The UNHCR and Refugee Return from Zaire to Rwanda

Posted: December 29, 2010 in Genocide Denial
Tags: , , , , , , , , , , , , ,

By Kurt Mills


A simmering humanitarian and security crisis in Eastern Zaire came to a head in late 1996 when a civil war broke out and hundreds of thousands of refugees returned to Rwanda. Many different actors were involved in the complex situation in this unstable part of Africa, including governments—both local and international—rebel groups, genocidaires, the UN, and NGOs. The focus of this chapter is to examine the role of one key actor—the UN High Commissioner for Refugees. As the de facto humanitarian coordinator for much of the crisis, UNHCR was a central actor in the humanitarian, political, and security milieu of the Great Lakes region of Africa. In one sense, UNHCR, as well as the rest of the international community, played a significant role in the maintenance of an extremely destabilizing situation in Eastern Zaire. It provided massive amounts of food and other materials to refugees in the camps. This assistance, as well as the very existence of the camps themselves, helped fuel the continuing conflict in the Great Lakes by supporting and providing the bases from which the militants conducted their destabilization campaign. UNHCR, which had primary responsibility for the camps, tried many ways to get the refugees to return. While being given much responsibility, it was also marginalized by all the major actors in the region and the role it played in the actual repatriation of the refugees was, in one sense, relatively small. It came under immense pressure from many different quarters, including governments and many nonstate actors. At times it was faced with situations which seemed to pose it with impossible choices between carrying out and violating different parts of its mandate.

The experience in the Great Lakes raises a host of questions related to the role of humanitarian actors in complex emergencies, in particular those having to do with the relationship between humanitarian action and political/security interests. For example, can humanitarian action inflame insecurity? How do humanitarian actors mediate between different parts of their mandate which may come into conflict? A key issue raised had to do with the applicability of the principle of nonrefoulement, which prohibits forcibly sending refugees back to a situation where they may be in danger. In the Great Lakes, many of the refugees were forced back to Rwanda. Yet, it seems most were refugees against their will and were, for a number of reasons, unable to actually make a decision whether to return or not.  This posed a seemingly intractable dilemma for UNHCR and others. However, this is a dilemma which will probably be seen again, so it is necessary to more fully investigate the situation and implications of various courses of action.

This chapter, then, will investigate the role UNHCR played in Eastern Zaire during the crisis in 1996 and early 1997. It will look at the pressures it was under and the dilemmas it faced, focusing on the efforts it undertook to protect the refugees and return them to Rwanda.

Background: 2 Million Refugees

On April 6, 1994, a chain of events began which ended in the slaughter of approximately 800,000 Rwandans in what was certainly the bloodiest few months since the end of World War II. As a result of the genocide and the ensuing fighting between the government and militia forces on the one hand, and rebel forces on the other, more than 2 million people became refugees outside of Rwanda, and many others were displaced within their country. About half of these fled to eastern Zaire. As the Tutsi-based Rwandese Patriotic Front (RPF) swept in from Uganda to challenge the Hutus who were engaging in the genocide, many Hutus fled the fighting and their perceived future in terror. These Hutus included many genocidaires—those who had participated in the mass slaughter—and many ordinary Rwandans who played no role in the genocide. The reason why the former group fled is obvious—they feared, probably rightly, what would happen to them if the Tutsi RPF caught up with them. The second group had more, and more complex, reasons for leaving. First, some, like refugees around the world, simply fled the brutal fighting. Second, they were escaping what they perceived to be a fate similar to those of the genocidaires. They were afraid that the RPF would indiscriminately go after any and all Hutu. This fear was exacerbated by propaganda put out by the genocidaires. Third, many Hutu were forced into refugeedom by their fellow Hutu, to serve as shields for the leaders who planned and executed the genocide.

These latter refugees, who formed the majority of Hutus in exile, were kept from returning to Rwanda for three reasons. The first was fear, the same fear that impelled them to leave in the first place—fear that the new Tutsi regime would take indiscriminate revenge on them. This fear was fueled by the genocidaires, who would certainly have been targets of retribution. Second, the refugees were kept from returning by force and intimidation on the part of the refugee leaders. Third, and relatedly, the refugees stayed as a result of a defining feature of Rwandan society—pervasive top-down control and decision-making. The individuals were not used to making decisions on an individual basis; rather, decisions were made for them, as part of a group. This combination of factors lead to one of the central dilemmas of the crisis which may characterize other refugees crises in the future. How can the principle of voluntary return be squared with a situation where it is not necessarily return but exile which is not voluntary?

This mass of people allowed the approximately 50,000 former Rwandan soldiers (ex-FAR) and militia [2] to hide out and organize themselves for a return to Rwanda to retake power. Exile was, as one commentator put it, “the continuation of the war by other means.” [3] Although there were formally elected leaders of the refugees in the camps, the real power in the camps lay with former high ranking officials in the Rwandan government, former members of the Rwandan Armed Forces (ex-FAR), and militia. They controlled the camps, prevented refugees from going back home, and organized the military actions that ultimately destabilized the border area and beyond.

The massive refugee camps provided cover for the genocidaires and those who wanted to regain power in Rwanda. Thus, a main responsibility for the leaders was to keep the refugees from returning. This task was facilitated in a couple of different ways. First, as noted above, the nature of Rwandan society is such that most decisions are collective decisions, made from the top and imposed on the rest of the people. There is very little individual decision-making. Because this is such an ingrained aspect of Rwandan society, it was very hard for individual refugees and families to make the decision to return home, absent such a decision by the leaders. This situation was further reinforced by the propaganda propagated by the leaders which misrepresented the situation in Rwanda at the time. As noted previously, most of the refugees were led or forced out of Rwanda by the genocidaires rather than being forced out by the advancing RPF forces. While perhaps the new government might have wanted to marginalize the returning Hutus, [4] it was the leaders who were in the most danger. Yet, in order to scare the refugees, the situation was made to sound significantly more dangerous than it actually was.

