Department of Sociology, California State University, Chico, Chico, CA 95929, USA
tel: (916) 898-6595. email:
(formerly Programme Officer, Tanganyika Christian Refugee Service, Ngara, TANZANIA)


Over 500,000 Rwandan refugees are currently in Tanzania, most of whom arrived in 1994. Current policy is that they are in Tanzania temporarily awaiting improvement of conditions in Rwanda so that they are able to voluntarily repatriate. This has been the explicit policy of international UN organizations, regional governments, and donor governments. As a result, other solutions to the refugee crisis, especially regional resettlement, are not being explored.

Unfortunately, voluntary repatriation is unlikely to provide a solution for the millions of refugees living in Tanzania, or for that matter, Zaire. In part, this is because the Rwandan government cannot jeopardize its own political legitimacy by re-admitting opponents of the current RPF regime. More important, though, the post-war Rwandan populations are again growing due to natural increase. This will result in the doubling of the 1993 pre-genocide population as early as 2013. As in the past, this population, should it be in Rwanda, will likely be dependent on agricultural land for subsistence, leading to a resumption of the tensions. As a consequence, it is argued that broader regional solutions involving repatriation and resettlement may be a more appropriate response to the Rwandan refugee crisis.


Early April 1994 was a time of hope for the warring Government of Rwanda, and Tutsi-dominated Rwandan Patriotic Front (RPF). The parties in the 4 year-old civil war reached yet another Tanzanian-brokered peace accord early that month. The agreement included plans for a coalition government, an integrated army, and plans for the re-integration of some 600,000-700,000 Tutsi refugees, most of whom had lived in Uganda and Burundi since the 1960s and who formed the base of RPF support. This was only the latest such agreement; earlier ones had been thwarted by Hutu extremists who continued to polarize the situation by massacring Tutsi. Nevertheless, Rwandan President Habyarimana, apparently believing that the difficult compromises necessary to form a viable coalition were complete returned to Kigali on April 6, 1994. On the approach to the airport, his airplane was shot down, probably by radical members of his own entourage opposed to peace negotiations. [1] Following the assassination, radical Hutu seized control of the government, the peace agreement immediately fell apart, and radical Hutu militia began assassinating political moderates. The RPF also began a military offensive two days later.

Widespread attacks on political moderates by Hutu militia quickly turned to attacks on the Tutsi minority living in Rwanda, some 8-14% of the population. In the meantime, the RPF army made quick advances toward the south as their government opponents focused on killing all Tutsi, rather than establishing military defenses. As a consequence, two types of refugees began to appear in Tanzania in mid-April: Hutu fleeing from the advancing RPF as they pushed southwards, and the few Tutsi who escaped the marauding Hutu militias. Tens of thousands fled across swamps into Karagwe District in Tanzania in April. Then, on April 29-30, as many as 170,000 refugees, mainly Hutu, crossed the bridge at Rasumo Falls into Tanzania. [2] This group crossed in the brief 25 hours after government troops abandoned the bridge, and an RPF unit moved into close again refugee access. Neither the RPF or Rwanda government permitted free access to the border to refugees in the area they controlled.

Nevertheless, by late June, the total number of refugees in Tanzania had exceeded 300,000, but more were still coming. Most crossed the swamps along the joint border which the RPF had yet to completely control. Others went into northern Burundi, and then moved into Tanzania across the land border with that country, into Ngara District where they were assigned to crowded and squalid refugee camps where they made huts from grass, sticks, and plastic.

In the meantime, the war in Rwanda continued. In July, there was an even larger exodus as over 1,000,000 refugees fled advancing RPF troops for Zaire. Then, in September the war was over. The militarily victorious RPF was then faced with governing a country destroyed by war, a people traumatized by genocide, and more problematic, a Hutu majority which had not supported the RPF conquest. The potential allies of the RPF–Rwanda’s resident Tutsi minority–had been killed by Hutu radicals during the previous months. Thus, by the end of 1994, the world was faced with the challenge of feeding roughly 2,000,000 refugees in Zaire, Tanzania, and Burundi.[3]

Quick Answers: Successfully Coping with the Humanitarian Crisis in Tanzania

By most accounts, the humanitarian response in Tanzania was effective. The presence of the United Nations High Commissioner for Refugees (UNHCR) in Ngara, as well as substantial food stocks pre-positioned by the International Committee for the Red Cross (ICRC), permitted a quick response, despite the size of the influx. Likewise, the condition of the refugees, who walked along a tarmac road to the campsite was better than those who earlier (and later) came across swamps. The food stocks they carried with them lasted until the first mass distributions were organized a week later.

