Refugees and Security in the Great Lakes Region of Africa

Refugees and Security in the Great Lakes Region of Africa1

Kurt Mills, Ph.D.

Politics Department

Mount Holyoke College

South Hadley, MA 01075

Richard J. Norton

Professor of National Security Affairs

United States Naval War College

Newport RI 02841

The history of the Rwandan genocide of 1994 has its roots, at least partly, in the displacement of thousands of Rwandan refugees. A similar displacement of refugees is also a critical component in understanding the aftermath of the genocide. The aftermath, among other things created the conditions for the eventual overthrow of the Zairian dictator Mobutu Sese Seko and what is referred to as Africa's first continental war. Since the 1950s Rwandan Tutsi refugees displaced during previous decades and living in exile in Uganda contributed to an increasingly unstable situation in Rwanda. Some took up arms and became refugee warriors. These warriors made up the bulk of the Rwandan Patriotic Front (RPF), which initiated military actions against the Hutu-dominated Rwandan government. The Tutsi refugees remaining in Uganda streamed back into Rwanda once the civil war was renewed and the genocide started, and the Rwandan Patriotic Front began achieving battlefield victories. These factors lead, in turn, to a massive outflow of Hutu refugees to all of the countries of the region. Some of these new refugees in their turn took up arms and returned to Rwanda to carry out attacks. This further destabilized the region and led to counterattacks by the new Tutsi-dominated government. The end result was a classic spiral of conflict that today spans half the continent, contributed to the conditions which resulted in the overthrow of Mobutu, the assassination of Laurent Kabila, of the Democratic Republic of Congo, thousands of dead, and a scavenger-like grab for resources as neighboring countries were drawn into the fray.

The focus of this article is the various linkages between large refugee flows and security. In particular, it will detail how extremely large refugee flows can sometimes lead to, or at least help create, conditions for massive insecurity. It concentrates specifically on refugees in the Great Lakes region of Africa, with a detailed emphasis on Rwanda and Eastern Zaire. It will illustrate how refugees can indeed constitute a security threat on many different dimensions and levels of analysis. These include the individual human security of the refugees themselves and the original residents in the areas to which they fled, societal security, the national security of Rwanda and other states, and international security as a civil war grew into a regional conflict.

The timeframe for this article covers the years leading up to the genocide, the period of the genocide itself in the middle of 1994, and two years following the genocide. First, a brief historical review of important events in Rwanda from the turn of the century through 1996 will be given. Then, an analysis of the interaction on multiple levels between the presence of large exiled refugee populations and security will be provided. The article will then present the conclusion that, in this case, the presence of large refugee populations, or more specifically, the presence of armed elements within the refugee populations, had significant and deleterious effects on security. The article will address "traditional" security elements and take a closer look at how emerging questions with regard to refugees are insinuating themselves into security concerns. It will be argued that issues such as massive refugee flows have led to a reconceptualization of the meaning of security. Finally, the chapter will also examine how the international community reacted to the Rwandan security and humanitarian crises by responding to the latter in order to make it seem that the former was being addressed as well.

Rwanda, approximately the size of the state of Maryland, occupies a portion of the African Great Lakes region. Broadly speaking, it is populated by two significant groups - the Hutu and the Tutsi. The Germans possessed Rwanda until 1916 and the Belgians until 1962, when Rwanda gained its independence. Neither European power was especially enlightened, and it could be argued that Rwanda had gone from bad to worse in terms of colonial rule. In an all too familiar pattern the Europeans introduced new notions of identity and preyed upon old tribal rivalries, in order to solidify their rule and profit (OECD 59 - 70). The Tutsi minority was used by the European powers as a local elite, holding power and favor over the more numerous Hutu (Keane, 16).

As independence approached, approximately 85 percent of the population was Hutu and 15 percent were Tutsis. The Hutu majority entered into armed rebellion in 1959 and brutal fighting ensued (Keane, 19). Thousands of Tutsis fled ethnic persecution and violence into neighboring Uganda. This process would be repeated at intervals as political violence periodically erupted in Rwanda. Eventually, many expatriate Rwandan Tutsis would join Museveni's Ugandan rebel forces (Keane 19-20). When Museveni won, his Tutsi allies left his army and formed the core of the Rwandan Patriotic Front (RPF). The RPF was an armed resistance group dedicated to overthrowing the Rwandan government. Although never formally endorsed by the Ugandan government, it was apparent that the RPF was drawing significant support from Uganda in the form of weapons, ammunition, supplies, intelligence and safe havens (Miskel & Norton, 222-223). For several years the RPF planned, rehearsed and eventually mounted attacks on Rwanda from these Ugandan safe havens. During the same period the RPF recruited heavily, if somewhat covertly, from the Rwanda Tutsi refugee population in Uganda.

In 1990 the RPF initiated another military attack on Rwanda. This effort met with greater success than previous actions and the result was a nasty, but inconclusive civil war that lasted for three years. In 1993 a cease-fire was arranged, due in no small part to French military intervention (Keane, 26). The cease-fire was welcomed by the Rwandan populace, but was anathema to the more influential supporters of the extremist Hutu State (Prunier, 160). The UN became involved in efforts to broker a peaceful settlement and, pursuant to UN Security Council Resolution 872, a small peacekeeping force known as the UN Assistance Mission for Rwanda (UNAMIR) was deployed in country. UNAMIR's mandate authorized traditional peace-keeping missions, including monitoring borders for the shipment of weapons, observing and reporting on the status of the cease-fire and so on. UNAMIR numbered only 2,500 personnel from several countries, with the most technologically sophisticated forces being supplied by Belgium. UNAMIR was not intended nor equipped to engage in wide-ranging, conventional military combat with other military or even paramilitary forces. Small in size, UNAMIR lacked the transportation; sensor and aviation assets that would have been needed to fulfill the assigned mandate, much less more ambitious orders. What transportation assets and heavy weapons that were available belonged to the Belgian contingent. (United Nations and Rwanda, 40) However it is important to note that while UNAMIR could not, in reality, accomplish its mission, it could, especially if the Belgians were involved, successfully establish and defend a modestly sized "safe haven."

