Journal of Humanitarian Assistance
5.1. Changing the mission: UNAMIR II
Not until the end of April was there some acknowledgement from the United Nations that it ought to respond more forcefully to the Rwanda conflict. The initiative did not come from the Security Council, where some members had registered strong disapproval of the proposal to scale down UNAMIR, but nevertheless had voted for it on 21 April. Rather, it was the Secretary-General who turned around and already on 29 April called on the Security Council to take "forceful action" in Rwanda. The move marked the beginning of a much more proactive role by the Secretariat towards the crisis. Recovering from its near-paralysis in the immediate aftermath of 6 April, the Secretariat assumed the role of a model executive branch: it asserted the moral and political need for action, assessed the problem, formulated an innovative and bold response, and actively sought to mobilize resources for its realization. This change in itself requires explanation.
As information about the situation in Rwanda and pressures on the UN to act were accumulating, the Secretary-General took a high-profile role to re-engage the UN militarily. The media, as noted, had scattered, inadequate coverage during most of April. Media pressure, therefore, was an unlikely factor in the reversal. The pressure from the NGOs was mounting. The ICRC had launched several appeals, although the principal humanitarian NGOs involved in Rwanda did not identify the massacres as a genocide until early May.[95] Probably more important, the Secretary-General was under increasing pressure from African countries to demonstrate that the United Nations took its African constituency seriously and did not give preference to European crises. African members of the UN had called for "forceful action" during the discussion of UNAMIR's fate after 6 April, and maintained this position as the months went by and the death toll mounted. Their demands were buttressed by growing information being fed to the Secretariat by UNAMIR and DHA about the nature and magnitude of the massacres. The newly established Operations Centre in DPKO/New York relayed graphic information about the slaughter. On 23 April, the Under Secretary-General for Humanitarian Affairs, Peter Hansen, travelled to Kigali, becoming the first New York-based high UN official to report back on the situation.
Elsewhere in the UN system, there was little impetus to act. The key human rights mechanisms had not been visible in early April, nor did they figure in this reversal of policy. The new position of High Commissioner for Human Rights was filled only one month before 6 April, and the new Commissioner did not physically take office until the day before; it is understandable that he did not immediately play an active role. A week after the Secretary-General had put Rwanda back on the Security Council's agenda, the High Commissioner, José Ayala Lasso, called on the UN Commission on Human Rights to convene an emergency session to consider the killings.[96] More striking was the passivity of the Commission itself. Only on 25 May - that is, several weeks after genocidal killings had commenced, one week after the Security Council had agreed to reverse course and send in a military force to protect civilians, and after an urgent appeal by the new High Commissioner for Human rights - did the Commission act. A Special Rapporteur was appointed to report on the human rights situation in Rwanda, and on 28 June he confirmed in a detailed assessment that genocide and other human rights violations had occurred (E/CN.4/1995/7). By this time, the genocide had claimed most of its victims. The OAU's African Commission on Human and People's Rights was completely inactive.
By taking a high-profile role in calling for a UN reversal on Rwanda, Boutros-Ghali put the onus of rejection or compliance on the Security Council. The immediate response from the Security Council on 30 April was cautiously supportive in principle, and even contained an indirect reference to the possibility that genocide might be unfolding in Rwanda (S/PRST/1994/21). It was evident, however, that the Permanent Five did not wish to engage directly. The Council passed the initiative to the African states by asking the Secretary-General to consult with the OAU and report back on further action. OAU's Secretary-General responded positively, and on 10 May provided firm offers of African troop contributions.[97] In the meantime, the Secretariat was developing a plan to expand UNAMIR and change its mandate to help protect civilians. The outlines of the innovative plan was ready on 13 May and was authorized by the Security Council on 17 May (Res.918 (1994)). In a surprise move just before the vote, however, the US delegation asked for more detailed preparations from the Secretariat before agreeing to the force. The final approval for establishing UNAMIR II came in the form of an implementing resolution on 8 June (Res. 925(1994)) - i.e. almost two months after the massacres had started and over a month after the Secretariat and the Security Council had started to reverse course.