The refugees themselves facilitated the orderly establishment of a community under difficult circumstances. The majority of the refugees in Tanzania were from two areas of Rwanda (Byumba and Kibungo) and reproduced the structures which had existed there, including the political leadership. This made food distributions easier to organize; it also made the recruitment of volunteer labor for the many tasks needed to establish the camps easier to recruit–no wages were paid for the first two weeks of Benaco. A problem with this approach, though was that it was known that some of these leaders had been involved in the massacres of Tutsi in Rwanda, a fact that made them distasteful for the UNHCR and other agencies to work with. The immediate humanitarian need to distribute food quickly in order to avoid major epidemics, control sanitation problems, and establish a source of clean water out-weighed this broader concern, and programs went ahead.

Quick Answers: Unsuccessfully Encouraging Voluntary Repatriation

While the short-term issues of relief were solved and the war fought, broader long-term questions of what would happen to the refugees was assumed to be one of quick voluntary repatriation. In support of this assumption, it was pointed out that a majority of the refugees were peasants who could be presumed to be apolitical, despite the fact that most were Hutu. The reasoning went that once the fighting stopped, and the rains resumed, the farmers would get the “itch” to return to their fields.

The fighting in Rwanda ended by September 1994, at the same time the rains began. The only refugees to return to Rwanda, though, were many of the “old caseload” refugees, most of whom were Tutsi RPF supporters who had lived in Uganda and Burundi since the 1960s. As long-term supporters of the RPF, they were encouraged to come back by the new government which had only a very small constituency within Rwanda itself. About 600,000 [4] of these old case-load refugees returned during the following year, and a disproportionately large number have assumed high positions in the new government. But, the 2,000,000 “new caseload” Hutu refugees, itchy or not, stayed put in the refugee camps of Zaire, Burundi, and Tanzania.

Nevertheless, by mid-1995, the UNHCR concluded that conditions had returned to normal in Rwanda, and that the majority of refugees, and in particular the peasants, could return without fear of persecution. Indeed, planning figures released in Tanzania in July 1995 indicated that 20-25% of the refugees would return in 1995, and a similar number in 1996. As a result, UNHCR began to implement an aggressive policy encouraging voluntary repatriation. This policy was based on what the UNHCR considered to be an improving human rights record within Rwanda itself, a record which was endorsed in 1996 by the Presidents of Tanzania, Zaire, Uganda, Burundi, and former US President Jimmy Carter.

More Quick Answers: Explaining Policy Failure

But the refugees have not returned, despite such planning assumptions. As a result, the UNHCR searched for reasons that the voluntary repatriation policies did not work. In doing this, they adopted a series of programs to counter what were viewed to be the reasons for the policy failure. Most aggressive were efforts to neutralize “intimidators” within the camps who it was claimed stopped refugees from participating in the voluntary repatriation program by spreading rumors and threatening returnees. To counter this, the UNHCR initiated radio programs, distributed newspapers, arranged cross-border visits, and showed video-tapes of happy returnees in the camps. Authorities in Tanzania and Zaire also arrested a few of the “intimidators” who could be positively identified. In Tanzania, the new RPF Governors of both Kibungo and Byumba (Rwanda) visited the camps at the request of the UNHCR to encourage repatriation and offer reassurances of safety for returnees. In February 1996, the Prime Minister of Rwanda himself made a speech from the back of a truck in Benaco assuring refugees that they would be well-treated in Rwanda.

Other refugees went on “go and see” visits to their home areas in Rwanda; they were permitted to visit family on unsupervised visits, and then return to the camps to “tell it like it is.” Social Services were purposely curtailed in the camps, particularly secondary schooling, in the belief that such programs would encourage people to remain refugees. Cost of living raises for refugee workers have also been denied refugees who continue to be paid a maximum of 22,000 shillings (US$ 44) per month. The camps themselves, while maintaining a minimum level of subsistence have turned into massive ghettoes in which refugee freedom is severely restricted.