At this point, a potential peaceful solution to the Civil War seemed at hand. Rwandan President Habyarimana was prepared to negotiate a power-sharing arrangement with the RPF, under conditions laid out in the Arusha Peace Agreements of 1993 (Jones, 131-56). This arrangement was unacceptable to many other leaders of the Rwandan government, especially the commanders of the Rwandan military. In April 1994, an aircraft with the Presidents of Rwanda and Burundi aboard was shot down as it left Kigali airport. Although not conclusively proven, most military and academic experts believe the shoot-down was conducted by the Rwandan Presidential Guard. (Keane 26, Norton & Miskel, 223, Laegreid, 235) Both Presidents were killed. The Rwandan military immediately reopened hostilities with the RPF. At the same time, Rwandan military and radical political leaders initiated a deliberate, pre-planned campaign of genocide against the Tutsi population and mass murder of moderate Hutus. The killing was carried out by elements of the Rwandan Army, specially trained militias - the Interahamwe - and both willing and ostensibly unwilling Hutu civilians who were pressed into service. There was no doubt that the genocide was intentional. Preparation had gone on for months, some would argue years, and included such activities as the use of hate radio and newsprint, as well as careful psychological preparation of the Hutu populace. Evidence of the genocide was so widespread that General Dallaire, the UNAMIR commander warned of the possibility on January 11, 1994, three months before the genocide began.2

On 6 April Rwandan Prime Minister Agathe Uwilingiyimana and ten Belgian peacekeepers who were serving as her bodyguard were taken, tortured and then killed by the Rwandan Presidential Guard (Laegreid, 235) Before the week was over, France, Belgium and the United States had evacuated their nationals from Rwanda. By 20 April, over the protests of the force commander, Belgium ordered and executed the removal of its troops from UNAMIR. The wisdom of this decision was highly debatable. At the initiation of hostilities, General Dallaire, UNAMIR force Commander, redeployed his troops to Kigali and secured the airfield. Despite a lack of engineering supplies, transport and appropriate combat power, Dallaire's troops were completely successful. In the process, it became clear that, despite the killing of the ten bodyguards, the Rwandan Army had no desire to fight the better trained and led UN forces. UNAMIR secured the airfield and provided a safe haven for as many as 14,000 Rwandans, until the Belgians were recalled. (Laegrid, 235 - 238).

Despite being initially caught off-guard, it was not long before international news organizations began providing the first photographic and video evidence of what was happening in Rwanda to a world audience. However, the killing was usually referred to as "tribal warfare," or "inter-ethnic killings," rather than genocide (Keane, 7) This misidentification may have led to lack of understanding of Rwandan events by the world public, a misunderstanding that was encouraged by the Clinton Administration's deliberate refusal - ostensibly at the prompting of the State Department - to use the term genocide in conjunction with the killing in Rwanda (Lake interview).

By May, the UNAMIR mandate had been expanded. The security council passed resolutions 918 and 955 called for increasing the size of UNAMIR by 5,000 troops. Had this force actually been deployed, at the very least an expanded safe haven could have been provided. However, few countries volunteered forces. Those that did, such as Nigeria, were woefully short of armored transport they said would require. The United States initially agreed to sell armored personnel carriers to Nigeria to meet this need, but a price could not be agreed upon. There was also a U.S. concern about UNAMIR's Rules of Engagement. The result was that UNAMIR was never expanded to an appreciable degree.

Many African states and the RPF viewed this intervention as nothing less than the French taking sides with the Rwandan Army. There was some reason for this viewpoint as there were very vocal, pro-Hutu politicians in France. (Callamard, 175). This deployment raised the possibility of a French - RPF clash. As early as 12 April RPF leaders had warned against any expansion of UNAMIR activities that might impede their military progress (Ibid.). In retrospect it seems clear that the leaders of the RPF were willing to endure the cost of the genocide in order to assure a complete military victory over the Rwandan Army.

And the RPF was making significant progress. In encounter after encounter, the Rwandan Army failed to halt the RPF. As the rebels advanced, the Hutu population took flight, fearing the consequences of an RPF occupation. This was not without logic, as some RPF units were conducting reprisals and a great many Hutu had participated in genocidal killings. Nor was the civilian population alone in fleeing the RPF. As many as 50,000 Rwandan Army soldiers may have eventually gone into refugee camps. (UN, 55)

The RPF continued to advance. Kigali was occupied and by July 21st the civil war ended with an RPF victory. By then, huge numbers of refugees, totaling more than two million people had fled to neighboring countries. Once there, former Hutu political and military leaders quickly re-established their authority and control of camp populations. At Goma and other camps, cholera broke out. On July 22nd, President Clinton ordered a limited-duration, humanitarian deployment of U.S. forces into the region.

As noted above, more than two million Hutu fled Rwanda, more than half going to Eastern Zaire, the majority of the rest going to Tanzania, and some to Burundi (which had been experiencing its own, somewhat smaller scale genocide, and which had also been generating refugees in the region). The Hutu fled the RPF for four reasons. First, some of the Hutu were genocidaires, participants in the slaughter, and they feared, probably rightly, for their lives. Others, who had not participated in the killing, also feared that they would be indiscriminately identified as killers and thus as legitimate targets for retribution from the RPF. Third, some fled the generalized fighting, like refugees all around the world. Finally, many were forced to leave by their fellow Hutu, who wanted to use them as shields and to hide amongst them in what turned out to be sprawling refugee camps in order to be able to escape justice and continue the conflict (Prunier 1995, pp. 265-7; African Rights 1995, pp. 1088-1100).

