaning of the new civil society that developed. The print media multiplied across the political spectrum. So did NGOs and various political movements. The difficulty of distinguishing rumours from facts and "hard" infomation from rumours was exacerbated by diffusion of propaganda that all sides resorted to as a weapon of war and political competition in this period.
Yet there was considerable and growing evidence of Hutu extremism. Human rights violations were documented by human rights organizations and concerned activists. The French had a military mission in Kigali, and the small foreign diplomatic corps observed the scene, reported home, and, for the most part, shared information with each other. An informal grouping of ambassadors from the United States, Belgium, Germany, France and the Vatican - also known as The Five Musketeers - met frequently.
It was well known at the time that the combined peace and democratization process faced strong opposition. Already in July 1991, a French intelligence assessment identified three critical circles of power in relation to the peace process (not to be confused with the three political tendencies discussed above), starting with the inner circle formed by the President's in-laws and associates, an outer circle of relatively more moderate members of the military and the cabinet, and, lastly, an amorphous grouping of younger officers and intellectuals willing to entertain change. Since the objective of the first circle was to retain power, any form of imposed negotiation would provoke their resistance. The analysis found its way to other donor states and was soon amplified by pointed statements and actions by the hardliners.
By late 1992 and especially in early 1993, a much more comprehensive body of evidence was available and repeatedly discussed in the diplomatic community in Kigali. Killings of Tutsi were more frequent, and there was increasing information about death squads as well as the existence of a Network Zero around the President, which reportedly was plotting to exterminate regime opponents and circulated death lists. The report of the International Commission of Inquiry on Human Rights in early 1993 was widely circulated (FIDH 1993). The Belgian government, for one, took it so seriously that it temporarily recalled its Ambassador from Kigali. The report documented the involvement of the government in what was described as systematic killings directed against the Tutsi, and estimated that about 2,000 persons had been murdered in the 1990-92 period. Briefing reporters on its finding, the Commission used the term "genocide" to describe the killings. This claim was retracted in their official report published in March because of fear that such a designation would be considered hysterical.[36] It was, however, upheld in a report by the UN Commission on Human Rights a few months later (see below).
The diffusion of hate propaganda in Rwanda was also noted as an indicator of increasing tension and mounting "Hutu power". However, interpretations of the message differed. As with the notorious Radio et télévision libre milles collines (RTLM) - founded soon after the signing of the peace agreement - the hate media in this period were so explicit and literal that some diplomats, both Western and African, afterwards said they tended to dismiss it.[37] Similar reactions were entertained even by the RPF, as a leading Front member later recalled: "What they said was so stupid; we did not take it seriously enough."
There were other telling pieces of evidence. At least two revealing government documents were leaked to the diplomatic community and circulated in Kigali. One was an internal report from a senior military officers' commission entitled "Definition and Identification of the Enemy" (21 September 1992). The report listed as enemies not only those Tutsi inside and outside Rwanda who supported the RPF, but also members of mixed marriages and moderate Hutu who opposed the hardliners within the government. The other was a letter from the (opposition member) Prime Minister to the Defense Minister, dated 25 March 1993. The Prime Minister gave details of illegal distribution of weapons to civilians and called for immediate action to stop it.
By this time, links between the civil violence and the peace process became more explicit. At every important juncture and setback for the government, Tutsi were murdered. Clearly discernable in retrospect (see Figure I: A Pattern of Violence), the coincidence between killings of Tutsi and the rhythm of negotiations in Arusha was also recognized at the time by several Kigali-based diplomats and variously interpreted. "We read the killings as a political negotiating tactic", one analyst later said. Others saw the violence more radically as an attempt to derail the entire peace process.
Simultaneously, diplomats in Kigali noticed that the extremist opposition to the Arusha talks was gaining strength. During the talks on power-sharing, Habyarimana was oscillating between moderates in his own delegation and the hardliners at home. The talks were halted in September, but when they resumed the Foreign Minister, Boniface Ngulinzira, was negotiating without the authority of the President, as an observer delegation noted at the time. Habyarimana soon made the point publicly by calling the accords merely "a piece of paper" in his famous 15 November speech. At Arusha, as noted, the CDR representative predicted "an apocalypse" when the power-sharing protocol was finalized. The message carried added significance since it was delivered by Col.Bagosora, the person widely believed to be behind the work of the death squads, including the Kibuye massacres in August 1992.