Assuming then that the assumption of the UNHCR was correct, i.e. that the political conditions inside Rwanda were ripe for voluntary repatriation, this “carrot and stick” approach would be expected to stimulate a large return.

While the programs did encourage small numbers of refugees to return they were not nearly as effective as 1995’s optimistic planning figures had hoped.. Between October 1995 and March 1996 6,000 refugees opted for voluntary repatriation. But in the context of the 600,000 refugee population, though, this was insignificant. The birth rate more than compensates for such numbers: a survey released in March 1996 [5] indicated that there were over 1500 live births in the greater Ngara area that month, meaning that annually there are 20,000-25,000 refugee children being born in Tanzania each year. Also contributing to an increase was a continuing influx of Rwandans. In January 1996, two Rwandan refugee camps in Burundi were forcibly closed by the Burundian Army. The refugees were told by the army that they were to return to Rwanda. They instead headed for the closed Tanzanian border where they were, after some days in the no-man’s-land on the border, given permission to enter Tanzania on humanitarian grounds. This preference by itself, was a fairly strong example of refugees “voting with their feet” about how they felt about the new Rwandan government.

Why did the UNHCR’s well-intentioned voluntary repatriation programs fail? The answer of course is not in a lack of effort, but in flawed planning assumptions about the willingness of refugees to return voluntarily. First, there not been a solution to the underlying political issues dividing Rwanda. [6] Second, it is obvious that refugees outside Rwanda, whether peasant, urban, or elite, continue to doubt the legitimacy of the RPF-dominated government which continues to be dominated by Ugandan-born Tutsi. Third, the RPF government continues to point out that reconciliation is unlikely given the massive crimes committed by the former government, some of whom remain outside the country unpunished.

Actions by the RPF government also discourage repatriation. Massacres conducted by the RPF government towards the end of the war also complicate reconciliation,. Bodies appeared regularly in the Kagera River until May 1995, long after the RPF gained control of the country. The violent shut-down of Kibeho center in April 1995 in which 4,000 Hutu displaced persons died, and the continued detention of 70,000 Hutu on genocide charges without trial contribute to the doubts of refugees in Tanzania and Zaire. Despite public protestations, the RPF government continues to do little to discourage such rumors. Refugees returning on official UNHCR “go and see” have been attacked and arrested as recently as January 1996. Meanwhile the new Rwandan government refuses to negotiate with refugee leaders some of whom, it is pointed out correctly, organized the genocidal massacres committed by the former Hutu government.

The fact that the fighting has ended and a semblance of peace returned to Rwanda is not enough to overcome this basic impasse. Current reasoning by some aid workers on the scene is that without negotiation and a political settlement, there will be no significant voluntary repatriation. Given the crimes of the former regime, and the problems that the RPF government has with legitimacy, negotiation may well be impossible for the foreseeable future. But the basic fact remains: without such negotiation, the refugees will stay right where they are and will not voluntarily repatriate.

If war crimes, negotiation, legitimacy, and minority rule were the only problems, a political solution might be possible in Rwanda’s refugee crisis.. But this is not the only problems in Rwanda. There is still a broader issue frustrating a successful repatriation program: the demographics of Rwanda’s population both inside and outside the country.

The Coming Demographic Crisis in Rwanda

In 1993, Rwanda was the most densely populated country in Africa, had a growth rate of 3.1% with a population density of 271 persons per square kilometer and a population dependent on subsistence agriculture. By way of comparison,. industrialized Germany’s had a population density of 222 per square kilometer. [7] Indeed, many observers believe that this demographic problem, with a large population of landless unemployed youth underlies the recent fighting in Rwanda. [8] The high population growth rate, and the lack of opportunities in the cities meant that a generation of youth emerged dividing up smaller and smaller pieces of the agricultural land. When the implications of this are examined more closely, even if a political solution were possible, practical considerations will make mass return difficult if not impossible.

Rwanda had a 3.1% growth rate in 1991, which means that the population will double in size every 23 years. The killing in 1994 was of course appalling, and for that year at least, the population growth rate was negative; most estimates indicate that from 500,000-1,000,000 Rwandans died, which was 6.7-13.4% of the pre-war population of about 7.5 million.