The UN maintained a slow pace even though time was of the essence. The Security Council had authoritative information to the effect that genocidal violence was unfolding at a rapid rate, with possibly thousands being killed every day. Already on 30 April, the Council itself acknowledged that "massacres and wanton killings...in a systematic manner" were occurring; these were further linked to genocide although in carefully non-committal terms.[98] The point was reiterated in mid-May. The Secretary-General used progressively stronger language to describe the violence in his communications to the Security Council, going from "massacres of innocent civilians ...on a massive scale" (29 April) to unequivocally declaring "genocide" (May 31). Reporting on a special mission he sent to assess the Rwandese situation during the second half of May, Boutros-Ghali used the figure of 250-500,000 killed (S/1994/640, May 31). In Washington, US administration officials had told the Congress in early May that an estimated 100-500,000 persons had been killed, although they considered the latter figure to be on the high side. Despite this information and the acknowledged severity of the situation, the Security Council itself did not push for action, and the United States slowed down efforts by the Secretariat to move more quickly by insisting on additional procedures before authorizing implementation.
Why was the UN so slow to act? Were the causes "systemic" characteristics of the UN, or the result of particular actions (or inactions) by individual actors?
While slow relative to the needs on the ground, the process was not so tardy by standards of multilateral diplomacy. The Secretariat had to identify troops, secure finances, and prepare a plan. When objecting to immediate authoriztion on 17 May, the United States delegation claimed that Boutros-Ghali's plan was inadequately prepared and lacked field assessments[99]; some regarded it as more of a public relations effort than a strategy for a serious mission. Yet neither the US nor any other of the Permanent Five that had the capacity to energize the process did so, reflecting a more fundamental, collective concern not to become involved in another UN operation in Rwanda.
The question of obligation to act was treated cautiously by Security Council members. With fast-moving violence on the ground, the UN's earlier commitment to implement the Arusha Accords had become irrelevant, at least in the short run. The genocidal violence against civilians did, however, raise the question of an obligaton to act under the Genocide Covention. Key members of the Security Council answered in the negative. The United States carefully sought to limit the obligations by terminological adjustments.[100] The British government "did not accept the term" genocide either, and the British Foreign Office was inclined to see a discussion of whether or not the massacres constituted a genocide as " sterile".[101] The French Foreign Minister, Alain Juppé, responded to the growing outrage in France over the killings by declaring it a genocide on 16 May, making France the only one of the Permanent Five to do so.[102] However, France did not translate this into a policy of expediting the deployment of UNAMIR II. The subsequent French unilateral intervention was launched as a humanitarian action that made made no reference to the Genocide Convention (see 5.3 below).
By omission and commission, the role of the United States was critical. As the world's last remaining superpower, the United States had repeatedly demonstrated that it could move the United Nations to take rapid and effective action. In the Gulf war (1990-91), the Administration had done this both by committing a significant part of its own military force, and by mobilizing other states through the US alliance structure and by exercising its formidable economic power. By not utilizing its capacity to act, and its formal leadership role in the United Nations, the US must take considerable responsibility for the overall failure of the UN to respond at this juncture. The same applies, though to a lesser degree, to the other major powers on the Security Council. France and Great Britain, in particular, had the capacity for rapid military action and considerable political resources to energize the Security Council. Even so, Britain could only muster 50 four-ton trucks at the height of the crisis. France demonstrated its capacity much later by undertaking a unilateral "humanitarian intervention", which the Security Council endorsed with utmost speed.
By not taking the initiative, the United States and others of the Permanent Five ensured that the Security Council proceeded along its customary and laborious path of consensual decision-making. More specifically, the explicit unwillingness of the major Western powers to commit troops to a revitalized UN force - which had been clear in the decision to scale down UNAMIR I and was reiterated in early May when UNAMIR II was planned - meant that the UN force had to be established on a mix-and-match basis. African countries would send troops, while industrialized states were to provide equipment and finances. To constitute a force along these lines was extremely time- consuming even under the best of circumstances.[103]
Problems of financing the force also caused delays and weakened the response. The role of the United States in this respect was particularly important since the US was automatically assessed 31 % of the costs of all UN peacekeeping, thereby acquiring more than a formal veto in the matter. US consent is essential to mount an operation whether or not American troops are involved. The United States made it quite clear during May that on purely financial grounds it would not readily endorse an expanded UNAMIR. US reluctance bogged down the UNAMIR discussion from the start. At one point, US demands for careful planning, prior troop commitments, and security of the mission - justified by the need to avoid wasting money on a failed mission - made the Security Council delay final authorization of the force for an additional two weeks, from 17 May to 8 June.
