Military history


Chapter Six

The Politics of Indigeneity in Uganda: Background to the RPF Invasion

THE Rwanda Patriotic Front (RPF) invasion of October 1990 occurred at a time of internal reform—and not repression—in Rwanda. Uganda-Rwanda negotiations on the right of refugees to claim Rwandan nationality had reached an advanced stage by then. An independent survey was to be conducted within Rwandese refugee camps in Uganda to see how many actually wanted to return to Rwanda and how many would prefer to remain in Uganda while claiming Rwandan nationality. On 28 September 1990, only three days before the RPF invasion, President Juvénal Habyarimana addressed the UN General Assembly in New York and announced two key concessions to refugees: Rwanda would grant citizenship and travel documents to those who did not desire naturalization in their countries of asylum, and it would repatriate many of those who wished to return. Supporters of the RPF have preferred to remain silent on this point, while opponents have highlighted it as evidence of the diabolical and power-hungry nature of the RPF. Neither view recognizes that the October invasion was more a response to developments within the region than to those inside Rwanda.

The more the population of Rwanda overflowed its boundaries, the more there developed a political diaspora—besides the earlier and larger cultural diaspora—outside Rwanda. In accepting that some of the 1959 refugees may have a right of return, Habyarimana had acted under pressure. The source of that pressure was not the internal opposition in Rwanda, but the knowledge that a section of the Tutsi political diaspora—that within Uganda—had, even more so than in 1959–63, come to constitute a significant armed and political force. The irony is that both Habyarimana’s concession and the RPF invasion that followed need to be understood within the context of developments in Uganda. The October invasion, I will argue, was the outcome of the first major political crisis the National Resistance Army (NRA) faced since coming to power in Uganda in 1986. At the root of the crisis was the political legacy of the colonial state: that citizenship be defined on the basis of indigeneity. Struggles around postcolonial entitlement had focused on the notion of indigeneity and had led to successive expulsions in independent Uganda. The best known of these was the expulsion of Asians in 1972. But the largest in scale was the expulsion of the Banyarwanda in 1982–83. My point is that it is events within Uganda, and not Rwanda, that set in motion the dynamic that explains both why the Tutsi diaspora in Uganda became a central force in the guerrilla struggle, and why it was marginalized in the conflict about entitlement that followed the victory of the guerrillas. At the root of both developments was the crisis of citizenship and indigeneity. With the invasion of October 1990, the NRA exported its crisis to Rwanda.

Without a regional perspective, it is not possible to understand either the dimensions of the crisis or the intensity it acquired. If the growing pool of refugees in the region can only be explained by internal developments in Rwanda, the organization of the refugees as an armed force and the timing of their armed return in October 1990 can only be explained by developments internal to Uganda. For what has come to be known as the RPF invasion of Rwanda also needs to be understood as an armed repatriation of Banyarwanda refugees from Uganda to Rwanda.


Tutsi refugees left Rwanda in three different waves: 1959–61, 1963–64, and 1973. The first wave followed the revolution. It involved mainly the Tutsi elite displaced from positions of power. The second wave was triggered by the repression that followed each of the major inyenzi attempts to restore Tutsi power through armed attacks. The outflow grew to include broader sections of Tutsi, as the repression expanded beyond perpetrators of the old order to target ordinary Tutsi just in case they may yet support inyenziefforts to restore the old order. The most detailed study of Rwandese refugees in the post-’59 period estimates that between 40 to 70 percent of Tutsi refugees fled Rwanda between 1959 and 1964.1 After 1964, there was a prolonged period of political stability as armed refugee attacks came to a halt from without and the First Republic came to be stabilized within. The political crisis of 1972–73 triggered a fresh outflow of Tutsi refugees from Rwanda, just as it led to a reconstituted political power within Rwanda. Until the October 1990 RPF invasion from Uganda, there was no further outflow of Tutsi refugees from Rwanda to neighboring countries.

How many Tutsi refugees were there in the region in 1990? Estimates vary radically. Most range between 400,000 and 600,000.2 Gérard Prunier gives a higher estimate of between 600,000 and 700,000.3 Captain Eugene M. Haguma of the RPF gives a wildly high estimate of 1.5 million as the number of “Rwandese living in exile” since 1959,4 thereby reproducing, without any explanation, assumptions widely held in Uganda in the early 1990s about the number of Rwandese refugees in the region. Students of political crisis in contemporary Africa are familiar with the wildly fluctuating estimates of those killed or displaced that tend to accompany reports of political crisis.5 Often, they are more a reflection of the political orientation of the writer than the nature of the phenomenon being described. At times they reflect a widely held notion that anything goes when it comes to reporting about small, faraway places. Given this, I have tended to give weight to estimates that are accompanied by a credible account of the procedure employed to arrive at figures. From this standpoint, Catherine Watson’s estimate of the total number of Banyarwanda refugees in 1990 at “probably about half a million” has a greater credibility than any other I have come across.6

If the total number of refugees in the region of the Great Lakes was between 500,000 and 600,000, not all of them were registered as refugees with the United Nations High Commission for Refugees (UNHCR). The registered refugees in the early 1990s numbered a little over two-thirds of this estimate. Most of these were registered in Burundi (266,000), and then in Uganda (82,000), Tanzania (22,000), and Zaire (13,000).7


The Banyarwanda form a distinct cultural group within Uganda. As speakers of the language Kinyarwanda, they constituted the sixth largest ethnic group within Uganda according to the 1959 census, surpassed only by the Baganda, the Itesot, the Banyankole, the Basoga, and the Bakiga. In 1990, the Banyarwanda formed “slightly over 1.3 million of the country’s total population of about 18 million.”8 From the standpoint of Rwanda, however, they constituted the Banyarwanda cultural diaspora within Uganda.