The Rwandan government did not help the situation, however. There were contradictory statements regarding how and when the refugees should return. Refugees who returned were regarded with suspicion, and returning refugees from Zaire was the last priority after returning the refugees from Burundi and Zaire. As the RPF consolidated its hold on power, it seemed that the Hutu population was being systematically marginalized, and the government was not making it easy for the refugees to return. [5] The RPF had engaged in massacres at the end of the war, and as one NGO worker observed, “[b]odies appeared regularly in the Kagera River until May 1995, long after the RPF gained control of the country.” [6] Further, the Rwandan jails were overflowing with suspected genocidaires, and the refugee leaders were able to capitalize on this to demonstrate the danger of returning. [7] They were thus able to maintain and tighten their grip on the camps, which served as the staging ground for destabilization.

Early Efforts to Return the Refugees

Within a couple of months after the refugees left, it was clear to most of the main actors in the region that the refugees needed to return. It was recognized that the camps were likely to become destabilizing. It was also felt that reconciliation would be easier if the refugees returned sooner rather than later because there would be less time for further hate and suspicion of the refugees to develop within Rwanda. The debate revolved around when. UNHCR argued for an early return. [8]

On July 23, 1994 the High Commissioner for Refugees, Sadako Ogata, stated that UNHCR wanted to encourage refugees to return. [9] UNHCR offered to take a group of 20 camp leaders from Goma back to Rwanda so they could see for themselves the conditions the refugees would be returning to and report back to the camps. They refused, an indication of the resistance and defiance among the leaders. [10] UNHCR first tried to repatriate refugees from Goma on August 23. This was stopped by the militias and camp leaders, who stepped up intimidation and violence to prevent the refugees from repatriating or even talk about repatriating. Even so, many thousands of refugees did repatriate from Goma during the last part of 1994, accompanied by a Zairian security escort to the border, although this represented only a very small percentage of the refugee population. At the same time, UNHCR suspended its repatriation assistance to Rwanda in September because of alleged human rights abuses there. [11]

Throughout this period, two issues permeated discussions within UNHCR about return. The first was the actual conditions in Rwanda. That is, was it safe for the refugees to go back? Conditions varied throughout the country, UNHCR did not have access to significant portions of Rwanda, and there was much contradictory information. It was thus hard for it to make a firm decision. [12] The second issue was that of the voluntary nature of the return. The principle of voluntary repatriation is generally recognized. [13] Yet, given the situation in the camps and the fact that the refugees, for the most part, were not able to make the decision to repatriate because of the intimidation and violence on the part of the militants, this principle took on new meaning. If the refugees themselves are not able to make the decision whether or not to return, almost by definition any returns, particularly on a large scale, will not be “voluntary.”

By October it was clear that the militants were expanding their organization and consolidating their grip over the camps. An October 21 UNHCR press release stated that “‘In some camps, the former authorities have virtually taken control of all food and relief distribution in order to consolidate their power and to manipulate and dominate the camp population.’” [14] Earlier in the month, Ogata had said that “‘The aim [of the camp militants] appears to be to control the refugee population, block their voluntary return to Rwanda and build resistance against the Government in Kigali.’” [15] By the end of the year, some NGOs pulled out of Goma and Ngara in Tanzania because of the worsening situation, and although UNHCR considered this, as Dennis McNamara, Director of the UNHCR Division of International Protection, has stated: “the agency’s mandate and the humanitarian imperative of caring for the majority of vulnerable and needy civilians, women and children, made a withdrawal impossible.” [16]

Since it was the genocidaires and militant leaders who were perceived as preventing a resolution of the refugee crisis, the obvious first step was to separate out these leaders from the rest of the population. Various requests and plans to accomplish this were floated as early as mid-August 1994, when the Zairian Prime Minister requested that 20,000 ex-FAR be relocated away from the border. When family members were included, the total number of people to be relocated climbed to 80-90,000. Identifying all of the ex-FAR would have been problematic, particularly since some of those wearing uniforms were not, in fact, soldiers. This plan would have been very costly, estimated at $90 to $125 million (whereas it was estimated that the ultimate solution—see below—cost about only about 2% of what a full-fledged peacekeeping operation would be). [17] Further, such an undertaking would have entailed deploying a significant number of troops, both in the camps to maintain security during the estimated six months it would have taken to prepare the sites for the transfer, as well as to separate the soldiers from the rest of the population. It was unclear how much resistance the armed elements would have put up, although this was an obvious concern. The international community showed almost no support for this and by January 1995 the UN Secretary-General admitted defeat in his attempt to put such a force together. Further, he put the responsibility for addressing the security issue on UNHCR, thus abdicating the responsibility of the UN to maintain international peace and security to its humanitarian arm. [18]

A lesser option was then pursued—deploying a security force in the camps. Its mandate would not include separating extremists. Rather, its focus would be to maintain security in the camps for the refuges and make sure that relief supplies could be distributed unhindered. However, it was hoped that the security force would undermine the hold the extremists had in the camps, making it possible for the refugees to make the decision to go home. In fact, this was the main reason UNHCR Goma requested this deployment. UNHCR negotiated with Zaire to create the Zairian Camp Security Contingent (ZCSC). Deployed in February 1995, it was comprised of 1,500 elite Zairian troops and paid for by UNHCR. The Zairian Minister of Defense referred to them as “‘Ogata’s troops.’” They of course were not, and could not be, under the command of UNHCR. While it could try to exercise influence in how the troops carried out their mandate, the orders came from the government. Obviously, however, the government wanted to be able to place the blame on UNHCR in case anything went wrong. [19] Further, Ogata, as head of a humanitarian agency, did not have a mandate for controlling military forces of any kind. This abdication of its security role by the Security Council highlights the fact that UNHCR, and humanitarian actors more generally, are being called upon to carry out functions that they are not equipped to do and which are the purview of international security actors, like the Security Council. The ZCSZ could never be “Ogata’s troops” because the international community has proven itself incapable of providing the support necessary for humanitarian agencies to carry out their work or, indeed, to obviate the need of humanitarian action in the first place. And, of course, there was no legal precedent for such a situation. Further, some felt that the ZCSC was a bad idea, precisely because they were not under UNHCR’s control, and when the war broke out, some of the troops fought. [20] Although the ZCSC did arrest a few “small fry” leaders, most leaders, although known to the ZCSC, remained unmolested by the force. [21] The Zairian leadership showed little interest in carrying out such activities and little outside pressure was exerted by the international community. In fact, the Zairian government provided resources to the militants and rather than wanting to undermine their position, worked to strengthen it. [22] Given the lack of support from all quarters to separate the extremists, there seemed little prospect for a quick return of the refugees. [23]