While the structure of UN assessment gives the US an inordinate power to influence UN peacekeeping, there are ways around this. In early May, the Secretary-General floated the idea of establishing a voluntary fund to finance an expanded UNAMIR. Having obtained troop commitments in principle from several African countries, Boutros-Ghali proposed an initial fund of US$50-80 million to launch the force. None of the economically powerful UN members came forward to subscribe, however, and the US received the idea coolly. By their inaction, other members of the Security Council ensured that the regular assessment structure would apply, thereby keeping the onus for action on the reluctant US.
US reluctance to support UNAMIR II reflected a progressively critical attitude of the Clinton Administration towards UN peacekeeping operations. With great deference to a hostile Congress, the Administration had just completed a review of UN peacekeeping operations that established stringent criteria for US support (PDD 25). To support UN peacekeeping in Rwanda, the Administration would have to determine that peace in this remote Central African country was critical to US national interests. In conventional strategic terms, this would be difficult, and the Administration did not try. Nor did it elaborate alternative criteria for national interest, e.g. to reduce massive international refugee flows or the principled need to react to genocide.[104] Congress had expressed concern over the escalating bill for UN peacekeeping when UNAMIR I was planned in September 1993. Since then, the operation in Somalia - deemed a disaster from a US perspective - had reinforced Congressional reluctance to finance UN peacekeeping anywhere, with or without US troops.
In reality, the budget for the projected UNAMIR II was quite small when judged against the "downstream costs" for relief and rehabilitation (estimated to US$1.4 billion in 1994), not to mention the human toll. The estimated budget for UNAMIR II was US$115 million for 6 months (or slightly under US$20 million a month).[105] This was the same rate as UNAMIR I (about US$10 million per month for a force of about half the size), and reflected attention to fiscal stringency despite the changed situation. Both this budget and an advance voluntary fund of US$50-80 million were well within the means of the industrialized states to finance.
The reluctance of the US and other members on the Security Council to embark on a military re-engagement reflected additional concern about the risks and practicality of intervention. On the ground, the Rwanda situation seemed confused and anarchic, raising the question of the mission's effectiveness, security of personnel, and fears that support operations might be needed for rescue or reinforcement. Again, behind these concerns lurked the Somalia experience, although the two situations in reality were quite different. There had been no systematic and massive killings of civilians in Somalia. The proactive orientation of UNOSOM to disarm armed factions and to arrest a prominent faction leader had been undertaken with a Chapter VII enforcement mandate. In Rwanda, by contrast, the proactive function would be to protect civilians. Although force could be used if necessary, a reduced-risk option was proposed whereby the units would move in from the border and remain in the protected zones in outlying areas.[106] On the other hand, there were obvious elements of risk. Throughout May, when the Security Council deliberated on the nature and mission of a military re-engagement in Rwanda, the members were fully aware that there was no cease-fire in the country or in Kigali; the airport was not secure; the Ugandan army had stopped all UNOMUR patrols on the Ugandan side of the border to keep arms flowing to the RPF; and it was not clear that the parties to the conflict would welcome, let alone respect, a UN force as a neutral, humanitarian entity.
The Rwandese parties were also responsible for the slow and hesitant UN deployment. During the critical Security Council deliberations in early May, the RPF sent ambiguous messages. While in principle accepting a humanitarian force, the Front feared that UNAMIR II at best would be irrelevant, and at worst a hindrance to their advance or protection for the government forces.[107] These concerns were consistent with their initial hostility to a UN force in early 1993 - which only changed to a lukewarm attitude prior to 6 April 1994 - and further strengthened the forces of caution in New York. On the government side, a UN force would prima facie be even less welcome insofar as its mission would be to protect civilians, who were targeted principally behind government lines.