There were more than 50,000 former Rwandan soldiers and militia hiding among the refugees. It was these individuals who were at the core of the security crisis. They kept most of the rest of the refugees in exile, preventing them from returning to Rwanda. They were able to do this by physical intimidation, spreading propaganda about the situation in Rwanda, and through the very nature of Rwandan society, which is very much oriented toward collective decisions-made at the top and imposed on the people. There was thus a situation where, for many reasons, the refugees could not make a decision whether or not to go home. It was the continuing lack of repatriation that set the stage for the expanding security crisis (Halvorsen 1999; Interviews).

In the months following the outflow of refugees into the neighboring states, the militants expanded their grip on the camps (particularly in eastern Zaire), creating bases for destabilizing raids into Rwanda. Many people saw this happening, and there was a push by UNHCR to return the refugees quickly, which also felt that the sooner the refugees returned, the more likely some kind of reconciliation could occur (Boutroue 1998, pp. 4-8). In July 1994, the High Commissioner announced that she wanted to encourage refugees to return. Its first attempt to repatriate refugees a month later was stopped by the militia and camp leaders through violence and other intimidation. A small number of refugees did, eventually repatriate during the last part of 1994, although UNHCR temporarily suspended its efforts in September because of alleged human rights abuses in Rwanda (Boutroue 1998, p.9; McNamara 1998; UNHCR 1995, p. 5).

The question of security throughout the region continued to be a major issue. UNHCR was not able adequately to assess the situation in Rwanda. It was unclear to what extent returning refugees would be safe in Rwanda (Human Rights Watch 1999; Halvorsen 1999, p. 314; Boutroue 1998, pp. 20-1). At the same time, they were increasingly insecure in the camps themselves. By October, the former Rwandan officials had taken control of resource distribution in the camps, and some NGOs deemed the situation so unsafe that they withdrew from camps in Eastern Zaire and Tanzania (McNamara 1998).

Thus, the core of the insecurity in the camps stemmed from the presence and control of the camps by the militants. There were various discussions regarding getting the militants out of the camps. As early as mid-August 1994 the Zairian Prime Minister asked that 20,000 former Rwandan soldiers be relocated further from the border. This would have had the effect of ensuring, first, that the militants could not control the camps, and, further, that they would be less able to conduct raids back into Rwanda. When family members were added into the equation, the total number was 80-90,000 people to be relocated, with an estimated price tag of $90-$125 million. Further, it would have been difficult to identify all of the ex-FAR, and it was unclear how much resistance they would put up. The entire operation would have taken perhaps 6 months and would have required the deployment of a substantial number of troops during this period. The international community showed no interest in committing to this operation, and by the beginning of 1995 the Secretary-General gave up his attempt to assemble such a force. Rather, he put the onus of maintaining security in the camps on UNHCR (McNamara 1998; Boutroue 1998, pp. 41-2; Interviews).

UNHCR responded by essentially hiring 1,500 elite Zairian troops to act as a security force in the camps. The main task for this Zairian Camp Security Force (ZCSC) would be to maintain general security in the camps and ensure the unhindered distribution of relief supplies. UNHCR also hoped that their presence would undermined the militants' hold on the camps, thus making it possible for the refugees to make a decision to repatriate. The Zairian government tried to portray these troops, paid for by UNHCR, as "Ogata's troops." Being a humanitarian organization, rather than a military one, UNHCR could not command a military force-even though it could try to influence the manner in which the troops carried out their mandate, final command rested with the government. However, by portraying them as such, the government was trying to shift responsibility to UNHCR, as well as the blame in case anything went wrong. This, coupled with the fact that Security Council abdicated its security function to UNHCR, demonstrates how little responsibility those whose job it was to ensure the overall security situation-the host state and the Security Council-were willing to accept. It also illustrates in high relief the willingness of the international community to use humanitarian actors as substitutes for concerted action to adequately respond to threats to international security (Halvorsen 1999, pp. 317-9; Interviews).

The ZCSC accomplished very little, particularly in wresting control of the camps away from the militants. Although it did arrest a few "small fry" leaders, most leaders, even though they were known to the ZCSC, were left free to continue on with their activities.3 Indeed, the Zairian government wanted to strengthen their position, and provided resources to the militants to that end. Thus, there was no real support for separating the militants and thus little prospect for the refugees to return to Rwanda in the near term (Boutroue 1998, pp. 42-7; Winter 1999).

There seemed to be no solution to the problem. The camps provided a safe haven for the militants who were destabilizing the region. The camps existed because of the refugees, and the refugees would not be free to go back until the militants lost their grip on the camps. It was a vicious circle to which there was no obvious way out. One possible way out came in August 1995, when 12-15,000 refugees were expelled from the camps in the Kivu region of eastern Zaire to Cyangugu and Gisenyi. Zaire felt implicated by a Human Rights Watch report released in May, which maintained that arms were reaching the militants in the camps. Further, a Security Council resolution in August 1996 lifted the arms embargo on Rwanda, which went against Zaire's position. In response, Zaire began forcibly repatriating refugees on August 19, which continued for a few days until it abruptly stopped (Human Rights Watch 1995; Boutroue 1998, pp. 47-50; McNamara).

Many officials, including in the relief community, felt that if a forcible repatriation could be organized, 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 (Boutroue 1998, p. 48).