The first forced repatriation from Zaire occurred in August 1995. The government felt put upon by the international community. A Human Rights Watch report [24] in May 1995 stated that arms were reaching the militants, and Zaire felt implicated. Further, Security Council resolution 1011 of August 16, lifting the arms embargo against Rwanda, directly contravened Zaire’s position. Consequently, approximately 12-15,000 refugees were expelled from the Kivu region to Cyangugu and Gisenyi over several days, beginning on August 19, when 181 refugees were sent by truck to Gisenyi. [25] It is unclear why the repatriations stopped. One senior UNHCR official claims that it was the result of international pressure. [26] The head of the UNHCR office in Goma at the time maintains that the repatriation was not seriously supported by the central government and that while it might have instigated the repatriation, the local officials were on their own. Under these conditions, there was no way that the repatriation could be sustained, and on August 20 the forcible returns ended. [27]

Although it was condemned as a violation of human rights and UNHCR could not support it because it contravened its mandate and other international law, it was still seen as a possibly positive development. As the head of UNHCR Goma at the time writes:

Many officials, including in the relief community, felt that if a forcible repatriation could be organised, with the Zairian armed forces acting as a deterrent against any violent retaliation by Hutu extremists, it would free the population from the grip of extremists and allow a voluntary return to take place…. It was felt, especially in UNHCR Goma (and Kinshasa) that a controlled coercion was much preferable to a violent outburst which was bound to happen if no viable solution was found for the refugees in eastern Zaire. [28]

There was relatively little violence, and the refugees, for the most part, seemed to be moving of their own free will. Those refugees who returned benefited in one significant way. Since they could claim to have been forced back to Rwanda by Zaire, they were not under the same cloud of suspicion as if they had returned more voluntarily. Yet, the hoped for momentum for further voluntary returns did not materialize, and some officials, particularly in Kigali, were coming to the conclusion that forced repatriation was the only way to get the refugees back. In other words, UNHCR should just allow this to happen, and then help the refugees when they were back in Rwanda. Some coercion, probably by Zairian forces, was necessary to break the stranglehold of the extremists allow the refugees to make their own decision (although it was assumed that most of them would decide to return). [29]

Soon thereafter, Zaire announced a deadline of December 31, 1995 for all of the refugees to leave Zaire. UNHCR did not oppose the deadline, and was roundly criticized. [30] Yet, as one UNHCR official intimately connected with the situation in Goma states:

This criticism rested on a traditional analysis of repatriation in a situation which was extreme and could not be dealt with, in a traditional manner…. the refugees wanted to return but were being prevented from doing so by intimidation and violence. Pressure on the leadership was thus necessary, even if it meant pressure on the population as a whole, since the leaders were hiding behind the mass of the refugees. The deadline maintained such a pressure. [31]

In other words, the only way to deal with the situation and allow the refugees to exercise what was assumed to be their option to return was to, in fact, force them to exercise that option. This is, at least on the surface, a significant breach of international norms and, on the one hand, it is rather shocking that UNHCR would not vigorously oppose such an eventuality. On the other, the circumstances in the Great Lakes may require a rethinking of the absolute application of these norms, or at least a realignment of strategies in pursuing refugees protection and solutions. [32]

For a variety of reasons the deadline came and went without the hoped for returns. On November 26, after a meeting with former President Carter, President Mobutu announced that the deadline was being suspended. The voluntary return of refugees, which had increased significantly the previous month, declined precipitously as the one significant tool UNHCR had to promote repatriation—pressure, partly in the form of the deadline—was taken away. [33]

1996: The Return of the Refugees

Throughout the first part of 1996, other strategies were pursued. The Zairian government restricted travel on the part of refugees. Further, Zaire and UNHCR agreed to other actions, coming under the heading “administrative closure.” Much economic activity was quashed by Zairian troops, including the closing of shops. This was very short-lived, however. [34]

The problem of repatriation appeared intractable. While attempting to implement voluntary repatriation, UNHCR Goma also followed another tact. Recognizing that at least for the time being, there was no prospect of the refugees returning to Rwanda, it decided to investigate another durable solution—integrating the refugees within Zaire. It proposed moving those refugees who did not wish to go back to settlements further from the border. It was recognized that this would not be looked on favorably by Zaire. Further, it would undermine efforts to encourage voluntary repatriation. It would help the leaders gain greater control over the refugees and a firmer position within Zaire from which to mount their hoped for return to power in Rwanda. At the same time, UNHCR hoped that it might put more pressure on Zaire to arrest the leaders as it had promised. However, it was difficult to envisage the entire refugee population staying and moving to new locations, and just moving the camps without addressing the issue of the militants would not help the situation. UNHCR considered moving some of the camps closest to the border to deprive the militants of easy access for their raids into Rwanda, while sidestepping the issue of integration. One camp considered was Kibumba, which had a relatively weak internal structure. Many in Kibumba wanted to stay there because they had easy access to their interests, such as property, in Rwanda. UNHCR hoped that moving the camp would undermine the camp structure and induce many of the refugees to repatriate. In turn, it was hoped, this would build momentum for repatriation from other camps. Yet, relocation, at the same time, could also have the effect of making staying in Zaire appear like a realistic possibility for the refugees, undermining UNHCR’s position that stability in the region could only be attained by the return of the refugees. Such a strategy would allow the international community to continue to abdicate its responsibility and allow the humanitarian sector to address the problem while it continued to ignore the issue, as it had for two years. In the end, none of these options were implemented. [35]