The Banyarwanda in Uganda comprised three distinct groups: nationals, migrants, and refugees. Nationals were those who were a part of Uganda at the time the country’s western borders were defined by the 1910 agreement between Germany and Britain, two colonial powers. Migrantscrossed that border, mainly during the colonial period, in search of a better livelihood. Refugees, in contrast, were mainly those displaced by political crisis in the postcolonial period. Thus, nationals were of precolonial and migrants of colonial vintage, but refugees were solely a postcolonial development. Whereas migrants were mainly economic, refugees were mainly political. All three were part of the Banyarwanda cultural diaspora, but only the refugees constituted a political diaspora.


Ugandan Banyarwanda—those I have termed nationals—live mainly in a single county in the western part of the country. Bufumbira County is part of the district of Kabale. Both the Bafumbira whose home is Kisoro district, and the Bakiga in the neighboring Kabale district, were drawn into the Nyabingi movement that constituted the spearhead of the political struggle against the Tutsi aristocracy and German and British colonialism and spanned both sides of the border. The Bafumbira were estimated at roughly 600,000 in the early 1990s. In the period from 1959 to 1990, however, they did not constitute an active political force.

Besides the Bafumbira, who were mostly impoverished cultivators, there was a second group of Banyarwanda nationals in Uganda. These were the nomadic Tutsi who had long been resident in Ankole, and some of whom still roamed the drylands with cattle. Others had bought land and practiced a mixed agriculture, while the rest had scattered, either moving to other pastoral areas to work as cattle keepers or adapting to an urban life. Catharine Watson estimated the Tutsi of Ankole—using this term for the original group that has now adapted to a range of preoccupations, from cattle keeping to mixed agriculture to urban work, whether within or outside Ankole—at roughly 120,000.9


Migrants constitute about half of the Banyarwanda cultural diaspora within Uganda, numbering between 500,000 and 700,000 in the early 1990s. Labor migrants were mainly Hutu peasants who began flowing across the border in the mid-1920s. But they also included petits Tutsi who came to work as cattle keepers for the Baganda in the south and for others in northern and eastern Uganda. Thousands more Tutsi came in the post-Second World War period, in the 1950s and ’60s, to work as cane cutters on Asian-owned sugar estates in Basoga and as miners in the Kilembe copper mines. Most labor migrants came in response to a double dynamic, on the one hand a push, on the other a pull. The push came from the tightening regime of forced labor and forced cultivation in Rwanda, identified with the newly appointed Tutsi chiefs who were charged with its implementation.

The pull factor was fresh employment opportunities, which varied over time. In the period between the two world wars, employment could be found on a big scale in two locations: the plantations of Busoga and rich peasant farms in Buganda. In the period after the Second World War, new opportunities came with the opening of the copper mines in Kilembe. “I left home because I wanted a job without beatings,” explained one migrant to the anthropologist Audrey Richards.10 For migrants, unlike for refugees, the move across the border was not a compulsion, but an option. From that starting point, they sought to become a part of the local society, gradually assimilating, taking on local names, clan affiliations, spouses, and even an overall identity. The agricultural Hutu assimilated more easily in the agrarian cultural context in Buganda, just as the pastoral Tutsi assimilated more easily in the context of the cattle-centered culture of the Bahima in Ankole. But the Tutsi assimilation in Ankole came to an end with the influx of large numbers of refugees in 1959.11

The preferred destination of agricultural migrants was the rich peasant economy of Buganda. There, they could begin as farm laborers and then aspire to acquire land as tenants, and finally even become land-holding bakopi (peasants). Buganda had a total population of 1,302,162 in 1948. Of these, 34 percent were immigrants. By the 1959 census, the total population in Buganda had risen to 1,834,128. Roughly 42 percent of these were immigrants. This means that whereas there was one immigrant for every two “indigenous” Baganda in 1948, there were three immigrants for every four “indigenous” Baganda in 1959.12 The largest single group of immigrants were from Rwanda. By 1990, many estimated that a quarter of the population in central districts in Buganda—such as Luwero, Mityana, Mpigi, and Mukono—was of Banyarwanda origin.