5.2. Deploying UNAMIR II
UNAMIR was reauthorized at the level of slightly more than 5,000. The new mandate was to help protect civilians by establishing humanitarian zones inside Rwanda, and by facilitating humanitarian relief to displaced and other needy persons. While designed under a Chapter VI mandate, the operation envisaged clearly went beyond conventional peacekeeping; for example, the final resolution authorized the force to "take action in self defence" against those who "threaten protected sites and populations emphasis added" (Res.918/1994). In effect a "6.5 mandate", the formulation permitted a flexible and firm response to protect civilians. It was a significant formalization and a further development of the practice undertaken on the ground by the rump UNAMIR contingent in Kigali, where UNAMIR soldiers were posted to sites that had become sanctuaries for threatened civilians.
The mandate and design of UNAMIR II seemed appropriate to the situation. It is reasonable to assume that UNAMIR II, if quickly and effectively deployed, could have had a significant impact in terms of saving lives. Even the minimal efforts by the UN forces on the ground, it will be recalled, had some protective effect. The most impressively cost-effective operation was at the Amahoro Stadium in Kigali, where, after the Bangladeshis left, 12 UN "blue helmets" armed only with hand weapons and barbed wire protected several thousand persons. Since the civilians mostly were not threatened by organized, well-equipped army units (as in Bosnia), but mainly by para-military gangs and mobs, a relatively modest force could have a significant deterrent effect.
However, the intervention was delayed in two ways that rendered it irrelevant to the most critical phases of the conflict. Over a month passed before the Security Council authorized the operation. The actual deployment took another two months, with the result that the force arrived in Rwanda after the civil war and the genocide were over, and after a unilateral French intervention with a similar mandate had taken place.[108]
Despite the risks of the situation and the novelty of the mission, African countries responded rapidly with troop commitments. OAU again emerged as a focal point for mobilizing support.[109] By the first week of May, Nigeria and Ghana - both with large armies and peacekeeping experience - had responded positively, and Tanzania, Senegal, Zimbabwe and Zambia had indicated interest. In addition to regular financing, however, all needed various kinds of equipment and logistical support; even Nigerian and Ghanian units lacked vehicles and heavy equipment.
As a result, the mix-and-match pattern of African troops and Western equipment developed. This had been practiced in other UN peacekeeping operations but on a smaller scale and not for units that formed the backbone of a force. Before being deployed, the troops committed to UNAMIR II had to be trained and equipped within a multinational and often two-layered bureaucratic structure (UN and bilateral). Donors of equipment grumbled that African countries used peacekeeping operations to outfit their forces, and worried that their equipment would be misused. Combined with "normal" bureaucratic procedures, this produced long delays.
The infamous APC-case is illustrative: The United States was to lease (with full renumeration from the UN) 50 armored personnel carriers (APCs) to a second Ghanian battalion earmarked for UNAMIR. The Pentagon received a formal UN request for these APCs on 19 May, which started a lengthy process. Paperwork for the lease had to be completed before the vehicles could be assigned. Several practical issues went back and forth regarding the type of vehicles (wheeled or tracked?), repainting them (white), finding a training partner (the US requested Egypt but the UN declined due to "extraordinary preconditions.") It took a full week to transport the 50 vehicles from a US base in Germany to Entebbe (Uganda), and when the last APCs were finally unloaded at Entebbe (30 June), they remained on the tarmac under UN guards for a month before moving on to Rwanda. Not until 30 July did the first APCs roll into Rwanda. By this time, the end of the war and the installation of a new government in Kigali rendered their original mission of securing the airport and protecting civilians quite irrelevant.