At the same time, the security situation in eastern Zaire was getting worse. Infiltration into Rwanda continued, and there were many clashes between the refugees and the local populations. Further, by October a widespread civil war had emerged in Eastern Zaire as a result of an attempt by the government to expel approximately 400,000 local Tutsi—the Banyamulenge. The Banyamulenge, and other forces hostile to the Mobutu regime, constituted themselves as the Alliance of Democratic Forces for the Liberation of Congo-Zaire (ADFL) and counterattacked. The ADFL, which was later to oust Mobutu from power, was also supported by Rwanda and Uganda, who were particularly concerned with ending the infiltrations from Zaire. As the ADFL swept through eastern Zaire, they attacked refugee camps, since they saw ex-FAR and militia as enemies and the camps as enemy bases. Thousands of refugees died, probably at the hands of all parties involved in the conflict—ADFL, ex-FAR and militia, and Zairian troops. And, from late-1996 onwards, all of the major camps were emptied. [36]

As the conflict spread, and the pressures on the refugees from many different quarters  increased, it was clear that UNHCR was involved in a situation which was far beyond its normal operating parameters and that it could not hope to adequately face the situation. In fact, as will be clear, it was, to a very large extent, marginalized as events overtook its capabilities. The High Commissioner told UNHCR’s Executive Committee on October 7 that “‘probably never before has my office found its humanitarian concerns in the midst of such a lethal quagmire of political and security interests.’” [37] As fighting between the AFDL and the Hutus and government forces increased, Zaire once again declared that the refugees should be expelled, and the US said that the camps should be closed. Refugees fled before their camps were destroyed, and eventually all international aid workers were evacuated from Bukavu and then Goma by November 2, thus losing access to information and the refugees themselves. On November 7, UNHCR called on the international community to raise a force to ensure humanitarian access. The debate over this force is covered elsewhere in this volume. Suffice it to say that UNHCR supported the proposed, but ultimately doomed, Multinational Force. [38] The High Commissioner also encouraged the refugees to return in a radio address in October. [39]

In the period November 15-20, 1996, as a result of an ADFL offensive on Goma, approximately 600,000 refugees repatriated to Rwanda from the Goma area, with UNHCR and other aid organizations virtually helpless to do anything until the refugees actually entered Rwanda. UNHCR was able to transport some of the refugees, although for most it was easier for them to walk than wait for trucks. The sheer number of people, as well as their relative proximity to the border, mitigated against a significant UNHCR role until the refugees crossed the border. At that point, UNHCR returned the refugees to their communes of origin. At least 200,000 more refugees fled further into Zaire. In fact, it appears that it was not necessarily an ADFL goal to push the refugees back to Rwanda. Rather, the aim was to destroy the camps, thus depriving the Hutu militants bases from which to launch their infiltrations. There is speculation that the ADFL may actually have wanted to push the refugees further into Zaire, away from the border, making it harder for the militants to operate and, more generally, to prevent the return of an extremely large number of Hutus. If they had wanted to push the refugees back to Rwanda, one observer notes, they would have attacked from the west, thus not giving the refugees that direction to escape, giving them only the option of going east into Rwanda. In any event, another group, the Mai Mai, who wanted to get rid of all Kinyarwanda speakers from the Kivu region, prevented most of the refugees from fleeing into Zaire. [40]

Some did, however, go further into Zaire, sometimes pushed by the Hutu militants who were, yet again, using them as human shields. [41] However, as soon as the refugees poured back into Rwanda, the Rwandan government, supported by the United States, declared the refugee problem solved, saying that no refugees were left in Zaire. There was also a debate regarding how many refugees actually repatriated, prompting the so-called “numbers game,” where Rwanda and the US declared all the refugees returned, and UNHCR, supported by intelligence from other countries, including France, claimed that there were still many refugees unaccounted for. [42] This in fact, became a major international controversy, which brought UNHCR into direct conflict with the biggest international player and the government of the refugees. The key questions are two-fold. First, why was there a conflict, and second, what was the truth?

Although, on the one hand, Rwanda had been periodically calling for the refugees to return home, there still may have been concern about the effect that such a massive repatriation would have on the country. Leaving a few hundred thousand refugees, mostly militants and their families, deep inside Zaire, where it would be difficult for them to engage in their attacks against Rwanda, may have seemed like a good strategy. Two observers who were in Zaire at the time speculate that transforming the Kivus into a buffer region may have been the Rwandan government’s initial goal in supporting the rebels. Further, it would have relieved a little pressure on the land, since the entire Rwandan population, including those that returned from Uganda and elsewhere, as well as those in the refugee camps, was probably back up to its pre-genocide levels as a result of the returns and the high fertility rate and it would have reduced the pressure on the prisons, whose populations were steadily increasing with genocidaires. [43] Thus, the best strategy would be to declare the refugee crisis over, and let the refugees come to whatever fate they might in the Zairian forest. The US supported Rwanda’s position, and in fact, has been a steadfast supporter of the new regime, even as other donor countries have become more wary as a result of apparent human rights abuses on the part of the Tutsi-led government. Much of the impetus for this support came from guilt arising from the US’ standing on the sidelines as the genocide occurred. Further, one strand of US officialdom held that the refugees were being given too much assistance, that this assistance perpetuated the security crisis, and that the refugees should be returned as soon as possible. There was significant criticism of UNHCR in this regard. This was the position of, among others, the Deputy Chief of Mission, Peter Wayle, and the new Ambassador in Kigali, Robert Gribben, who took up his post in mid-1996. Further, by this time the donor countries were suffering from donor fatigue and wanted to resolve the refugee situation so they would not have to support such large refugee populations. [44] In addition, declaring the refugee crisis over undermined support for and eventually killed the proposed MNF, thus relieving pressure for a mission that almost nobody wanted. [45]