Migrant relations with indigenous Baganda (nationals) went through three different phases in the period between the heyday of migration after the 1920s and the NRA coming to power in 1986. The first phase closed with the anti-chief and anti-landlord peasant uprisings of 1945 and 1949 in Buganda. Even though half of the tenants in many of the central counties of Buganda were immigrants from Rwanda (and Burundi) by the end of the Second World War,13 social relations in rural Buganda did not pit migrants against residents, but landlords against tenants. The central trend in Buganda of the 1940s was to join all tenants, indigenous or immigrant, into a common anti-landlord movement called the Bataka Movement. The context for the second phase was set by the combination of colonial repression and counterreform that followed the peasant uprising after the Second World War. The point of colonial reform was to reinforce a landlord initiative set up as an “indigenous” electoral coalition in the name of the king, Kabaka Yekka (the King Only), to participate in the coming elections. It was a political initiative designed to draw a sharp political distinction between all those defined as indigenous, and those not. It was successful in dismantling the coalition between indigenous Baganda peasants and migrant Banyarwanda tenants and laborers. This trend was not reversed until the 1980s, when the NRA reconstituted the alliance of the indigenous and the immigrants, this time against government-installed chiefs and not local landlords, ushering a third phase in relations between migrants and nationals. Although migrants supported and joined the NRA, few joined the RPF in the October 1990 invasion of Rwanda.14 In contrast, refugees formed the backbone of the RPF.


The number of Tutsi who fled to Uganda between 1959 and 1964 were estimated at between 50,000 and 70,000. By 1990, they had swollen to around 200,000, although only about 82,000 had registered with UNHCR as refugees.15 The social destiny of refugees turned out to be markedly different from that of migrants. For unlike migrants who crossed the border to take up particular jobs, refugees were destined for camps in which they lived under a commandant, in circumstances that clearly divided them from the host society. Unlike migrants who were mainly Hutu, refugees were almost all Tutsi. The political division between the Hutu and the Tutsi inside Rwanda was reproduced outside Rwanda as a social difference between migrants and refugees. We shall see that the RPF was a force socially anchored in refugees more than in migrants.

The lot of the Rwandese refugees in Uganda was arguably the worst in the region. In Tanzania they could take up citizenship relatively easily, and the government actually demarcated a separate district for them where they could even have access to land for livelihood. In Zaire, they were at times offered citizenship, though—as we shall later see—the offer was withdrawn at other times. By contrast, successive Ugandan governments considered even the children of refugees to be refugees. The closest parallel I can think of to their predicament was that of Palestinian refugees in the Middle East: once a refugee, always a refugee. They were the butt of popular prejudice and official discrimination, the readily available explanation for any situation difficult to explain, from poverty to sabotage. Not surprisingly, a refugee self-consciousness developed first and foremost in response to anti-refugee prejudice promoted by the state and shared by many in the society at large. This is also why the mainly Tutsi refugees in Uganda came to think of themselves as Banyarwanda (Rwandese), and not as Tutsi.

The refugees who crossed the border into Uganda between 1959 and 1964 were settled in seven camps established in western Uganda, with an eighth set up only in 1982. In three of the camps, refugees were able to earn a livelihood by grazing herds on land. In the rest, they were forced to become cultivators. As their numbers multiplied, they overflowed camp boundaries. Those who had remained herders were among the first to leave the camps, moving their cattle to adjacent land. Thus, they entered both Lake Mburu National Park and adjacent ranches. Five square miles on the average, each of these ranches had been demarcated and allocated to bureaucrats in the 1960s as part of the USAID and World Bank–financed Ankole-Masaka Ranching Scheme.16 Both locations would become flashpoints of political conflict in the 1980s and would bring the Banyarwanda refugee question to the forefront of Ugandan politics, forming the immediate backdrop to the RPF invasion of October 1990.

Whereas the downside of a refugee status lay in its precarious political position, its upside was a consequence of UN recognition. Those recognized by the UN as refugees were entitled to UN aid the way the Ugandan poor were not. The eight refugee camps held mainly first-generation refugees. The only young people to be found in the camps were those poorly educated or those who had opted to stay with their parents and help them tend cattle or cultivate gardens. Most of the younger people had taken advantage of UNHCR scholarships, gone to schools and even to universities, and moved out of camps, mostly to towns like Fort Portal, Mbarara, and Kampala in southern Uganda, or to Nairobi in Kenya, or in some cases even to Europe or North America. The fact of a scholarship bounty that locals considered sure evidence of preferential treatment by UN agencies, and the success it facilitated, bred local resentment against successful refugees. The result was that even when successive generations left the camps, went to schools and some to universities, and came to form an elite educated stratum whose members could be found in the professions, business and the civil service, they could not escape being branded and set apart as refugees. Many had to bribe their way into institutions. Others had to hide their identity and pretend to be what they were not: Banyankole, Baganda, Banyoro. Socially, Banyarwanda refugees came to be distinguished from both Ugandan Banyarwanda (the nationals) and migrants by this single fact: many of their children were educated and successful. But even in the moment of their success, they could not escape the social stigma of being refugees.


It was the political and not the cultural diaspora, the community of refugees and exiles and not that of migrants and residents, that formed the first Banyarwanda organizations in Uganda.17 The refugee intelligentsia in Uganda set up the first political refugee organization in the region, the Rwandese Alliance for National Unity (RANU) in Kampala in 1980, openly discussing the question of a return of exiles to Rwanda. RANU was primarily an organization of intellectuals, given largely to debate, discussion, and representation. It was similar to a host of organizations set up in the Banyarwanda cultural diaspora, whether in Nairobi or North America. But in the fast developing situation in Uganda—where every minority defined as nonindigenous was fast becoming the target of state hostility and popular suspicion—RANU was soon to come under great stress.