Signs clearly pointed to the consolidation of the power of the Hutu supremacists - through the formation of what amounted to para-statal organizations. The CDR was formed in March 1992, bringing together military and political extremists from circles inside and connected to the Presidential Palace. The party militias (the interahamwe and impuzamugambi) were also formed at this time with direct support from the CDR, the ruling government party, and the Presidential Guard. Both operated openly, and diplomats in Kigali readily recognized both as instruments of the ruling party. Yet their precise function was open to various interpretations. Some observers accepted at face value that the militias were formed to defend the country against the invading RPF troops, according to the doctrine of village self-defense. Subsequent intelligence information suggested that up until possibly late 1993, this was indeed their main purpose. In a cable from Kigali to New York dated 11 January 1994 - later to become known as The Black File - the UN force commander in Rwanda relayed a statement from a high-level government informant as saying that the "principal aim of interahamwe in the past was to protect Kigali from RPF". With a peace agreement in place and the arrival of a UN peacekeeping force to help implement it, the militias started planning for the extermination of Tutsi (see chapter 3).
There was a certain confusion about who constituted the principal target. Some diplomats, including the Canadian Ambassador in 1994, read the violence primarily as politically directed against regime opponents regardless of ethnicity, and were totally unprepared for the genocidal dimension of the violence.[38] Others saw Rwanda's Tutsi minority as a hostage to the war between the Habyarimana-regime and the RPF, and thus collectively being at great risk. This was the pattern of the early 1960s, when attacks by a previous generation of Uganda-based Tutsi refugees-become-guerillas brought fierce retaliation against the Tutsi who remained in Rwanda.
2.2.2. The conditionality question
Acknowledging the widespread evidence of a deteriorating situation, donor countries made some efforts to correct it. While the United States did not stress, and France did not include, human rights issues in their policy towards Rwanda,[39] other donors did. Belgium formally incorporated human rights criteria in its aid policy after the Socialists came to power in 1992. Canada, another major donor, had done so all along. Also the representatives of Switzerland and the Vatican spent considerable time and efforts raising human rights issues, and made repeated representations to the President and the Minister of Justice (when there was one). More generally, diplomats from states concerned with democratization put their faith in political reforms: promoting multipartisme and good governance, it was argued, would help address human rights abuses, including the critical issue of legal impunity.
There was some progress towards democratization. The regime's progressive move towards multi-partisme was partly due to Western persuasion and pressure. High-profile pressure by foreign diplomats and quiet work by ICRC helped to bring the release of nearly 10,000 persons who were detained by the government when the RPF invaded in 1990. A few individual human rights cases were addressed.
In principle, most Western donors made economic aid conditional upon observance of human rights. In practice, however, most donors did not reduce aid with specific reference to human rights violations, even though these were recognized as severe. For instance, the Belgian government took steps considered extreme in the language of diplomacy when the international human rights commission published its incriminating report in March 1993. The Belgian Ambassador in Kigali was "recalled for consultations", and the Rwandese Ambassador in Brussels was told that Belgium would reconsider its economic and military aid unless steps were taken to rectify the situation. However, Habyarimana made conciliatory statements and Belgian aid continued. The Canadian government announced in Kigali that, largely for human rights reasons, but also for budgetary reasons, Rwanda would be phased out as a major aid recipient from 1992-93 onwards. In Ottawa the decision was explained largely with reference to budget cuts and changes in bilateral programming.[40]
Among the African countries, only the Ambassador of Tanzania took a strong, outspoken position on human rights violations. As a result - and in consequence of Tanzania's weak economic leverage - Habyarimana asked in late 1993 that the Ambassador be replaced with another one. Tanzania complied. The OAU's Commission on Human and People's Rights visited Kigali once and generally "performed abysmally", an OAU observer concluded.
Why was human rights conditionality - an arrow in the quiver available to the international community - not pursued more decisively? Conditionality on economic policy was all along precisely defined and consistently promoted. The principal demands for structural adjustment and fiscal reform government to comply in the area of economic reform possibly reduced the incentives for donors to insist on human rights issues. There were also other concerns. With a traditionally efficient local administration and functioning infrastructure, Rwanda had a capacity for absorbing aid that partly accounted for large transfers in previous years. In the early 1990s, it was still a country where foreign aid bureaucracies could reasonably disburse money. Project feasibility helps to explain the continued inflow from two traditional donors, France and Belgium.[41]
There was a more fundamental reason. International pressure to democratize and institute good governance developed a special rationale in Rwanda: just as the Arusha process was a solution to civil war, so democratization came to be seen as a solution to the growing problem of civil violence. Support for democratization and the related peace process implied continuous economic and diplomatic engagement in Rwanda. From this perspective, the threat of ultimately imposing sanctions by withdrawing aid - as Western human rights organizations called for in 1992-93 - was counter-productive.[42] Donors thus became hostage to their own policies.