There was a valid debate over the actual numbers of displaced persons in Zaire after the repatriations. [46] However, in the aftermath of the return, there continued to be a politically motivated debate on the numbers of refugees left in Zaire. The argument seemed to revolve around a couple of issues. The US argued that 600,000 refugees returned in the mass return, while UNHCR estimated that 500,000 had returned (although the UN provided many different figures in the first days of the crisis [47] ). The US and UNHCR also had different estimates of how many refugees were in the camps in the first place. And, the US argued that whatever refugees were not accounted for had been killed in the fighting. While UNHCR conceded that some may have been killed as a result of fighting, it maintained that there were still hundreds of thousands somewhere in the jungle. It steadfastly refused to alter its estimates, although it did, eventually, accept the US’ figure of 600,000 returned. [48] Part of the problem with UNHCR’s numbers was that it was not able to get an adequate census of the refugees in the camps before the return, thus the numbers were estimates, with no firm ground beneath them. Other observers thought that UNHCR’s numbers were high, although they conceded that there were probably still substantial numbers of refugees in the Zairian forests. The problem was finding them. They were in extremely inhospitable terrain, and it was very difficult to physically search for them. Security was also an issue, since there was still much fighting going on. The US had been providing satellite data, which had been considered reliable, to UNHCR. This stopped, however, once the debate began. The US began to insist that its aerial intelligence, generated by overflights of eastern Zaire showed no refugees. One humanitarian official claims that there was no way that the flights out of Rwanda, which were said to provide the aerial intelligence, could have missed the refugees. Further, the US claim was contradicted by French intelligence, which for a while just hinted at discrepancies. As this official claims, UNHCR “got significantly different information” from different donor governments, thus fueling the debate. [49] Many different sides had an interest in the numbers debate, and as one donor representative argues, nobody knows the exact figures, although he maintains that the correct number was probably somewhere in the middle of the varying estimates. [50]

Regardless of the actual figures, the mystery refugees did start reappearing. An additional 85,000 repatriated from Zaire by the end of 1996. [51] However, many of the refugees had gone north and west, some appearing as far away as the Republic of Congo, where approximately 15,000, many armed, settled in remote villages in the north of the country. About 10% of these were eventually repatriated by UNHCR. Further, approximately 12,000 refugees were repatriated by UNHCR by air from Mbandaka, near the Republic of Congo between May and November 1997. Another 3,700 refugees were repatriated from the area of Mbujimaya, which is south and west of the Kivu region toward Angola. [52] The vast majority of the refugees, however, appeared at Tingi-Tingi, where 160-170,000 refugees lined the road by the end of February 1997. This region, which was north and west of Goma, approximately halfway to the Central African Republic, was very remote and hard to get to. UNHCR planes had to land on the road since there was no airstrip. There was no government authority in the region, which led to a power vacuum and a lack of law and order. Further, there were arms in the camp and UNHCR personnel had little protection, although the government did bring in a couple of flights of arms. UNHCR’s operation collapsed with the advance of Kabila’s troops at the end of February. The refugees fled north and west on February 28 toward Ubundu and Kisangani (where they were not let in). The camp was overrun on March 1, and all aid personnel were expelled on March 3 and 4. UNHCR was prevented from gaining access to the refugees in the Kisangani area. It was declared an Operational Zone by the ADFL, and many refugees were killed. UNHCR was able to reestablish its presence in Kisangani by the beginning of April on the orders of Kabila, although it was far from an ideal situation. UNHCR was given 60 days to repatriate all the refugees. Troops followed UNHCR personnel as they sought out and visited refugees. They were very rough and many refugees were killed. At the same time, UNHCR was not in a position to investigate the situation. The situation was also bad in other ways. The mortality rate in Kisangani was an astounding 62 per 10,000. By way of comparison, the normal mortality rate in refugee camps is .06 per 10,000. A rate of 2 per 10,000 constitutes an emergency, and 4 per 10,000 is declared an emergency out of control. Most of the refugees were begging to go back. It was obvious that they would die if left in the forest. It was seen as a situation of rescue. UNHCR personnel were uneasy because they did not have access to and know the situation in the prefectures from which the refugees came. Yet, the situation was very dire and UNHCR mounted a massive airlift, flying approximately 43,000 refugees back to Rwanda from Kisangani between April and June 1997. A total of more than 60,000 refugees were repatriated from Zaire by air. Another 130,000 repatriated by land during 1997 through Gisenyi, across the border from Goma, and Cyangugu, across the border from Bukavu. UNHCR estimated that by the end of 1997, there were still approximately 37,000 refugees in what was, by then, the Democratic Republic of Congo. [53]