Indigeneity became a central issue in postcolonial Uganda, as it had in Rwanda in 1959. At its heart, the question of indigeneity was a question of entitlement: Who was entitled to justice in the postcolonial period? Who was the postcolonial subject? The mainstream nationalist response to this question was to turn the colonial world upside down, but without fundamentally changing its terms of reference. If the terms of the colonial world set the indigenous apart from the nonindigenous in a racial sense, and then privileged the nonindigenous in a perverse way, the postcolonial response was to stand this world on its head, so as to privilege the indigenous against the nonindigenous. Even if they had run away from Rwanda, the refugees would find that they had not moved out of the postcolonial world where entitlement was tied to indigeneity.

The question of entitlement became acute at times of political crisis. It is during the political crisis of the late 1960s that the Obote I government began to draw a boundary between nationals and immigrants: to target the latter as trespassers so as to encourage the former to act as gatekeepers—and hope to become beneficiaries—in an independent Uganda. The more it faced opposition, the more the ruling party tried to paint the opposition as illegitimate, claiming that many refugees and migrants were entering the political process as nationals. Such was the rationale behind demands for tighter surveillance over refugees and immigrants. A series of measures followed, each designed to achieve precisely these objectives. To ensure the legal basis of all-around control of refugees, the government passed a bill called the Control of Alien Refugees Act. The legislation made the Rwandese a special class of residents subject to arbitrary questioning or even detention.18 Among the immigrants, the Rwandese were particularly vulnerable. Those who spoke Kinyarwanda could be found in all three categories: refugee, migrant, and national. This meant both that refugees and migrants could pose as nationals, and that nationals could be dismissed as refugees or migrants. When Obote ordered the removal of all un-skilled foreigners from public employment in 1969, thousands of Banyarwanda were among those affected. That same year, the government issued an invidious order: to conduct a census of all ethnic Banyarwanda. Catherine Watson claims that the objective was to exclude “both citizens and refugees from the political process and possibly even (to) expel(ling) them from Uganda.”19 Before the exercise could be completed, however, Obote was overthrown by his army commander, Idi Amin.

Not surprisingly, many Banyarwanda welcomed the overthrow of Obote—as did many members of the other embattled nonindigenous minority in Uganda, the Asians. Amin reciprocated, bringing the deposed Tutsi king, Mwami Kigeri, from Nairobi to Uganda and allocating a house and a car for his use. Many believed that Amin had promised to help reestablish the monarchy in Rwanda and, in that manner, ensure the return of refugees. Not for the first time in the history of an embattled minority, it sought refuge in a close relationship with an isolated power. A number of Banyarwanda refugees joined Amin’s army and secret service. Following Amin’s overthrow, there were claims that next to the southern Sudanese the Banyarwanda constituted the biggest number of aliens in his security forces, especially in the dreaded intelligence organization, the State Research Bureau.20

If several Banyarwanda refugees joined the Amin regime, there were also others, particularly students, who joined the anti-Amin forces. The most prominent of these was Fred Rwigyema, a teenage secondary-school boy recruited by Yoweri Museveni into the Front for National Salvation (FRONASA) in the mid-1970s. In 1990, Rwigyema would lead the RPF invasion of Rwanda. When FRONASA moved into Uganda behind Tanzanian forces in the anti-Amin war of 1979, Museveni began a mass recruitment that included Banyarwanda. In the post-Amin reorganization of the Ugandan state, FRONASA merged with Obote’s Kikosi Maalum (KM) to form the Uganda National Liberation Army (UNLA). Following the rigged election of 1980, which paved the way to the Obote II regime, there were defections from the UNLA as the guerrilla war against the Obote II regime gathered momentum. The Obote regime, in turn, baptized the “Museveni soldiers” as “Banyarwanda.” The more the repression of the Banyarwanda was stepped up, the more Banyarwanda soldiers joined Museveni and the NRA in the bush. The regime’s claim was fast turning into a reality.21

From the point of view of the Obote II regime, the Banyarwanda had sinned twice: not only had they joined the repressive apparatus of the discredited Amin regime, they were now providing the backbone of Museveni’s guerrillas in the ongoing bush war. This kind of perspective led to a massive state-organized repression of the Banyarwanda refugees in western Uganda. It all began as an exercise by Ugandan authorities in February 1982 to force Tutsi refugees to return to camps they were said to have left during the Amin period. When the attempt was unsuccessful, the ruling party’s regional leadership decided to organize what can only be described as a pogrom. The ensuing repression had a paramilitary and populist flavor about it. Organized by leading members in the Obote regime, it was implemented by chiefs in the local authority and paramilitary youth wingers on the ground, with even a unit of the paramilitary Special Forces joining in. Together, they targeted not only Banyarwanda refugees, but all Banyarwanda, whether nationals or refugees, and even some Bahima and Bakiga—in short, any and all who were ethnically identified with the opposition in the western part of the country. Houses claimed as Banyarwanda-occupied were looted and set aflame, or walls were pushed in and corrugated roofs were stolen. Occupants fled with their cattle and what they could carry. Local authorities tried to broaden popular support for the repression by redistributing confiscated land, cattle, and petty property to followers and promising to expand the circle of beneficiaries through further confiscation. The expulsion, Obote claimed in an interview, was in reality an uprising of Ugandans against Rwandans.22