There was little violence and the refugees appeared to be moving more or less of their own free will, thus indicating that they wanted to return. The refugees benefited from this forced return because since they were forced back, they were not under the same suspicion by Rwandan authorities and others as if they had come back voluntary. This experience did not lead to further returns, and it appeared that the only solution was some sort of forced coercion, which would break the grip of the militants and allow the refugees to decide whether or not to go home (although it was assumed that they would decide to return) (Boutroue 1998, pp. 49-51).

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 (Boutroue 1998, p. 51).

During the first part of 1996, other possible options were pursued. The Zairian government restricted the travel of refugees and, with the agreement of UNHCR, attempted to stop much economic activity, thus increasing hardship in the camps, although this did not last long. There was also discussion of moving the camps further from the border. However, this was not supported by Zaire, and it was recognized that this might undermine attempts to encourage voluntary repatriation. Further, without separating the militants, the situation would not improve. In the end, there was no action taken as a result of these discussions (Boutroue 1998, pp. 54-9).

In Burundi, the refugees were in a similar situation. The majority of them were manipulated by the leaders to provide cover. They were subjected to intimidation, propaganda, and psychological manipulation, and there were significant questions about how voluntary their continued presence in Burundi was. UNHCR could not guarantee their safety, and Burundi was not happy about hosting tens of thousands of refugees. Further, the Rwandan government had decided that it wanted to get the refugees back, and decided to force the issue. Rwanda had told Burundi that the refugees should be returned or else it would do it itself. Information is contradictory on exactly what happened, but according to one account, the Counselor to the President of Rwanda and point man on refugee issues came to Burundi in mid-July 1996 demanding that the refugees be returned. Burundi was also sympathetic to this position, since the refugee camps were seen as a security threat. He went to Ngozi where many of the refugees were and two camps were dismantled in one afternoon and the refugees were loaded onto trucks. The Burundian military participated in this, and there were reports that Rwandan troops also entered Burundi to participate in the forced repatriation. Between July 15 and July 20 15,000 Rwandan refugees were forcibly trucked back to Rwanda. UNHCR protested, and the returns stopped or slowed down, for a week or two, but then resumed. By September, 65,000-70,000 had returned to Rwanda (Interviews).

Meanwhile, the security situation in eastern Zaire continued to degrade. The militants continued their infiltrations into Rwanda, and refugees and the local population clashed on an ongoing basis. Throughout the entire region, there was increasing pressure for the refugees to return. And, by October a full-scale civil war had emerged in eastern Zaire as a result of government attempts to expel 400,000 local Tutsi-the Banyamulenge. The Banyamulenge and other forces hostile to Mobutu formed the Alliance of Democratic Forces for the Liberation of Congo-Zaire (ADFL) and counterattacked. They were supported by the Rwanda and Uganda, who wanted to end the infiltrations from the camps. Since they saw the ex-FAR as enemies and the camps as their bases, the ADFL attacked the camps. Refugees fled, aid workers left, and UNHCR called for an interventionary force to address the situation. In the end, between November 15 and 20, 1996 approximately 600,000 refugees returned to Rwanda from the Goma area, while more than 200,000 more fled further into Zaire. Most of this latter group disappeared for a while, although eventually many of them showed up again. A further 85,000 repatriated by the end of 1996, and approximately 190,000 repatriated from Zaire during 1997. By the end of that year, UNHCR estimated that there were still 37,000 refugees in Zaire (by then known as the Democratic Republic of Congo after the ADFL overthrew the Mobutu regime), including many militants who continued to have a destabilizing, if less so, effect on the region. As a result of the continued destabilization, a new civil war erupted in August 1998, which eventually drew several neighboring countries into the fighting on one side or the other (Prunier 1997; Reed 1998, pp. 145-50; McNamara 1998; Boutroue 1998, pp. 13, 59; UNHCR 1998, p. 26; UNHCR 1997; Interviews).

The forced return from Zaire helped set the stage for another forced return in December from Tanzania. Several factors created an untenable situation for the Rwandan refugees in Tanzania. One was the effect that the refugees had on the surrounding area, including resources. After two years, forests had been destroyed by refugees in search of building materials and fuel wood, crops were stolen, and local leaders spent less time on dealing with local concerns than before the refugee influx. Security was also an issue. Ngara, for example, where many of the refugees were, used to be a quiet place, but with the refugees came armed robberies and killings. There was a general feeling among the population that, while the refugees were receiving many resources, they were being neglected and that they should be cared for more. The broader security situation was somewhat different than in other places because the numbers of genocidaires in the camps around Ngara were relatively small (although nobody really knows how many there were). Although there were allegations of military training on the part of some of the refugees, these were never proven. To lessen security concerns and the possibility that the military elements in the camps might be engaged in cross-border activities against Rwanda, UNHCR requested sites campsites further from the border, but the government refused. Further, while there was talk of separating the leaders from the rest of the population, nothing was ever done. UNHCR maintains that it was the government's responsibility to do this, but the government took little action, even though this was an expressed concern of theirs. Doing so would have created more categories of refugees and thus would have required more resources. UNHCR also wanted to screen some of the refugees, but the government said no, preferring that they retain their more unofficial status, which, otherwise, would have probably created more obligations for the Tanzanian government. Thus, as was the case elsewhere, the government did not help the situation by taking actions, such as agreeing to move the camps and attempting to separate out the leaders, which might have at least partially mitigated the situation. Eventually, the government, in conjunction with UNHCR, set a deadline of December 31, 1996 for the refugees to repatriate. The actual return occurred a couple weeks earlier. In response to an attempt by the refugees in the Ngara region to move further into Tanzania on December 13th, the Tanzanian military intervened and pushed the refugees back into Rwanda (Interviews; USCR 1997, pp. 98-100).