Analysis: The Role of UNHCR

The experience of the Great Lakes, and in particular eastern Zaire, highlights three related trends with regard to humanitarian action which may be prototypical for other humanitarian emergencies in the future. First, humanitarian action has become a substitute for international security action, part of what I call neo-humanitarianism. [54] Rwanda is certainly not the first instance of this. The international response to Bosnia-Herzegovina was a very clear example where humanitarian actors, and in particular UNHCR, were used by the international community to demonstrate its will to respond to a severe humanitarian and security crisis. The international community, until very late, did not commit the necessary security resources to address the ethnic cleansing nor did it provide the necessary security protection to allow UNHCR and other humanitarian agencies to adequately carry out their activities. In the Great Lakes, the onus for almost the entire international response to the refugee and security crisis (which, of course, were significantly the same) fell on UNHCR. The humanitarian community was responsible for the care and feeding of more than 1 million Rwandan refugees in eastern Zaire. This humanitarian activity helped to fuel the destabilization of the region by the exile militants by feeding them and providing them with bases from which to carry out there activities and to hide themselves (i.e. the camps). It is hard to see how the humanitarian community, could have acted otherwise, however, given the scale of human suffering. Further, the blame for the situation in the camps and the resultant destabilization cannot be put on the humanitarian actors. It was the responsibility of other actors—either the government of Zaire or the UN—to prevent or address the activities of the militants by enforcing security within the camps and separating the militants. Zaire, as seen elsewhere in this volume, had little interest in truly addressing the situation. The UN balked at providing the financial and military resources necessary to do the job. The job of providing security for itself and the camp more generally fell squarely upon the shoulders of UNHCR, which basically rented a small portion of the Zairian army to do something that either the army should have been doing already or the UN should have done. And even then, the ZCSC had no real mandate to address the most pressing issue—separating the militants. Thus, the whole of the international community abdicated it responsibility for ensuring peace and security in the region to the humanitarian community which, of course, did not have the capabilities, resources, and mandate to carry out these activities.

Second, as elsewhere in the Great Lakes, UNHCR was faced with the dilemma of forced repatriation. In Burundi, Tanzania, as well as Zaire, militants had a firm grip on the camps, although the camps in the first two countries were not nearly as destabilizing as those in Zaire. Nonetheless, there was still much debate about how to break the grip in these camps and allow the refugees to make a decision to go home. In Burundi, during July 1996, 15,000 refugees were forcibly returned. UNHCR protested, and the repatriations stopped for a while, although this set the stage for the return of most of the rest of the Rwandan refugees in Burundi. In Tanzania, UNHCR, under pressure from many sides, signed a Memorandum of Understanding with the Tanzanian government which said that the refugees should return by the end of the year. Although UNHCR hoped to use this as a way to promote the return of refugees in an orderly fashion, it did not turn out this way. As the deadline approached, the refugees in the camps in and around Ngara, at the urging of their leaders, actually started moving toward the interior of Tanzania. At this point, UNHCR called on the Tanzanian government to intervene, which it did, essentially taking the opportunity to push all the refugees back to Rwanda. As this operation was in progress, UNHCR felt it had no option but to help the refugees as best it could under the circumstances, thus participating in the forced return of 300,000 refugees. [55]

In the case of Zaire, UNHCR found itself in a situation where it was faced with a set of bad choices. There was an unstable situation where refugees were unable to repatriate because of pressure from the militants. The only way to get them to go back seemed to be through some sort of non-voluntary repatriation. Yet, this would violate the principles of voluntary return and nonrefoulement. The latter was operative because under this principle refugees could not be returned to a place where they might be in danger. UNHCR was, in fact, unable to investigate and guarantee that many of the home communities of the refugees were, indeed, safe. It thus found itself in a position where it publicly condemned forced repatriation but privately felt, at least in some form, this was the only viable way to get the refugees back. In this particular instance, return was seen as the only real long-term solution. However, it also highlights a more central trend where repatriation, rather than being seen as just one of three durable solutions—the other two being integration into the country of first asylum and resettlement in a third country—has now come to be seen as the primary solution to refugee situations. [56]

With regard to the issue of forced return, UNHCR was attempting to face a reality which will likely reoccur. While the Great Lakes was atypical in many ways, the apparent expansion of communal conflicts where forced movements of large numbers of people coincide with rebel movements who may take refuge in, or indeed arise out of, refugee populations would seem to indicate that similar situations will arise in the future. More thought needs to be given to how to address these situations, including making reference to the exclusion and cessation clauses found in international refugee law, whereby asylum seekers can be excluded from refugee status as a result of other activities they have engaged in or where the situation in the country of origin has changed such that it is safe for them to go back and thus the refugees are no longer in need of international protection. Regarding the first, certainly participation in genocide and other crimes against humanity would qualify. In Zaire, the main impediments to invoking the exclusion clause were the large numbers of people and the lack of resources. Individual status determination for 1 million people is not a realistic possibility. The only way exclusion could have worked was if the suspected leaders were separated from the rest of the population and held in a secure situation where the more formal legal procedures could be invoked. Separation did not happen, was not seriously tried, and some argue that the situation was such that separation was only a realistic possibility at the beginning of the crisis when the refugees first flowed into eastern Zaire. [57]

Regarding the cessation clause, it would first have to be determined that the situation was indeed better, which was by no means clear in Rwanda, at least in certain parts of the country. In order to do this, UNHCR would also need access to all areas of the country, which it did not have in Rwanda. Once such a determination is made, however, it must then be decided how the refugees should go back. Unceremoniously pushing them back with the use of military force would not seem to be the humanitarian answer. However, the experience of the Great Lakes indicated that such tactics could work. First, the use of force on the part of state militaries in Tanzania and Burundi did get refugees back in a relatively mild manner with relatively few incidents. And, it must be noted that most of the refugees were happy to be going back. In Zaire, it was a somewhat different story, where refugees returned as a result of forces other than national military forces, which might have at least some constraints put on them by international public opinion. Most of those who returned to Rwanda in the initial mass repatriation were happy to be going back. The problem came with those who did not return to Rwanda immediately. They found themselves in horrendous humanitarian situations, and an unknown number were killed, mostly by the ADFL. Further, while UNHCR was able to play at least some role in the returns from Tanzania, it was excluded from any action to protect and aid the refugees until they returned to Rwanda or until they emerged from the forest, particularly in Tingi Tingi and then Kisangani, in a very desperate situation. Clearly, in the case of eastern Zaire, the refugees would have benefited, and the overall security situation would have changed significantly, if the refugees had returned earlier. UNHCR recognized this and this is why UNHCR Goma quietly supported some sort of forced repatriation to break the grip of the militants and thus allow the refugees to return. [58] Things certainly did not turn out how these UNHCR personnel imagined, but one can wonder what would have happened if the international community had, in fact, listened to, and acted upon, such views.