Estimates from the period suggest that some 40,000 had crossed into Rwanda, driving some 25,000 cattle ahead of them. Thousands more preferred to cross into Tanzania. When Rwanda closed its border in November of 1982, another 4,000 were said to be trapped on the Ugandan side of the border, while almost 30,000 were under supervision in camps in Uganda.23 On a conservative estimate, 40,000 heads of cattle and 16,000 homes were destroyed.24 Rwanda stated that it recognized only 4,000 persons as entitled to Rwandan nationality and would settle them throughout the country. Uganda said it would take back no more than 1,000. This meant that 35,000 continued to languish in border camps, their future uncertain.25 In time, they divided into two groups: adults fled to camps, as many a youth headed for the growing army of NRA guerrillas in the bush.

In March 1983, Rwanda and Uganda came to an agreement under international pressure. Rwanda agreed to resettle more than 30,000, and Uganda agreed to set up an additional refugee settlement, Kyaka II, to relieve pressure from overcrowding in the old settlements. In addition, Uganda agreed to a joint screening exercise to determine the citizenship of those displaced and to consider compensation to those who had suffered loss.26 But there was no compensation. The perpetrators of the violence and the beneficiaries of the looting went unpunished. If anything, the arch of violence broadened. In December 1983, local chiefs and (UPC) youth wingers evicted over 19,000 Banyarwanda from Rakai and Masaka districts. With the Rwanda border closed, half fled into Tanzania, and the other half divided, once again the older ones heading for the security of camps, with the younger lured by the promise of guerrilla ranks. About the same time, “a similar but less calculated attack on the Banyarwanda was taking place” in Teso in the east and Lango in the north, grazing areas where Banyarwanda pastoralists had moved, either in search of grass for cattle, or in search of employment more suited to their skills.27

In July 1984, Uganda and Tanzania signed an agreement whereby Uganda agreed to take back 10,000 Banyarwanda refugees who had crossed the border a year earlier.28 In Uganda, however, the repression of refugees and attack on refugee settlements had become part of the official response to the spreading guerrilla war. In January 1985, the army entered Kyaka II settlement. As the harassment spread, the international staff left the camp and many Banyarwanda followed. It is unclear how many of the more than 30,000 that Rwanda had agreed to settle within its borders were actually so settled. For in mid-November 1985, after the Obote II regime was overthrown and the Lutwa regime was still in power, there was a repatriation, whereby 30,000 refugees expelled from Uganda in 1982 were returned, this time to NRA-controlled areas.29 One more time, the guerrillas harvested youthful recruits from the victimized refugee population. Two months later, when the victorious NRA entered the city of Kampala to take power in January 1986, roughly a quarter of their ranks of 16,000 were composed of Banyarwanda.


The counterpoint to the wave of refugee expulsions was the spreading guerrilla war in the Luwero Triangle. The guerrillas faced more than a military challenge. Their real challenge was political: to define an alternative to the politics of indigeneity. The issue was central both because the state claimed that the guerrillas were a nonindigenous force, a mainly Banyarwanda movement, and because of the very heterogeneous social composition of the Luwero Triangle: between a third and a half of the resident population in the Triangle were non-Baganda migrants. The largest group among the migrants were the Banyarwanda. Unlike the Banyarwanda in the refugee settlements, these were not refugees but residents who had immigrated at different times since the 1920s. As the guerrillas constructed an alternative political power in the areas they controlled, they needed to define the new political subject. Who was to be the bearer of rights and the beneficiary of entitlements under the new political arrangement? It was a question that could not be answered without coming to terms with the dominant mode of state politics in postcolonial Uganda: the politics of indigeneity. For, without a more inclusive alternative, the guerrillas would be faced with a population divided between those indigenous and those not.

The answer evolved over a period of time. It was arrived at through trial and error rather than implemented as a blueprint. As they came to grips with a string of concrete dilemmas, guerrilla leaders evolved an alternative. In contrast to the dominant mode of state politics which drew the line between those indigenous and those not, the NRA’s politics turned on the distinction between residents and nonresidents. Whereas the state used descent as the litmus test for defining the political subject, the test for the NRA was that of residence. The effect was to distinguish the political community from the cultural community and the future from the past, no matter what the overlap. If a cultural community was an outcome of historical processes, the political community was defined more with an eye on the future. From this point of view, the members of the political community were not necessarily those who shared a common history, whether defined by biological descent or a common culture. They were, rather, those who labored in the same community and intended to live together under a common political roof, and thus to forge a common future. The question of difference in their circumstances of origin was relegated to a secondary status.