It is not the intention of this article to enter a lengthy discussion over the concept of security. This concept, over the last decade, has been the focus of intense academic debate. Much of the debate has revolved around the central issue of just what is supposed to be protected. The ensuing debate has challenged one of the central assumptions of Realist thought that holds protection of the state, it's borders, resources, population and so on is the essence of security. Security was provided primarily through military force, backed by a robust economy and stable political leadership. During the Cold War, the linkage between military activity and security became so strong that an issue that did not involve military force was simply not a security issue. (Baldwin 1997, 9)

In the wake of the Cold War, the state-centric notion of security has been heavily challenged (Buzan et al 1998; Matthew 1989; Booth 1991; Baldwin 1997; Graham and Poku 2000). Among the alternatives to military dominated, state-centric concepts of security are: Societal security (Waever 1995); Individual, or human, security (Booth 1991, 319 - 322, UNDP 1999, 36); and International security (Ullman, 1983, 133). As they apply to the situation of refugees in the Great Lakes region, each of these concepts, as well as the more traditional notion of state-centric security, shall be examined.

The establishment of refugee communities in Uganda in the 1950s and the activities of primarily Tutsi refugee-warriors had a major impact on individual, state and regional security in the period prior to the Rwandan Civil War. Following the RPF's victory in 1994, the presence of massive refugee camps in eastern Zaire, and to a lesser extent, in Tanzania and Burundi created an extremely destabilizing and insecure situation throughout the entire Great Lakes region. In this section, the exact nature of this insecurity will be elaborated within a framework that encompasses four different levels of, or ways of conceptualizing, security-human security, societal security, national security, and international security. The nature of, and response to the threats will be examined for each level of security, accompanied suggestions as to what might have been done differently.

At their most basic level, refugee crises are caused by threats to human security.4 People become refugees because they fear for their personal safety and the personal safety of their loved ones. They may fear persecution, they may be fleeing war, or there may be some other threat to which they are vulnerable. As people become displaced, they are also become vulnerable to other threats to their personal security. They will likely lack access to food, water, medicine, and shelter, and may become dependent upon others for these vital resources. They also frequently lack legal and physical protection and may be at increased risk of intimidation, robbery, physical attack, and other threats from various quarters.

This was certainly the case for Rwandan refugees whether they fled in the late 1950s or the mid-1990s. As noted above, the Hutu refugees were fleeing as a result of a combination of fear of reprisal attacks, the war, and because of the fear instilled in them by the genocidaires who pushed many of them out of the country. While in exile, they were dependent upon the international community for assistance, as well as on militant camp leaders who frequently were able to control the distribution of humanitarian assistance. The predominately Tutsi refugees who had earlier settled in Uganda had fled pogroms, persecution and war as well. (Norton, 91 -92)

Unlike the Tutsi in Uganda, the Hutu refugees generally had little security while they were in the camps. As noted, the militants basically controlled many of the camps, including distribution of resources. Further, they also intimidated the refugees through force and propaganda, and took advantage of traditional societal structures to prevent the refugees from returning to Rwanda. In addition, although the reports of insecurity back in Rwanda were significantly overblown by the militants, there were still lingering questions during much of this time period about how safe at least certain parts of the country were for returning Hutus. Thus, the refugees themselves were in a completely insecure situation. The host governments, particularly in Zaire, could not or did not want to protect them. The international community did not provide the resources necessary to protect the refugees and separate out the militants which would have created a much secure environment for the refugees and allowed them to decide whether or not they wanted to go home. This was again in contrast to the situation among the Tutsi in Uganda, who not confined in camps and where the militants were respected and supported by the refugee population.

In general, the international community failed from the very beginning to ensure human security in the region, beginning with its failure to deploy troops to stop the genocide to not taking any action to ferret out the militants in the camps. It has been argued that the genocide fell upon the Tutsi population with such speed and surprise that there was little or nothing the United States, arguably the most capable military power in the world could have done, even if a decision to intervene militarily had been reached at the onset of the crisis. (Kuperman, 105 -107). This is simply false. Although this argument should be taken to task in many areas, its most significant shortcomings are those involving the probability of success for a military intervention. The argument fails to take into account actual, demonstrated U.S. military capabilities or UN success at establishing and maintaining safe havens during the initial days of the genocide. (Sullivan, Hancher, Palmer, Ikins interviews, Norton 98).

When examining refugee developments after the end of the civil war, the question of using military forces to remove Hutu militants from the refugee camps arises. While it is unclear to what extent a military force could have been successful at separating the militants, such action was seriously under consideration in late 1996. The fact that this option was never even tried represents a failure for the UN. It seems more likely that separation might have had a better chance of working early in the crisis, as refugees were fleeing Rwanda in July 1994, or in the ensuing few months before the militants had consolidated their grip on the camps, but again, there was little international will to attempt or even consider such an option, even as the situation was becoming clearer. The only actors who played a significantly positive role are the humanitarian workers, who, while not being able to provide security in the same way as a military force could, may have sometimes provided partial individual security just by their presence. They have been criticized for actually worsening the situation by supporting the conditions that made the continuation of the camps viable-providing humanitarian aid-but, although some groups did pull out, there was a general feeling (probably correct) that doing so would just worsen the situation by removing the one lifeline the refugees had. Further, it was not their job to ensure security. And, they, themselves, faced significant threats to their personal security during the day-to-day carrying out of their duties.

It also needs to be pointed out that humanitarian organizations lack the training, equipment and doctrine to effectively provide the physical security the camps required. This expertise does exist however, with the Military Police and Civil Affairs Battalions of the United States Army and similar military organizations (Newton interview). Given the conditions existing in the camps in 1996, it is believed that regional, UN or state forces of this caliber would seemed to have had the greatest chance of success.