The "systemic" and individual factors that caused the crucial delays in establishing UNAMIR II were the same that had produced a halting deployment of UNAMIR I. Requiring consensus from key members of the Security Council and being dependent on national sources for men and equipment, the UN needs 3-6 months to form a peacekeeping force of this size. A determined political push by member states, however, can cut through the bureaucratic tangles to speed up the decision-making process, or come up with missing force components from national sources. In this case, strategic disinterest in Rwanda, combined with fear of a failed or costly mission, made key members of the Security Council hold back. Since it is reasonable to assume that a more rapid deployment of UNAMIR II could have protected a large number of civilians, the consequences of delay are the equivalent loss of life, and, quite possibly, a larger outflow of refugees who otherwise might have remained in protected areas, as happened in the French humanitarian zone (see 5.2).
Critics have argued that the UN presence and activities in Rwanda throughout the crisis were marked by great concern for neutrality (African Rights, 1995). Given that events in Rwanda constituted genocide - as the Secretary-General called it on 31 May (S/1994/640), and as did the Special Rapporteur of the UN Commission on Human Rights a month later (E/CN.4/1995) - then neutrality is untenable on moral and legal grounds (Brauman, 1994). However, the more fundamental issue was the lack of commitment: the main obstacle to an effective response in Rwanda was not that the UN insisted on neutrality, but that it was not there.
5.3. Opération Turquoise
While the UN was struggling to mount troops and equipment for an expanded UN presence in Rwanda, on 19 June France suddenly offered to send its own force, but under French command and control. Officially designated as a humanitarian intervention to protect civilians and hinder a mass outflow of refugees,[110] Opération Turquoise was launched and executed in a manner suggesting mixed motives. Partly for that reasons, the consequences were both humanitarian and political in nature.
At the UN, the Secretary-General presented the French offer as a stop-gap measure until UNAMIR II was fully deployed, then expected to take another three months.[111] Critics have argued that France could readily have speeded up this process by placing its Turquoise-designated units at UNAMIR's disposal. However, strong RPF opposition to accepting a UN force with a French contingent made this difficult as long as UNAMIR operated under a Chapter VI mandate requiring at least minimal consent from the local parties. When France decided to intervene unilaterally, it was regardless of RPF permission and with a heavily armed force authorized to use "all means necessary" under Chapter VII of the UN. The RPF chose not to challenge the interventionist force.
By retaining national command, the French government was able to conduct and claim credit for an operation that on purely military grounds was generally judged as well planned and superbly executed under the command of General Lafourcade (Connaughton, 1994). The political credits were also "national"; the ably run Opération Turquoise came to be juxtaposed with the uncertain and fumbling UN peacekeeping operation. At the same time, the French had requested and quickly obtained a Security Council endorsement of the intervention (Res. 929(1994), 22 June). The operation consequently confirmed the growing precedent for unilateral interventions initiated by a state and conducted under national command, but legitimized by the UN.[112] The overall effect was to further undermine the institution of collective international intervention as delegated to the UN and envisaged in its Charter.
These wider consequences partly flowed from the judgement that the operation had been a success: As planned, the units moved rapidly and mostly avoided armed confrontations with the RPF or any of the other parties. They could claim with some reason that the intervention saved lives and stabilized the situation in the south-western corner of Rwanda, thereby reducing the outflow of refugees. The French presence facilitated relief work carried out by humanitarian organizations. Moreover, the force accomplished all this without incurring any casualties and withdrew as planned within the two-month time limit specified at the outset.
Each of these claims have some validity, although their significance must be judged in a wider context. Opération Turquoise did save some lives if it is assumed that these otherwise would have been killed before the RPF could reach the area. By 2 July, the French had established control over the south-western corner of Rwanda, equivalent to one-fifth of the national territory, and proclaimed it a "safe humanitarian zone". The RPF claims that its forces could have "liberated" this area within a short period of time, and that the Tutsi who survived for over two months to be rescued by the French would equally have been there later if RPF units had arrived. As noted, however, there was a dynamic at work that tended to speed up the killings of Tutsi behind government lines as the RPF advanced. What would have happened in the south-west without the French therefore remains unclear.
The numbers are also uncertain. French government claims that Opération Turquoise had saved "tens of thousands" seem exaggerated. Probably the vast majority of the Tutsi population and associated Hutu in the south-west had been killed by the time French troops entered.[113] Some 13-15,000 persons remained in camps and sites guarded by a hostile gendarmerie, including 11-12,000 in the Nyarushishi death camp in Cyangugu (Prunier 1995, ICRC, interviews 1995, African Rights 1995:1147). These were freed by the French.