Many of the dead though have been replaced. In 1994-5, about 600,000 old caseload Tutsi refugees, “returned” to Rwanda from homes in Uganda, Burundi, Tanzania, and Zaire. [9] These refugees were resettled in both the homes of the dead, as well as the homes abandoned by refugees now living in the squalid camps in Tanzania and Zaire.

In addition, all of the Rwandan populations, including those in Tanzanian and Zairian refugee camps, as well as Rwanda, have probably exceeded pre-war fertility. This is because both the refugee populations and that in Rwanda itself include unusually large proportions of child-bearing age women. Estimates of the post-war population of Rwanda itself indicate that about 2/3 of the population is female, presumably due to the larger numbers of males killed during the massacres and fighting. [10 ] Indications from Tanzania are that while the gender ratio is even, very few small children made the trip from Rwanda to Tanzania. In the only demographic survey taken of arriving populations in 1994, it was found that only 12.7% of the population was under 5 years old. Normally, African populations have approximately 20% in this age group. Again the consequence that there will be an unusually large proportion of fertile women in the refugee population without small children. Forecasts from the age pyramid indicate that an especially large number of girls now aged 7-14 enter child-bearing age between 1995 and 2002 (see Figure 1). The will lead this expansion from now until the end of their child-bearing years in 2025.

Figure 1, which is an “age pyramid” from Chabalisa 2 refugee camp in Karagwe District of Tanzania and illustrates why such populations will have a higher than normal birth rate. “Normal” African populations typically have a pyramid shape due to 20% of children below age 5. The arriving population in Chabalisa 2 had very few children in this age group, with the largest proportion aged from 5-20 years old. This means that the number of women in the child-bearing years (i.e. 15-45 years old) will be higher than in the “normal” population in coming years. [11]

Thus in the coming 10 years, the Rwandan populations in Rwanda, Zaire, and Tanzania will expand more quickly as large numbers of women enter child-bearing years. Table 1 shows how fast each population will grow if there is a growth rate of 4% for the coming 10 years, and then 3.5% for the following years. Extrapolations are made to the year 2013, which is 20 years from the last pre-massacre year of 1993.

The sum of this exercise is to conclude that in 2013, the combined population of people claiming a right to live in Rwanda–meaning both those in Rwanda and in refugee camps ringing Rwanda–will be 13.3 million, which is almost double what it was in 1993. In Rwanda itself, there will be over 9,000,000 which, again, is by itself more than the country supported in 1993. Also remarkable: by mid-1998, there will be more Rwandans than there were in 1993 before the massacres. As a matter of explicit international policy, each of these will have a right to claim a home in Rwanda, and no right to seek a new home elsewhere.

In sum, without considering the complicated ethnic loyalties , post-genocide problems of reconciliation, justice, and vengeance, or the difficulties the current Rwanda government has in legitimizing its rule, there will be major problems in Rwanda. The question is not whether these problems will emerge, but how they will be best dealt with. But then, if repatriation is not the solution, then what is?

The Magnitude of the Solution: Making 2,000,000 Refugees Productive

The trick of course is to figure out how to make the 2,000,000 refugees independent and productive. Repatriation may continue to hold some promise for some, but the past record, and the future prospects are not good. Continuing to rely on repatriation as the sole “solution” is a policy which will only lead to millions of refugees sitting on Rwanda’s borders scheming about an eventual return. Ironically, this is what the Rwandan Tutsi who were denied Ugandan or Burundi citizenship did this for thirty years before launching invasion in 1990. This is not a healthy situation for either Rwanda or the poor host countries which will need to devote more and more resources to controlling a hostile alien population housed in squalid refugee camps.

Other problems emerge when it is realized that the refugee camp populations are rapidly being “deskilled” from productive activity. In the refugee camps, land is even rarer than in Rwanda, and subsistence farming is impossible in what are, in terms of population density, clearly urban environments. One result is that a group of young people starting to emerge who eventually will not be able to farm well Total skill loss in 1996 is still probably fairly minor, however after 5 or 10 years of refugee camp life the consequences for young deskilled peasant youth in particular will be severe. The typical response to such a situation–education into trades suitable for an urban economy–is also discouraged due to the insistence on the voluntary repatriation which assume a return to the ante-war status quo of subsistence farming.