The question of societal issues - which includes identity/culture, social cohesion, and social and economic well being-is of obvious importance since it was the lack of social cohesion and the perceived irreconcilable nature of identity which led to the exodus of Tutsi refugees in the 1950s and 1960s as well as the 1994 genocide and subsequent refugee flows in the first place. Thus, a general lack of societal security and social instability was a basic cause of the refugee crisis in the Great Lakes and is likely to remain so. The social conflict in eastern Zaire where the Banyamulenge were targeted also affected the refugee situation as it drew outside actors into the fighting in the area which had the ultimate effect of breaking up the camps and pushing the majority of the refugees back to Rwanda.

The presence of the refugees itself had significant effects on the entire region. It brought increased crime, and, in some regions added increased economic and environmental stress. Whole forests were decimated as refugees searched for firewood and building materials. Refugees received resources from humanitarian actors, whereas local residents frequently did not. Incidents of both petty and more major crimes increased in many areas. All of this created resentment on the part of some local residents and officials. This, combined with a perceived lack of attention and resources from the international community led to increased pressure to expel the refugees, particularly in western Tanzania.

During the pre-genocide period, the question of societal security was not raised by any of the regional actors, with the exception of the genocidaires. The Tutsi minority, it was argued, was a threat to the state and to Hutu culture. Genocide was the prescribed solution to this threat. It is the potential to use the argument of societal security as justification for persecution of refugees or indigenous minorities that is most troubling. The arguments for societal security also sound a great deal like those put forth by Japanese whalers who argue that shokubunka, or food culture, should override ecological concerns in regards to hunting Minke whales. (Moshavi, A4) However, as noted earlier, issues related to identity and social cohesion were at the core of what was going on in the Great Lakes. The saga of genocide and its aftermath in the region suggests that while understanding these components is critical, defining them in security terms may result in the grossest human rights abuses imaginable. At the present time the idea of societal security is underdeveloped and thus is of undetermined utility.

Examining threats to state security involves looking at threats to territorial integrity, political autonomy, internal stability, and economic well being. Throughout all phases of the conflict, the presence of large numbers of refugees, to varying degrees, constituted threats to the state security of all of the countries in the region. Within all the Rwandan refugee flows, including those of the late 1950s and 1960s, were individuals committed to carrying out military actions across borders - the refugee warriors. In Uganda, the rise of the RPF represented a two-edged threat. It clearly raised the possibility of involvement with Rwanda in an inter-state conflict, but the RPF was also seen as posing a potential internal threat to the Ugandan regime (Otunnu, 3-49). Refugee communities were the chief recruiting sources for the RPF and also served as safe havens when combat operations in Rwanda were not successful. These refugee communities clearly posed a threat to Rwanda's territorial integrity, economic well being and political stability, as evidenced by a series of armed incursions and eventually the civil war. As the civil war came to a close, Hutu refugee flows contained military personnel and many individuals who had directed, executed and participated in crimes against humanity and who wasted no time in establishing dictatorial control of their camps. As the realization that there were large numbers of de facto combatants within these flows grew, so too did concern that these faux refugees would export conflict to neighboring countries, as well as back into Rwanda.

These concerns were obviously well placed. The presence of the refugee camps, particularly in eastern Zaire, posed major threats to state security. The camps allowed the militants to have a base from which to carry out attacks against the new Rwandan government. This created a continuing security crisis for the Tutsi-led government as it tried to stabilize the country in the aftermath of the genocide. Attacks occurred mainly from eastern Zaire, although there was also movement across the borders from Tanzania and Burundi. And, as will be discussed below, the presence of the camps, which housed militants in eastern Zaire eventually provoked a response, which resulted in the overthrow of longtime Zairian dictator Mobutu Sese Seko. The issue of state security in this context of the Great Lakes cannot separated from the broader issue of international security, and thus will be dealt with in more depth in the next section.

The international security aspects of the Rwandan refugee crisis were and are widespread and continuing, ranging from cross-border destabilization to Africa's first continental war. There were attempts to deal with the crisis by a variety of actors. Regionally, the OAU and affected states tried to deal with the issues through unilateral and collective action as well as appeals for international action. In Uganda, the RPF was given support. In addition to rewarding former allies, and removing the RPF from Ugandan territory, an RPF victory would result in a solution to Uganda's refugee question altogether. It can be argued that this strategy was highly successful, at least in the short term. Other states tried to deal with the refugee flows by establishing camps and restricting refugee movements within their borders. Combined with allowing international aid organizations to address the needs to the refugees, this seemed to offer an interim solution to the problem.

The OAU attempted to use diplomacy to re-establish the cease-fire in the wake of the death of the Presidents (Miskel & Norton, 224). The attempt failed. It was bound to fail as members of the Habyarimana regime were already committed to a course of genocide. Attempts at military intervention were unsuccessful because states in the region either chose not to intervene, such as Uganda, or were unable to intervene, such as Nigeria.5 The majority of the regional actors lacked both the security training and the security tools to effectively oppose Rwandan genocide and so stop the refugee flows at their source.

Regional responses were therefore inadequate. They were inadequate due to a lack of pre-war consensus on how to deal with the RPF-Rwanda conflict and a lack of the traditional security tools that might have stopped the refugee flows and associated security problems before they became overwhelming. The OAU lacked the cohesion, power and capabilities of other regional alliances. It also lacked the will to act. To be fair, given the capabilities of the regional actors, it is hard to see what else could have been done without involving other international actors.