The alternative was never theorized in the course of the guerrilla war, and was only partially theorized by the National Commission of Inquiry into Local Government, set up by the NRA after it came to power in 1986.30 But the question first arose in the course of establishing organs of local power in areas captured by the guerrillas. Called Resistance Committees, this institutional form really set apart the political community the guerrillas set out to build from the inheritance they intended to set aside. The establishment of the Resistance Committee was key to the reform of chiefship, the institution that symbolized unaccountable power, fused into a single fist by twentieth century indirect-rule colonialism and reproduced by the post-colonial state as “customary” and “traditional.” The thrust of the NRA reform was not to replace the agent of power but to reform power itself. It was not to replace one chief by another but to open the fist and to distinguish between different moments of power—the legislative, the executive, the judicial, the administrative—so as to assign each to a different authority. In this arrangement, the Resistance Council, the council of all adults sitting together in an assembly, was to be the holder of legislative power. The question then arose: Who was entitled to sit in the council, to deliberate in it, to vote in it, and to be elected to its executive organ, the Resistance Committee? Any resident adult? Or only the indigenous Baganda?

The truly radical side of the NRA’s response was to sublate this colonial inheritance by altering the line that distinguished the political subject from the nonsubject. In doing so, it distinguished politics from culture, and future from history, but without delinking the two. Without ignoring history, it refused to become a prisoner of history. Key to this enterprise to tap the creative potential of politics was the endeavor to redefine the political subject and the political community. By redefining the political subject as the resident, and by considering the historical fact of migration as politically superfluous, the reform moved away from the inherited world of the settler and the native in one single stroke. Yet, the reform was both partial and tentative. The decision to define rights on the basis of residence, and not ancestry, held at the lower levels, being the first two levels of village organization (Resistance Committees 1 and 2), as it did at the highest levels, that of the guerrilla leadership. But it did not hold at the middle level—the level of the district. There, the practice was to invest leadership in only those with an ancestral claim to the land. It was a temporary compromise, an unstable one, sure to be challenged whatever the bigger outcome, defeat or victory. The reform held for the duration of the guerrilla war, but not much longer.


The end of the guerrilla war was not an end to the struggle for power. Mindful that its organization was limited to less than half the country, the NRA gave priority to coalition building after victory; it termed this the creation of a “broad base.” Seeking to continue the struggle by other means, the opposition looked to identify the point at which the NRA would be most vulnerable to pressure. The issue it chose was that of entitlement. In demanding that indigenous Ugandans receive priority in the new political order, the opposition brought maximum pressure on the heterogeneity of the NRA, both at the leadership level and at the base. As individuals and factions within the leadership vied against one another for promotions, tensions multiplied at the top. At the same time, social struggles generated tensions at the popular level. The key struggle in this context was that between ranchers and squatters in the southwest of the country. Its larger significance flowed from the fact that while ranchers had been closely identified with every previous regime in the history of the country, squatters had been the source of many a recruit into the ranks of the guerrillas. Those threatened with the possibility of being left out in the cold when it came to jostling for positions and resources highlighted the distinguishing feature of many a squatter recruit into the ranks of the guerrillas: they were Banyarwanda, not indigenous.

Tensions within the NRA

From the very outset, the presence of Rwandese in the guerrilla war was important. Two of the twenty-seven persons who were said to have begun the guerrilla war with the ambush at Kabamba in 1981 were Banyarwanda: Fred Rwigyema and Paul Kagame. They had grown up together in the Kahunge refugee camp in western Uganda, meeting in 1963 and separating in 1972 when they joined different high schools.31 Both had been active in RANU. When the NRA entered Kampala in 1986, Rwigyema was its deputy commander. In 1987, he was appointed deputy minister of defense in Kampala. Paul Kagame became the acting chief of military intelligence of the NRA. Besides Rwigyema and Kagame, there were others, lesser known but still highly placed. Peter Baingana was head of NRA medical services. Chris Bunyenyezi was commander of the 306th Brigade. All in all, six senior officers were part of the senior command of RPF when it crossed into Rwanda: Lieutenant Colonel Wasswa, and Majors Kagame, Baingana, Kaka, Bunyenyezi, and Nduguta.32 And they were just the tip of the iceberg.

Had they joined the guerrilla struggle in Luwero, as some argue, to gather skills and weapons, and build an organization, so they could move to Kigali at the first available opportunity?33 Or was the move to Kigali a result of developments that took place after the NRA assumed power in Kampala, developments that would bring home to refugees a bitter truth: that in Africa today, once a refugee, always a refugee? To accept the first proposition is tantamount to assuming that there was a conspiracy to invade Kigali and take power, one that simply took time and effort to hatch. That would be to ignore the real debate that unfolded within the ranks of the refugees from the time they joined the NRA. No doubt refugee ranks included some who were convinced from the outset that there was no alternative but to return. Just as surely, these ranks could not have grown without others who had earlier championed naturalization in Uganda and in the countries of the larger region as the more viable alternative, and later changed their minds.