This brings me to my third point. Even as humanitarian actors and UNHCR in particular, are given the lead role as substitutes for concerted international action, they are marginalized. This occurs in two ways. First, UNHCR was not given the resources it needed by the international community to carry out its mandate. Second, it became a backseat observer as events passed beyond its grasp. As the ADFL swept through eastern Zaire, it had to evacuate and could not carry out its activities. As the Tanzanian government pushed refugees back to Rwanda it felt it had to go along, thereby undermining the principle of voluntary repatriation. It could not do everything that needed to be done to protect refugees and maintain all of their rights—which would have entailed separating the militants—because it did not have the mandate or resources to do so. Insofar as it was able to aid millions of people and engage in some protection activities, UNHCR could be considered to have been successful to a certain extent in the Great Lakes. Yet, it was not, in the end, allowed to be completely successful by all parties to the conflict, including the rebels, the Zairian government, Rwanda, and various international actors, including the United States, which, particularly once most of the refugees had returned, became obstructionist rather than helpful.

UNHCR has been extremely flexible and entrepreneurial in many circumstances, pushing the boundaries of it formal mandate to assist great numbers of people, negotiating access, at times being outspoken on the international stage. [59] Yet, the experience in the Great Lakes demonstrates two important points with regard to UNHCR as an international humanitarian organization. First, it has spent more and more of its resources and energies providing humanitarian assistance rather than protection. Some would argue that this indicates that UNHCR has strayed too far from its intended path to protect refugees. [60] I am not sure this is true. The two activities cannot necessarily be divorced from each other. To the extent its protection activities may have suffered, this may have more to do with whether it has been allowed and given resources to provide that protection. This leads to the second, and related, point. UNHCR is part of an international governmental organization and is thus beholden to state interests. When those state interests do not coincide with real protection activities, UNHCR will obviously be constrained. It cannot raise its own military force to separate out militants, there can be no “Ogata’s Troops.” If it is not given the best intelligence, it may have a harder time finding refugees, as when the US cut off its reconnaissance information after November 1996. UNHCR has perhaps the most developed capability to provide humanitarian assistance in the world. In some cases this may not be adequate to address a complex humanitarian emergency without other requisite resources which it, itself, cannot provide.


In eastern Zaire, UNHCR, as a humanitarian organization, found itself to be out of its depth. Beyond pure humanitarian issues, UNHCR was faced with security and political issues which it had neither the mandate nor the resources to deal with. Humanitarian crises never occur within a political vacuum, and UNHCR has been able, to varying degrees of success, to deal with the political milieu in which it has found itself on many occasions. The Great Lakes, and in particular eastern Zaire, was unique, however, in the severity and brutality of many of the main actors who created and fed the humanitarian nightmare. UNHCR, as well as the many other humanitarian actors, was expected to work in, and indeed address, an exceedingly complex and insecure situation with virtually no backing from the international community which had set it the task in the first place. This was not the first time the international community has used humanitarian action to avoid squarely facing a situation. It perhaps, however, represented an extreme. In light of its refusal to stop the genocide and then to address the situation which resulted as a result of its inaction, it is hard to see how the international community could become any more cowardly. It set up UNHCR to fail, even if not intentionally. Yet, given the extreme situation, it is hard to rate UNHCR as a failure in eastern Zaire. It has come under great criticism for creating and perpetuating the instability in the region by maintaining the camps. Yet, this is the job it was sent to do, and when it tried to truly address the broader situation, its pleas fell on deaf ears.

[1] This paper is partly based on more than 50 interviews conducted during the summer of 1998 in Tanzania, Burundi, Rwanda, and Geneva with UNHCR personnel, NGO officials, and governmental representatives. Because of the nature of much of the information provided here, including its great sensitivity, individuals are not identified. I would like to thank UNHCR for providing me with significant logistical support during my time in the Great Lakes, and the individuals who I interviewed, who were most helpful in answering my questions.

[2] US Committee for Refugees, World Refugee Survey 1997, (Washington, DC: US Committee for Refugees, 1997): 103.

[3] Joel Boutroue, “Missed Opportunities: The Role of the International Community in the Return of  the Rwandan Refugees from Eastern Zaire—July 1994-December 1996,” Massachusetts Institute of Technology­-Center for International Studies/UNHCR, February 1998, p. 4.

[4] Ibid., p. 7.

[5] Ibid., pp. 9-13.

[6] Tony Waters, “The Coming Rwandan Demographic Crisis, or Why Current Repatriation Policies Will Not Solve Tanzania’s (or Zaire’s) Refugee Problems,” Journal of Humanitarian Assistance [Online], (July 4, 1997), Available:

[7] They could also capitalize on a speech by an official from the Ministry of Justice who said that since there were one million people killed, the government would like to have one million people in prison. Boutroue, p. 13. For more on the situation in the prisons and continued insecurity in Rwanda see Human Rights Watch, “Rwanda: The Crisis Continues,” April 1995, Available: and Amnesty International, “Rwanda and Burundi—The Return Home: Rumours and Realities,” 20 February 1996, Available:\RWANDA.

[8] Boutroue, pp. 4-8.

[9] Ibid., p. 19.

[10] Dennis McNamara, Statement Before the House Committee on International Relations, Sub-Committee on International Operations and Human Rights, Hearing on “Rwanda: Genocide and the Continuing Cycle of Violence,” May 5, 1998.

[11] UNHCR, “Impact of Military Personnel and the Militia Presence in Rwandese Refugee Camps and Settlements,” prepared for the “Regional Conference on Assistance to Refugees, Returnees and Displaced Persons in the Great Lakes Region,” Bujumbura, February 12-17, 1995, p. 5.

[12] Boutroue, pp. 20-1.