As the economic situation deteriorated in 1993 and renewed war plus drought increased the number of internally displaced people to around one million, the international community pledged large amounts of humanitarian aid. Simultaneously, some donors shifted development aid into relief assistance. The issue of conditionality thereby became more remote. The call for pledges of international humanitarian assistance went out in the UN system in March, just at the time when the International NGO Commission published its report detailing the government's involvement in massive human rights abuses and calling for human rights conditionality in aid policy. The UN Department of Humanitarian Affairs put the question on the table by asking in an internal memorandum how the human rights issue should be handled at the forthcoming consolidated appeal for humanitarian assistance. The answer suggested by the DHA was to distribute copies of the report before the meeting.[43]
Given the problematic nature of sanctions, donors instead looked to "positive conditionality", in European Union terms, to promote democratization and human rights, and later to have the Arusha Accords implemented. Aid was allocated to strengthen particular sectors. Thus Canada, the United States and Belgium supported projects to improve administration in the Ministry of Justice with advisors, seminars and training, and to strengthen civil society by supporting local human rights organizations and encouraging a free press.
Despite these efforts, the quality of governance declined markedly. In the Ministry of Justice, for instance, the assignment of foreign advisors could not make up for the fact that for most of 1993 there simply was no Minister of Justice. The one who was finally appointed towards the end of the year had a shady reputation which was subsequently confirmed.[44] Under the impact of multipartisme and fiercely conflictual politics in late 1993 over the implementation of the Arusha agreement, administrative agencies turned into political fiefdoms of a more extreme kind than before. There was no accountability and project implementation had become impossible, USAID concluded, suspending its program in 1994. Germany did the same, citing insecurity in the countryside and the erosion of administrative efficiency. The European Union also put most of its 4-year aid allocation to Rwanda on hold.
By suspending aid with reference to bookkeeping and resource restraints rather than human rights criteria, donors sent the message that human rights conditionality was preached but not practiced. The same applied to the infamous Radio mille collines. Unable to agree among themselves on the significance of the radio and how to silence it, the donors limited themselves to making yet another representation to Habyarimana. The latter responded as usual by promising to look into it, and there the matter rested.[45] Half-hearted conditionality of this kind probably succeeded only in eroding the credibility of diplomatic suasion.
By late 1993, the limits of diplomacy were becoming obvious, yet the prevailing sense among concerned members of the diplomatic corps was that the only alternative was to soldier on. Arguably, a window of opportunity for dealing with the extremists had existed around 1992, before their strength developed in earnest and while the donors still had commitments to development aid that could be adjusted as needed. Although information about the extremists at that time was less alarming, human rights violations were significant indicators of trouble, as well as important in and of themselves.
2.2.3. The UN human rights machinery
In this period, the UN was peripherally involved in the conflict management process in Rwanda, while humanitarian agencies, as noted, provided significant aid. The UN also had a formal mandate to monitor and pronounce upon human rights violations in member states. This it did in Rwanda, but in a very limited fashion compared to the vigorous and thorough monitoring carried out by NGOs.
The UN Commission on Human Rights discussed Rwanda in both 1992 and 1993 under its "1503 procedure", the confidentiality clause used to initiate discussion on countries where serious problems are noticed.[46] Customarily, the 1503 procedure is a warning signal that alerts the Commission to the fact that human rights conditions within a country merit particular attention.
As the discussion was protected by the confidentiality clause, little is known of what transpired when the Commission's Committee of Five examined Rwanda in 1992, and again in 1993. The first year, only a pro-government spokesman appeared, a local judge known to Western diplomats for his grossly inadequate handling of cases brought against government officials. By the time of the second meeting, Rwanda had a coalition government, and a member of the opposition MDR party was in the delegation, the then Minister of Primary and Secondary Education, Agathe Uwilingiyamana. Mrs. Uwilingiyamana - who was the first high-level official to be murdered after 6 April - admitted to human rights problems and said the government would try to rectify the situation.[47]
Rwanda was also discussed in several thematic reports presented to the Commission as a whole in 1992 and 1993.48 In early 1993, human rights groups in Rwanda asked the Special Rapporteur on extrajudicial executions, Mr. B.W. Ndiaye, to participate in an international commission of inquiry into human rights violations in the country. While declining on procedural grounds, Ndiaye was sufficiently concerned to visit Rwanda. His 10-day visit produced a report in August 1993 (E/CN.4.1994/7), which was presented to the Commission early in 1994, i.e. shortly before the genocide took place. The report was given only routine consideration in the Commission. As a political body composed of UN members, the Commission rarely discusses particular countries except those singled out for attention by a Special Rapporteur. Not being singled out, Rwanda was folded into a report the Commission discussed thematically rather than according to country cases.