Initially, the NRA victory strengthened both tendencies. The electric example of a home-grown guerrilla movement that had defeated an internationally recognized government without substantial external support was not lost on RPF leaders. As one of them put it, “If the NRM could liberate Uganda, the RPF began to ask why it could not do the same in Rwanda.”34 At the same time, precisely because Banyarwanda refugees had played a vital role in the struggle, there was a radical and sympathetic shift in the official Ugandan position on naturalizing Banyarwanda refugees. All along, the Uganda government had insisted on very strict proof of descent as grounds for citizenship. In the Gabiro Accords of October 1982 with Rwanda, for example, Uganda authorities had insisted that no one could acquire citizenship without proof that the individual’s father, grandfather, and great-grandfather were born in Uganda.35 In July 1986, just six months after coming to power, President Museveni reversed policy radically: he announced that Banyarwanda who had been resident in Uganda for more than ten years would automatically be entitled to Uganda citizenship.36This meant that all the 1959 refugees would become Uganda citizens. In Rwanda, too, confident that the tide was turning irrevocably in favor of erstwhile refugees being granted citizenship in Uganda, the government was adopting a softer line on the return of Tutsi refugees: it agreed that individual refugees with resources to be self-supporting could return. If one is to go by the resolutions of the conference of Rwandese exiles that met in Washington, D.C., in 1988, the thrust of refugee opinion was also clearly in favor of naturalization in the country of their residence.37 In another few years, however, the balance of opinion would tilt decisively in the reverse direction. A senior RPA commander put it to me in Kigali in 1995: “You stake your life and at the end of the day you recognize that no amount of contribution can make you what you are not. You can’t buy it, not even with blood.” What had happened to leave such a bitter aftertaste?

Once guerrillas returned to society from the bush, they returned to a world of citizens and refugees, a world they thought they had forever left behind. Those who had lost the armed struggle, and thus power in the central state, lost no time in applying pressure on what they saw as an unholy power-hungry coalition of indigenous and nonindigenous factions. With hardly any delay, there were further expulsions of Banyarwanda. Though each was the result of a fresh local initiative, it is hard not to conclude that these did not involve some level of central coordination. The most notable of these was the forced eviction of Banyarwanda cattle keepers from Teso in 1986. Yet, the presence of Banyarwanda in the military leadership seemed to be increasing, rather than decreasing, in the first years of NRA rule. In the context of a simmering civil war in the north, their role in the NRA became even more important strategically. While the ratio of Banyarwanda in the army declined sharply—from 3,000 out of 14,000 in 1986 to roughly 4,000 out of a reported 80,000 to a 100,000 in 198938—Banyarwanda veterans formed a rapidly expanding core in the officer corps. The reason for this was simple: a disproportionate number of Banyarwanda had joined the struggle early on and thus had a greater battle experience than most others. The more fighting experience they gathered, the greater were the chances of their being commissioned. The opposition press began to scan the ranks of the NRA for evidence of officers who were Banyarwanda. The issue became politically charged in 1989 when two Banyarwanda officers, Majors Chris Bunyenyezi and Stephen Nduguta, were accused of having committed gross human rights violations during antiguerrilla operations in Teso. More than any other, one fact highlighted the centrality of Banyarwanda refugees in the leadership of the NRA: a Munyarwanda refugee, Major-General Fred Rwigyema, now occupied the position of deputy army commander-in-chief and deputy minister of defense.39 The only person senior to him in the military hierarchy was the president and minister of defense, Yoweri Museveni.

Soon, when it came to promotions, especially where a Munyarwanda was concerned, descent began to count more than anything else. The NRA tradition of giving preference to residence became distant memory. To Banyarwanda officers and men and women in the NRA, the period after 1986 seemed a betrayal by their former comrades-in-arms. One of the founders of RANU recalled the shifting mood: “The NRA experience was a catalyst in mobilizing the Banyarwanda in NRA. As far as 1983, our position was that people should join the struggle in Uganda voluntarily. It was worthwhile. It was not a deliberate effort to organize an army inside an army. The discrimination and harassment puzzled them, made them look for alternatives. They turned to senior RANU members, like Baingana. The discrimination did mobilize quite a few for us.”40

The RPF was born of this predicament in 1987.41 Faced with repression under Obote II, the old RANU had migrated to Nairobi in 1981. Once the NRA took power in Kampala in 1986, the Ugandan capital became the focal point of the political diaspora. RANU held its seventh congress in Kampala in December 1987 and rechristened itself the Rwanda Patriotic Front (RPF). Even though it had a Hutu as its head—ironically, just as had the monarchist UNAR in 1959 and as if to symbolize the nominal character of his presidency—the leadership of the RPF was predominantly Tutsi. More than that, it was predominantly from the “Ugandan” Tutsi, particularly from those who had been part of the NRA. The RPF crystallized two points of consensus: that the leadership of the refugee struggle would come from Banyarwanda in the NRA, and that the return home could only be armed. Unlike the left-leaning ideological tendency that had been RANU, RPF was to be a broad front. Though socially anchored in the broader refugee population and driven by an activist impulse, there was still no agreement within RPF on the future of the refugees—return or naturalization. The issue was at the center of discussions in the 1988 conference of the Banyarwanda political diaspora in Washington, D.C. The outcome of that conference showed that the weight of the opinion in the exile community was still on the side of naturalization in the countries of their residence for those who so wished, and return through peaceful negotiations for the others.

Then came the 1990 debate on ranches. Conducted during a three-day special session of parliament and chaired by none other than President Museveni, the debate followed a squatter uprising in southwestern Uganda. Its consequence, though, went beyond the conflict between squatters and ranch lords to once again placing the question of indigeneity center stage in the politics of Uganda.