The continuing conflict in Rwanda officially became a matter of wider international security when the UN deployed UNAMIR in 1993 in support of the cease-fire. It came to the fore again when the UN Security Council acted on April 21, two weeks after the genocide began, by withdrawing most of UNAMIR from Rwanda. UN Resolutions in May and June noted that the situation was a threat to international peace and security. Opération Turquoise, launched by the French in the southwest part of the country where there were relatively few Tutsi left to protect (although the public goal was to stop the genocide), was intended at least partly to prevent the destabilization of northern Zaire which, the French perceived, would occur with significant refugees flows to the region. In July, the Security Council directly linked the refugee flows to regional stability (Prunier 1999; Mills 1998, pp. 409-10).

Throughout the entire crisis, in fact, it was recognized by all of the major players that the refugee situation could be extremely destabilizing. Some, and in particular the US representatives in Kigali, argued that the refugees were receiving too much assistance, which was perpetuating the security crisis. Their argument was that the mere presence of the camps created a destabilizing situation because, as did happen, the rebels could use them as cover. According to this logic, if the food, shelter, and other assistance were simply taken away, the refugees would head back immediately. Those subscribing to this theory blamed UNHCR, although they failed to take any real action to address the crisis. In fact, this view was very simplistic and shortsighted, ignoring the very real fear (whether well-founded or not) that many of the refugees had, and the circumstances, which led them to flee in the first place, including coercion by the genocidaires. It also ignores the dire humanitarian situation which did exist and which had to be addressed. (Interviews).

As noted above, the security situation in eastern Zaire reached a particular point of crisis in the fall of 1996. The situation of the Hutu incursions into Rwanda was merged with the attempt to push out the Tutsi Banyamulenge population from the region. As the situation became unstable, there were increasing calls for the international community to send in a force to stabilize the situation, protect the refugees, and separate out the militants from the general refugee population. Canada initially agreed to lead such a force, although without the task of separating the militants. Before such a force could be constituted, and probably to the great relief of all concerned, the situation partially took care of itself when the ADFL, supported by the Rwandan government, broke up the camps and sent most of the refugees back to Rwanda. Some of the refugees stayed in Zaire, including some militants, and continued to be an irritant to Rwanda by carrying out attacks in Rwanda.

In the meantime, the ADFL, led by Laurent Kabila, which initially only had the goal of clearing out eastern Zaire, went on to take over power in Kinshasa, overthrowing Mobutu Sese Seko. Yet, soon after taking power, Kabila seemed to have shed his earlier support of the Tutsi population in Zaire and did little to reign in the remaining militants. In August 1998, more than four years after the genocide and the initial refugee flows to eastern Zaire, widespread civil war broke out again in eastern Zaire, with several different groups being supported by Rwanda and Uganda, who continued to be subjected to attacks from what was then the Democratic Republic of Congo (DRC). Eventually, most of the other countries in the region, including Burundi, Tanzania, Sudan, Angola, Namibia, Zimbabwe, and Zambia have been drawn into the fighting in one way or another. Motives for intervention include supporting the rebels and trying to oust the militants, supporting Kabila, and gaining access to the vast diamond deposits in the DRC (Fisher and Onishi 2000; Gnamo 1999). More recently, Kabila was assassinated. His son, Joseph Kabila, was sworn is as the new leader. It remains to be seen how this will affect the conflict, although he has taken some steps to resolve the conflict. What is clear, however, is that although there were substantial fractures within Zairean society before the outpouring of refugees after the Rwandan genocide, the presence of the rebels in Zaire provided not only a pretext for action by Rwanda and others, but also almost the necessity of action on the part of these states to ensure their own security.

Thus, the unaddressed situation of the refugee camps and militants forces hiding within them, led not only to destabilizing activities against Rwanda, but also to the conditions for the overthrow of the long-time ruler of Zaire and a conflict spanning a significant portion of the continent. Certainly it provided a pretext for the intervention of many parties in the conflict in the DRC (Rwanda and Burundi to address threats to their security, Zimbabwe, Zambia, and Angola to support the government of the DRC, while all the while military leaders from all of the countries are making money from the plundering the resources of the DRC).

The lack of response by the United States and most Western Powers can be explained through the use of a traditional security calculus. There was no real chance that the violence within Rwanda would acquire an intercontinental dimension. None of the combatants were a super-power proxy. While it was clear that France was interested in the conflict there was no chance that French involvement would widen the conflict outside the immediate region. Regional actors that could be expected to become involved in a widening conflict were not of sufficient importance to warrant intervention by the majority of the international community.6 While the merits of this narrow interest-based calculus can be, and should be disputed, it never the less sheds some light as to why the broader international community failed to act.

However, it can be argued that there is another dimension to this level of analysis, that of support for international norms. The existence of the Declaration of Universal Human Rights and the Convention for the Prevention of Genocide indicates that there is a widely held consensus that some human rights violations must not be allowed to stand. In regards to the United States, a belief in universal human rights has been identified as one of seven enduring values of the U.S. population (Adams). The initial refugee flows were caused by genocide. Once the genocide commenced any measures short of the use of force would not have been successful in stopping it. It also seems clear that knowledge of genocide was present in the world community, if evidenced only by the refusal of the Clinton Administration to use, or discuss the use of, the term. Could genocide have been prevented or mitigated? The answer is clearly yes. The success of UNAMIR, and (despite their motives and the potential guilt of those sheltered) Operation Turquoise, indicates that a rapid deployment of combat forces into Rwanda could have stopped genocide - despite the desires of the Rwandan Army to carry such acts out and the desire of the RPF not to be delayed in the military campaign. Although faced with commitments elsewhere, the United States possessed the lift and combat assets required to deal with the problem. It also possessed the political clout and talent to gather international support for such action. It is true that protracted U.S. combat operations would require congressional approval and that such support was problematic. However, there were at least thirty days in which the Administration could have acted and simultaneously made the case for acting to the U.S. public and the world community. It is the authors' personal opinion that this value-driven aspect of security should have prevailed and collective intervention carried out with all possible speed. The fact that the United States and the United Nations failed to do is seen as a shameful abrogation of responsibility.