Because the Ndiaye report raised the question of genocide, the failure of the Human Rights Commission to give it more than routine recognition has later been criticized. One explanation lies in the nature of the report. Ndiaye supported the substance of the allegations of the NGO Commission (FIDH 1993), in particular concerning official involvement in the massacres of civilians (para. 28). He further noted that the question of whether the massacres may be termed genocide "has often been raised" (para.78). After a brief discussion - contained in 4 out of the report's 86 paragraphs - Ndiaye offered an "initial reply": since Tutsi have been the victims in the overwhelming majority of cases, and have been targeted solely qua Tutsi, the Genocide Convention's Art.II (a) and (b) would apply.
A Western diplomat in Kigali who read the report, and himself was nearly killed by the Hutu extremists, said it hardly prepared him mentally for the massive slaughter that was to come. Rather, the Ndiaye report was read as a technical, legal analysis, not an alert that raised the specter of the killings of half a million or more people. The magnitude of deaths associated with genocide in popular opinion was absent.[49] Moreover, the report presented the unfolding violence in the context of "business-as-usual": "...the country has already experienced many massacres of an ethnic character. Such acts of violence recur periodically, and the persons responsible, who in most cases are known to everyone, go unpunished. The Special Rapporteur became aware ... that this situation was part and parcel of everyday life for many Rwandese people, and that ethnic violence had become a practice which, if not accepted, was at least firmly rooted in Rwandese folk memory." (para 46)
The failure by the international community to make effective demands on its relationship with a regime whose involvement in organized killings of civilians was generally suspected and carefully documented in two 1993 reports - one from a joint NGO mission and the other by a Special Rapporteur of the UN Commission on Human Rights - was an act of omission that carries at least some moral co-responsibility for subsequent events in Rwanda. Acts of commission carry heavier responsibility.
French aid was justified with reference to the continuing war with the RPF. The nature of the conflict meant, however, that France helped to strengthen the military apparatus of a regime that was deeply implicated in serious human rights violations, and some of whose members subsequently became deeply implicated in the genocide as well. French military aid also strengthened the latter more directly in that some weapons provided to the FAR were diverted to the militias. Milita members were believed to be among the army recruits French coopérants militaires were training (Prunier 1995). Diplomats in Kigali observed French officers with interahamwe units in a national park area.[50] On the eve of the civil war, and even a couple of days after the killings had started in April 1994, French planes reportedly flew in arms to the Rwanda government (see 3.2.).
Information about the human rights situation was generally available; given its close relations with Rwanda, France had better access than most to obtain intelligence on the extremists. Yet it is striking that senior officials in the Ministry of Cooperation in 1993 regarded Rwanda as "the Switzerland of Africa", as one official later said. To the extent that human rights violations were noted, they were considered as not particularly bad "by African standards". Evidently, the Office of Military Assistance (DAMI), which is located in the Ministry of Cooperation and reports to it, either did not observe the mounting force of Hutu extremism - which is possible given its preoccupation with monitoring the designated enemy, the RPF - or observed but did not report. Either way, for the ministry most directly and broadly involved in Rwanda, it amounted to a massive intelligence failure.
French policy in Rwanda was hardly the result of a grand design. More likely, it evolved as the outcome of multifarious and partly uncoordinated initiatives, each of which was shaped by the prevailing sense of France's purpose and role in Africa. The result was to give encouragement to forces that were essentially evil, yet, the process itself demonstrated in a sense "the banality of evil" (Callamard, 1995). The absence of a grand design does not, however, obviate responsibility for the consequences. The French have been their own most severe critics in articulating the political and ethical implications of France's Rwanda policy (e.g. Bayart 1994, Brauman 1994, Verschave 1994). Even from the criteria of Realpolitik it seems questionable. As the noted scholar Jean-François Bayart asked, if French policy aimed to stabilize the situation, how could that be achieved by supporting a regime that permitted, or even organized, assassinations and mass murder?