The issue of lives saved is central in evaluating the mission because it was launched and endorsed by the United Nations as a "strictly humanitarian" operation (Res.929(1994). By that criterion, the mission was not ineffective, but the timing made its potential underutilized. If a similar effort to save lives had been undertaken in April rather than in late June, the number of lives saved undoubtedly would have been much higher.
The late date of the intervention fuelled widespread suspicions that the main purpose of Opération Turquoise was to save France's erstwhile clients in Rwanda from total defeat. A careful analysis of the planning of the operation suggests that this was not the principal political aim, at least initially (Prunier 1995).[114] Moreover, the intervention was first proposed by the Minister of Foreign Affairs, Alain Juppé, evidently in response to mounting pressure from NGOs and the media to stop the slaughter. Simultaneously, critics focused on France's role in arming and supporting the government on whose side the atrocities were being perpetrated. An intervention designed to save lives and help bring relief supplies to displaced persons would be an atonement of sorts. Media coverage was an important element in this respect, and official guidelines for the operation directed the troops to "demonstrate great understanding towards journalists and humanitarian organizations."[115]
Commonly offered, the atonement explanation seems plausible. A (London) Times journalist captured the sense when reporting on 1 July that "a week after starting the rescue mission, French Special Forces finally discovered 250 Tutsi people yesterday." It also helps to explain why the French pulled out as promised after two months, leaving the RPF to take full control of the zone. In this respect, French policy was consistent with the trend observable since 1990 of rendering only half-hearted support to the regime: the Habyarimana government had received significant aid, but not enough to hold back the RPF.
If Opération Turquoise was a humanitarian gesture for some, this did not exclude consequences of another kind. As the mission was implemented, it became clear that it had the effect of giving some protection to retreating government forces. By late June, government troops were falling back as RPF units moved quickly into the south-east region, and there was a similar rout in the rest of the country. Strong pro-FAR circles in the Ministry of Defense saw the French intervention as an opportunity to help the Rwandese army, which for years they had been training and equipping (Prunier 1995). Soon after a "safe humanitarian zone" had been declared on 4 July, French commanders announced that RPF rebel forces would not be allowed to enter, and attempts to do so would be met with force.[116] To RPF Secretary-General Theogène Rudasingwa and more neutral observers as well, this was clear evidence that the operation was exceeding its UN mandate, which specifically prohibited an "interpositional force". Force Commander General Lafourcade sharpened the partisan edge of the intervention by declaring on 11 July that members of the "interim government" - who were directly linked to the genocide - would be allowed to seek asylum in the French zone. A contrary statement from the Ministry of Foreign Affairs declared that they would be interned, but the military were on the ground and in fact made no arrests. Instead, French forces were observed by journalists and foreign relief personnel escorting members of the former regime out of the zone.
For a humanitarian intervention force of 3,000, the unit was heavily armed (including air support from four Jaguar fighter-bombers and four Mirage ground attack planes based in Goma, Zaire). This discouraged the RPF from challenging the intervention militarily, even though the RPF feared that the French planned to divide the country and dig in. The RPF instead accelerated its advance so as to secure Butare and Kigali (2-4 July) before the French could do so. After 4 July, a dividing line between the two armies running just west of Butare was tacitly accepted. Apart from two incidents, there were no military confrontations: the French ceased their advance, and the RPF did not press forward into the zone. Effective and early communication between the two parties was instrumental in avoiding an escalating conflict that neither wanted. Established in Paris just prior to the operation and continued throughout, the communications structures permitted careful management of a conflictual relationship.
Within the zone, the French operated by and large in a friendly environment; indeed, the embrace of local Hutu leaders proved something of an embarrassment. This removed some of the security concerns that had weighed on the planning of UNAMIR II. With a formal welcome from the Hutu in the south-east and tacit agreement from the RPF, the intervention became a relatively low-risk venture.[117] The point is important since the efficient execution of Opération Turquoise is often held up as a standard by which the more reticent UNAMIR II can be judged. In fact, the two operated in different security environments, and as they also commanded very different fire-power, they are not fully comparable.