[13] For a discussion of the legal status of the principle of voluntary protection, see Michael Barutciski, “Involuntary Repatriation when Refugee Protection is no Longer Necessary: Moving Forward after the 48th Session of the Executive Committee,” International Journal of Refugee Law 10 (1/2 1998): 247-51.

[14] Quoted in McNamara.

[15] Ibid.

[16] McNamara.

[17] Interviews.

[18] McNamara; Boutroue, pp. 41-2.

[19] It is interesting to note similar situations elsewhere. For example, in Ngara, Tanzania, UNHCR also pays for the police presence, and the public face is similar. The District Commissioner, for example, stated that “they [UNHCR] are our employers” and that UNHCR and the police are “sailing in one boat.” At the same time, however, the police withhold information from UNHCR and do not always act in UNHCR’s interest. Further, blame is always put on UNHCR for any problems because of a lack of resources that they say are needed from UNHCR. Interviews.

[20] Interviews.

[21] In fact, some familiar with the situation argue that a UN peacekeeping operation would not have been able to do any better with the same mandate—that is, without a mandate to separate intimidators which, throughout the entire crisis was never seriously in the offing. Interviews.

[22] See the chapter by Roger Winter in this volume.

[23] Boutroue, pp. 42-7. A similar agreement was also signed with the government of Tanzania. UNHCR, “Impact of Military Personnel,” pp. 9-10.

[24] Human Rights Watch/Arms Project, Rwanda/Zaire. Rearming with Impunity. International Support for the Perpetrators of the Rwandan Genocide. Washington, DC, 1995.

[25] Boutroue, pp. 47-8; McNamara.

[26] McNamara.

[27] Boutroue, pp. 48-50. The local officials had so little support from the government that at one point they requested (only partly seriously) that UNHCR loan them some trucks (which UNHCR refused to do) because they did not have enough money to rent the trucks necessary to carry out the expulsions. (p. 49)

[28] Boutroue, p. 48.

[29] Ibid., pp. 49-51.

[30] Ibid., p. 51.

[31] Ibid.

[32] This will be returned to briefly below.

[33] Boutroue, pp. 52-3.

[34] Ibid., pp. 54-5.

[35] Ibid., pp. 56-9.

[36] For a brief overview of the so-called Kivu Crisis in mid- to late-1996 see Gérard Prunier, “The Geopolitical Situation in the Great Lakes Area in Light of the Kivu Crisis,” WRITENET Country Papers, UNHCR, February 1997, Available: On the ADFL, see Wm Cyrus Reed, “Guerrillas in the Midst: The Former Government of Rwanda & the Alliance of Democratic Forces for the Liberation of Congo-Zaire in Eastern Zaire,” in Christopher Clapham, ed., African Guerrillas (Bloomington, IN: Indiana University Press, 1998):145-50.

[37] Quoted in McNamara.

[38] McNamara.

[39] Boutroue, p. 24.

[40] Ibid., pp. 13, 59; McNamara; Interviews.

[41] McNamara.

[42] Boutroue, p. 60; Interviews.

[43] Boutroue, p. 59; Waters; Interviews.

[44] Interviews.

[45] Interviews.

[46] “How Many Refugees Are In Eastern Zaire? Why Estimates Vary Widely,” US Committee for Refugees [Online], (November 26, 1996), Available: See the chapter by Susanne Schmeidl in this volume for a much more in-depth analysis of the debate over refugee numbers.

[47] Prunier.

[48] Boutroue, p. 60; Prunier; Interviews.

[49] Interviews.

[50] Interviews.

[51] UNHCR, “UNHCR’s Great Lakes Operation and the Refugee and Returnee Operation in Rwanda, 1997 Progress Report and the 1998 Programme,” Geneva, 1998, p. 26

[52] Interviews; UNHCR, “UNHCR’s Great Lakes Operation,”  p. 4.

[53] Interviews; “Rwandan and Burundi Refugees—Locations, Movements and Dispersal in Eastern Democratic Republic of the Congo (October 1996-June 1997) [map], June 20, 1997; UNHCR, “UNHCR’s Great Lakes Operation,” pp. 4, 26-7.

[54] I define neo-humanitariaism as “the embeddedness of humanitarianism within, rather than at the margins of, contemporary conflict, and the explicit manipulation of humanitarianism for political or military gain on the ground in a conflict or as a substitute for political and military action.” See Kurt Mills, “Neo-Humanitarianism: The Role of International Humanitarian Organizations in the Emerging Global Order,” presented at the Annual Meeting of the International Studies Association, March 2002, Available:

[55] The situation was, of course, more complicated than this, and involved many actors and pressure from many quarters, and, as in Zaire, did not lend itself to easy answers. Interviews.

[56] As the 1997 General Conclusion on International Protection adopted by the UNHCR Executive Committee states: “voluntary repatriation is the most preferred solution, when feasible.”

[57] Interviews. For a discussion of issues related to cessation, see Barutciski.

[58] In fact, as the head of UNHCR Goma observes, “UNHCR’s concern in the Great Lakes was related much more to security inside Rwanda than to strict voluntariness” of return. Boutroue, p. 20.

[59] On its entrepreneurial role in the former Yugoslavia, for example, see Thomas Weiss and Amir Pasic, “Reinventing UNHCR: Enterprising Humanitarians in the Former Yugoslavia, 1991-1995,” Global Governance 3 (January-April 1997): 41-57. On its innovative “cross-border” and “cross-mandate” operations in Somalia, see Kurt Mills, Human Rights in the Emerging Global Order: A New Sovereignty? (London: Macmillan, 1998): 110-1.

[60] See, for example, “The UNHCR Note on International Protection You Won’t See,” International Journal of Refugee Law 9 (2 1997): 267-73.

(This was published as chapter in Howard Adelman, ed. Analyzing and Evaluating Intervention in Zaire, 1996-97, Africa World Press/The Red Sea Press, 2003)

Comments are closed.