Another scenario has been entertained by the Tanzanians is permanent resettlement. Indeed, the Rwandan government approached Tanzania, Zaire, and even Gabon about the possibilities of organized resettlement of up to 1,000,000 peasant farmers in the late 1970s and 1980s. [12] These proposals were rejected for lack of funding for the population transfers. Many Tanzanians would agree that a small proportion of the Rwandans could be resettled here–say 100,000 refugees. Tanzania has implemented such programs for Burundian, Rwandan, and Mozambican refugees in the recent past.

But, the large number of Rwandan refugees currently in Tanzania and Zaire has though blunted discussion of resettlement, and both the Zairean and Tanzania governments are publicly hostile to discussion of resettlement plans. The Tanzanian government is also wary of the political problems that past settlement policies have led to. Both the resettled Rwandan from the 1960s and Burundian refugees resettled in the 1970s have used their settlements as bases for exile return movements.

However, unlike the demographic crisis, the issues of resettlement are probably surmountable, particularly if the problem is viewed not as a local problem for Rwanda and adjacent areas, but as a regional problem. In the context of Tanzania and Zaire only, the resettlement of 2,000,000 refugees in rural areas does indeed appear insurmountable. But, in a broader context, and with a little imagination, such problems becomes more manageable, particularly when the magnitude of migration within Africa and elsewhere is considered.

The Precedents for Solving a Medium-sized Refugee Problem

The speed and drama with which the Rwanda refugee crisis emerged was of course impressive. However, in terms of refugee movements, or more precisely population movements, it is basically a medium sized problem. Many problems have been more massive: the 10,000,000 displaced persons in Europe at the end of World War II, the massive population exchange between India and Pakistan in 1947-8, and Afghan refugee flight in the 1980s were all larger and politically more complex. While the policy, itself was controversial, it is also relevant that Tanzania in the 1970s successfully resettled 10,000,000 Tanzanians from dispersed clearings into socialist Ujamaa villages.

Refugee problems of similar size to the Rwandan situation include the Indochinese refugee crisis in which well over a million refugees were resettled in the United States and other countries, and the Mozambican refugee flight and subsequent repatriation from Tanzania, Malawi, Zambia, and Swaziland. Taiwan, Hong Kong, and Israel are all societies founded on the reception of millions of refugees. Germany in the 1950s accepted several million refugees from East Germany. In the 1990s Germany accepted over one million more ethnic Germans (aussiedler) and asylum seekers from the former Communist bloc countries.. The point is well made: the movement of 2,000,000 is not unprecedented in terms of either repatriation or resettlement.

From such a context, the Rwandan refugee crisis merely becomes another medium sized migration movement; the only major difference being that it did not emerge over a period of 3-10 years as the other movements, but quickly between April and November 1994. But when viewed from this perspective, older examples can offer ideas about where and when each solution is relevant. This requires some subtlety of thought; and continuous discussion among the governments and peoples of the region. But for it to be successful, the blinders of the past two years, particularly the over-reliance on voluntary repatriation policy must be discarded. Refugee crises are solved with both sponsored repatriation and resettlement programs. Large numbers have also been accommodated with unofficial, “look the other way” repatriation and resettlement programs. Indeed, these “unofficial” movements of Rwandan refugees are probably already a significant part of the “solution” to the Rwandan problem. The magnitude of unofficial movement is unknown. However, east African cities like Mwanza, Nairobi, and Dar Es Salaam, are already quietly hosting large numbers of Rwandan refugees. Other refugees are quietly moving back and forth from the refugee camps to Rwanda.

Rwanda cannot and will not be able to support peacefully a population of 13,000,000 agricultural people in 2013. To assume, as current policy does, that this will happen only exacerbates the already fragile peace which has emerged in Rwanda. And yet the assumption of Current policy is that it voluntary repatriation will be somehow possible. It is not, and failure to recognize this means that the demographic time bomb is being ignored.