Longer-term consequences of the intervention stem from the friendly relations between the French and the people in their zone. French forces were instructed to disarm the militia.[118] This they did when encountering paramilitary groups, but apparently did not follow a consistent and proactive disarmament strategy based on cordon-and-search operations (Prunier 1995). Importantly, there were no official instructions to disarm government troops. As a result, the French zone facilitated an orderly retreat of FAR units through the zone and into the South Kivu area of Zaire with their weapons intact. Two major military camps - as distinct from designated refugee camps - were formed near Bukavu, the main destination for refugees moving out of the south-west. A UNHCR report estimated that the Bukavu camps held 10,000 soldiers, or about one-third of the Rwandese army (UNHCR/ FRS/ A/04: para.5).[119]
The French did not "rescue" either the government forces or the militia in the sense claimed by RPF and other critics. If the French zone had not been declared, both soldiers and militiamen would probably have moved across the border sooner rather than later. A more fundamental problem lay elsewhere. Neither the UN mandate for Opération Turquoise nor the objectives the French government set for the mission included disarming soldiers. Yet the very concept of a safe humanitarian zone implies a demilitarized site, and has developed as such in practice by humanitarian agencies (ICRC, 1993). The failure of the French proactively to disarm government troops within the zone must count as the cost of a significant opportunity lost, or rather, deliberately bypassed. Even if partially successful, this would have helped the successor government deal with the defeated enemy in an orderly manner and in accordance with legal norms. Similarly, French policy in the zone was not to screen or arrest suspects in connection with the genocide, nor was this included in the mandate requested from the Security Council. General Lafourcade's staff claimed to have compiled a list of suspects, but these were not turned over to UNAMIR, which moved in when the French pulled out in late August (Human Rights Watch, 1995).
On the other hand, the French intervention clearly helped reduce the refugee outflow from the south-west, thus accomplishing a major objective of the mission. This greatly facilitated relief operations (see Study III and IV). Given the huge outflow of the time, this issue was so significant that the success or failure of Opération Turquoise in large part came to be judged against its capacity to stabilize the situation in the south-west and stem the outflow. The reverse side of this dynamic was that militant communities were established inside the French-protected zone instead. Opération Turquoise thus served to internalize the problem of "refugee-warrior" communities. In some respects, these camps posed fewer problems for the new government than the large concentrations in exile. The last and critical phase of dismantling them turned into a disaster when RPF forces massacred thousands at the Kibeho camp (see chapter 6.3). The French intervention made the Kibeho massacre possible, though certainly not inevitable.
Opération Turquoise also had more long-term and indirect consequences. In bilateral terms, the intervention further worsened relations between France and the new government in Rwanda. The uneasy mix between official humanitarian objectives and political motives at first worried humanitarian organizations, which feared the political shadow would compromise their relief operations. As the intervention turned out to facilitate their work, however, they came to applaud it. Others concluded that it was a disingenuous coupling of national interests to a relief operation, which served to discredit both the UN and the very concept of a humanitarian intervention. For several African leaders, it was additional evidence that a major European power could manipulate the UN and humanitarian operations to demonstrate its own power in the region. The five developing countries on the Security Council marked their scepticism by pointedly abstaining on the resolution authorizing the force. In this perspective, Opération Turquoise was yet another case of powerful Western states using the UN to legitimize their interventions in the South.
The long-term costs of the mission are intangible. The demonstrative success of a national operation further undermined the peacekeeping operations of the United Nations, possibly lending support to potentially destabilizing unilateral interventions. Use of different mandates in the same country added to the confusion and undermined a principled use of international law. To have free hands, the French obtained from the Security Council a mandate that invoked a "threat to international peace and security" under the UN Charter's Chapter VII. However, the UN force already in Rwanda, and the expanded UNAMIR being planned, had only a Chapter VI mandate, the implication being that Rwanda was not for this purpose a threat to international peace and security. The result may have strengthened the forces of cynicism and cast doubt on the very concept of humanitarian intervention.