The United States, by virtue of its world position was the one actor that could have most influenced a successful response to the genocide and war that engulfed Rwanda in 1994. Had the Clinton Administration been guided by their own professed values of respect for human rights and opposition to genocide, a substantially different outcome could have been achieved. As the scope of the Rwandan disaster became apparent to the world population, the Administration claimed that a lack of early warning was primarily to blame for their failure to react to the crisis. In reality there was ample warning (Keane, 26; Laegreid, 231 - 253). There were many contributing factors to the failure of the United States to more forcefully respond to the genocide and refugee flows in Rwanda. The United States and many of the other members of the Security Council had been greatly disillusioned by what was becoming increasingly seen as a disaster in Somalia. The situation in Rwanda appeared to have an even greater potential for disaster. Killing was rampant, and members of the civilian populace were carrying out the majority of the murders. A civil war was raging and it was difficult to distinguish the combatants. The possibility that an intervening military force would come under fire was high. There were other issues, that better fit a traditional security calculus, or that were closer to home, that also clamored for attention. These included Bosnia, Haiti, continuing problems with Iraq and the Chiapas rebellion in Mexico. There was a growing public perception among the U.S. population that the Administration was botching foreign policy. There was no concerted effort by U.S. special interest groups to become involved in Rwanda and clear political signals from the Republicans that any such action would be immediately attacked. All of these circumstances pressed the Clinton Administration toward non-intervention.

The international community generally, and all of the major players specifically, failed to adequately respond to a highly volatile security crisis in the Great Lakes. From the very beginning, instead of beefing up UNAMIR, the UN withdrew most of it. Instead of taking action to separate out the militants from the general population, the Security Council and host governments took almost no action. The question becomes whether there was anything that could have been done to avert what has grown into a continental crisis.

There is debate on what could have been done to stop the genocide in the first place. As argued earlier, the authors are convinced that prompt military action, would have greatly reduced the scope and duration of the genocide and saved as many as several hundred thousand lives. However, regardless of the possible efficacy of military intervention, once the refugees had fled to the neighboring countries, what could the various actors have done? Breaking the militant's hold on the camps would have allowed most of the refugees to decide whether or not to go home (the authors presume most of the refugees would have decided to do so). In the process the militants would have been stripped of the cover from which they carried out their internal and cross-border attacks. Thus, the most effective action would have been to separate the militants as they initially crossed the border from Rwanda. However, given the nature of the mass exodus, this was probably unfeasible. There were too many people crossing the border too quickly. However, in the ensuing months as it became increasingly clear that the militants were increasing their hold on the camps, but had not yet solidified their control, a well trained military police force could have regained control of the camps, ending this security threat before it fully developed. Yet, neither the Security Council, nor state actors in the region, such as Zaire, took any such action. Instead, they either blamed UNHCR for the problem, as was the case with the U.S. mission in Kigali, or called for the refugees to be sent back (Zaire). Given the rapidly evolving nature of the situation at that time, it somewhat unclear what might have been accomplished, but there was not even an attempt to address the situation.

This continued for the next two years. As the situation solidified, it seems that there was less chance of any kind of intervention working to break the grip of the militants, and all of the actors showed no real interest or will to take the necessary actions, even while those same actors placed blame. In the end, the situation was addressed by a series of forced repatriations, and then only partially since many of the militants were able to remain active. The repatriations, carried out by both state actors (in the case of Burundi and Tanzania, and Zaire in 1995) and nonstate actors (in the case of Zaire in 1996-although supported by Rwanda) were a general violation of international norms. They also had divergent effects. For example, significantly more people were killed as a result of the conflict and repatriations from Zaire than from Burundi and Tanzania. This may be explained partly by the characteristics of the actors and the differing situations. In Burundi and Tanzania, while there was some brutality involved, the repatriations went relatively smoothly. Being states, the actors involved had to pay at least some attention to international norms and public opinion. This was especially true of Tanzania, which has a significant stake in being seen as a responsible member of the international community. The ADFL, being a rebel group (albeit supported by Rwanda) and thus outside of international norms to a certain extent, has less of a compulsion in this regard.

Yet, in all of these cases a major goal-returning the refugees-was accomplished, in a situation where nothing else had been able to achieve that outcome. In similar fashion, military action that defeated the Rwandan Army (the RPF) ended the genocide. Timely military intervention by the United States and other major powers could have also produced this result. These facts raise several disquieting questions. Should military units be committed to action in areas of internal conflict in order to curb human rights abuses, knowing that achieving this result will mean inflicting and taking potentially significant number of casualties? Might there also be some instances where the general norm against forced repatriation may have to show a bit of flexibility? That is, in situations where the presence of large refugee camps pose a significant threat to international peace and security and where the camps are under the sway of hostile forces, perhaps the international community should look at ways that the threat from the camps can be reduced by breaking them up in some manner, without, of course, inflicting greater loss of life. This would probably require a concerted effort by international government organizations, rather than just leaving it to states to act in the best interest of the refugees. Such actions should obviously not be entered into lightly, but given a situation similar to that in the Great Lakes, perhaps deserve more consideration. If anything, the case of Rwanda from 1994 to 1996 clearly demonstrates that under some conditions, refugee flows, or at least the conditions arising out of large concentrations of refugees, may be security threats. Furthermore, under these conditions they may be threats that will need to be dealt with by military forces. In some cases military forces may have to be used to address the lack of human security that lies at the source of the flow, or to take actions such as disarming and separating combatants from other refugees.

Refugees and Security in the

Great Lakes Region of Africa