1. It has never been established who actually shot down the plane which also carried the President of Burundi. Circumstantial evidence points to radical Hutu. However, others also had incentives to de-rail the peace process, including radical Tutsi in the RPF. See Rene Lemarchand (1995), the Rationality of Genocide, Issue 23(2):8-11. Also, Gerard Prunier, The Rwanda Crisis, History of a Genocide, Columbia University Press, 1995.2. The most commonly reported figure for number of people crossing the bridge was 250,000. As a result of a census in early July 1994, this figure was downgraded to somewhere between 70,000 and 170,000 refugees. See Humanitarian Aid and its Effects, Joint Evaluation of Emergency Response to Rwanda, vol. 3 pp. 105-106, for a discussion of how such figures were routinely inflated during the crisis.

3. Burundi, with its own internal problems, began to quietly ease its Rwandan refugees toward Tanzania. As a result, the Rwandan exodus toward Tanzania did not end until April 1995 when Tanzania formally closed its borders with Burundi. By that time, the total refugee population in Tanzania, mostly Rwandans, with a sprinkling of Burundians, had reached 600,000.

4. Catherine Newbury (op cit.) discusses the social origins of the “Rwandans abroad.” Besides the Tutsi refugees from the late 1950s and 1960s, there were also Tutsi who had lived for several generations outside Rwanda. Those in Uganda were also identified with the Rwandan refugees, and as a result often suffered from the same discrimination within Ugandan society experienced by the refugees. Prunier√≠s book (op cit.) is also relevant to this discussion.

5. Statistics released by the UNHCR Medical Coordinator in Ngara indicated that there was a 4.3% annualized birth rate in the Rwandan and Burundian camps in the Ngara area which had a total population of 440,000. Live births were 4.2%. No annualized death rates were available from the Ngara camps. This statistic was for March 1996 only. Further press releases put the rate as high as 6%, a rate which is beyond the capacity of a normal human population to sustain. In any event, the rate is very high.

6. Ironically, The Economist noticed the collapse of any compromise in an article about Rwanda in their September 2, 1996 when they noted that “it is fear of arbitrary arrest and reprisal killings that has kept most of the refugees from returning.” (pp. 36, 41)

7. This has been widely discussed, but the implications never completely analyzed. See Joint Evaluation of Emergency Assistance to Rwanda, vol. 1 pp. 14-16.

See Catherine Newbury, op cit. for brief a discussion of the landlessness problem in pre 1994 Rwanda.

The returnees from Uganda, many of whom were pastoral people also brought with them 600,000 hungry cows putting even more pressure on Rwanda’s diminishing agricultural resources. The RPF government has attempted to solve this problem by assigning the herders to the formerly uninhabited (by people and cows) Akagera National Park. This though will only briefly postpone one of Rwanda’s most pressing problems: the pressure on land by the use of extensive agriculture practices, of which cattle herding the most problematic. Control of cattle herding will be complicated by the fact that the Ugandan Tutsi returnees are among the long-time core constituency of the RPF.

8. The UNHCR reported a 4.3% birth rate in Ngara area camps in March 1994. In Chabalisa 2, birthrates were about 30 per week in a population of 32,000 in late 1995.

9. Refugees themselves are not aware of the threatening statistics. However, they do acknowledge the high birth rates, explaining them as necessary to replace the large number of children lost during the fighting, massacres, and subsequent flight.

10. The plan to move 1,000,000 Rwandans into Tanzania was agreed to by the Tanzanian and Rwandan governments. The program reached the planning stages and the Tanganyika Christian Refugee Service/Lutheran World Federation was to assist with the implementation. The plan was shelved in 1979 or 1980 due to lack of donor interest. These plans were also complicated by the insistence of the Rwandan government that the migrants retain Rwandan citizenship. The result would have been effective extra-territorial rights beyond Rwanda’s borders. See Catherine Newbury, Background to Genocide: Rwanda, Issue (Africa Studies Association) 23(2):12-17.

Figure 1–This is the age pyramid for Chabalisa 2 refugee camp in Karagwe District, Tanzania. Data were collected when the refugees arrived in September-October 1994. Sample size was 7,380 (3,752 males and 3,628 females) out of a population size of approximately 37,000. Sampling was done by cluster samples of blocks from each of the two communes in the camp (Rukara and Kigarama). Summary tables from which the table was constructed is in Appendix 1. Raw data available form author upon request.Table 1–Population of Rwandans in Rwanda, Tanzania, Zaire, Burundi and Total between 1993 and 2017.

Appendix 1–Raw data upon which the age pyramids for Chabalisa 2 was constructed.


Comments are closed.