The Trigger: Ranchers versus Squatters

The month-long squatter uprising of August to September 1990 took place in Mawogola County in Sembabule subdistrict, Masaka district, in southern Uganda. Classified as public land and grazed by pastoralists through the colonial period, these grasslands were divided and distributed to prospective ranchers by postindependence governments.42

The beneficiaries of the process included not only entrepreneurs, but also a combination of businessmen and bureaucrats, politicians and military men. Thus came into being a class that the 1988 Commission of Inquiry into Government Ranching Schemes referred to as “telephone ranchers.” The Mawogola ranches had all been allocated in the latter phase when allocation had turned into lucrative political patronage.43 An anomaly in the arrangement was that while cultivators on the land were guaranteed compensation, pastoralists were not pledged similar treatment since they were not deemed to own property that justified compensation. Deemed squatters, the pastoralists were a mixed lot who had been drawn to these grasslands at different historical periods and in different circumstances. Three different groups traced their arrival in the county to three different periods: the 1920s, 1979–80, and 1983–84. The first were the Bahima settlers who had come to the area as early as the 1920s. Then came migrants who had been displaced from areas bordering northern Tanzania during the 1979 anti-Amin war. Finally, there were mainly Banyarwanda cattle keepers, evicted from their homes and pastures around Lake Mburo in 1982–83, accused of being active supporters of guerrillas.44 Of the estimated 200,000 squatters, refugees were said to constitute around 80,000.45

The combination of absentee proprietors and propertyless squatters made for a flourishing of rental relations. The more squatters moved to idle ranch land to graze their cattle, the more ranchers were able to press home the legal fact of ownership and increase rent. As ranchers upped their demand, rent went from one cow per year to one for every fifty grazed, in addition to labor for digging dams, constructing dips, or clearing thick bush. Rents generally doubled between 1980 and 1984. When a serious drought hit Masaka District in the mid-1980s, a new exploitative relationship—water lordism—was added to this suffocating land lordism: ranchers began to demand rent for water, not just pastures. It was this latter development that broke the proverbial back of the pastoral camel.46

Squatters, whether Ugandan or Rwandese, had been an important source of guerrilla recruits for the NRA. When the NRA came to power in January 1986, some of these recruits retired and returned to the pastoral community, bringing to it a capacity to organize, a familiarity with arms, and a sense of self-assertiveness. Others remained within guerrilla ranks, constituting important allies of squatters. That the squatters constituted an important social base for the NRA was clear from one single fact: one of the ten points in the National Resistance Movement (NRM) program explicitly referred to the need to settle displaced peoples. Within the first year of assuming power, the NRA government abolished rent in all ranches, without being specific about the distinction between pasture and water rent. Acknowledging the right of the state to regulate the use of pastures since it was the owner of land, ranchers argued they had the right to charge for water from the dams they had built in the first place. Thus, ranchers suspended pasture rent and hiked up water rent.

In response, the government appointed a commission that recommended that over a quarter (57) of ranches be repossessed by government and reallocated to squatters, and that close to another quarter (50) be repossessed if they fail “to demonstrate (their) ability to develop” in a year.47 The commission’s recommendations were publicly seen as no less than a government-inspired agenda to dispossess swiftly half of the rancher class and empower squatters. Not surprisingly, these recommendations turned out to be a curtain raiser for a fuller conflict between ranchers and squatters, into which were drawn both the political and military leadership of the state. The expectation that they were about to be allocated land precipitated an influx of squatters from neighboring districts, from as far as the Tanzanian border to Masaka, concentrating numbers in the area.48 In response, the Masaka-based ranchers—being those under the most severe pressure—organized as the Masaka Livestock Farmers Association (MALIFA). They wondered why nineteen ranches were to be repossessed from Masaka but only three from Ankole. The implication was clear: Ankole ranchers must have been protected since they belonged to the president’s ethnic group. The wider implication was that the real shift in entitlement under the new order was not from those indigenous to all residents but to the president’s ethnic group and their Banyarwanda cousins across the border. MALIFA declared that it considered the government’s decision “null and void” and threatened to take the government to court, which precipitated a meeting of the National Executive Council of Parliament. The result was a government declaration upholding the policy to distribute ranch land to squatters.49 The Mawogola uprising broke out in less than two weeks.

Pushed in a corner, ranchers armed themselves, some bringing armed security guards into ranches. Before government could act to repossess and subdivide ranches among squatters, ranchers began a campaign of terror to get rid of squatters. The state response was immediate and united. Speaking in parliament, President Museveni condemned the killing of squatters as the work of organized terrorists who intended to precipitate a stampede from the county. An army representative alleged that leaflets warning squatters that the area was to become a war zone had been dropped for precisely this purpose. To guarantee their safety and convince the squatters to stay, government deployed a battalion in Mawogola County. The army forbade Resistance Councils, police, and the judiciary—all seen as heavily disposed in favor of ranchers—from intervening in the conflict.50

With local organs of the state neutralized, and with an entire battalion deployed in the county to keep ranchers in check, squatters counterattacked. Armed with guns, spears, and machetes, they stormed a number of ranches and seized them by force.51 The attacks did not simply target individual ranches, but showed a higher degree of coordination and organization: squatters mounted roadblocks closing off all routes by which ranchers could summon armed reinforcement. “The degree of determination, organisation and consistency exhibited by the squatters,” wrote Expedit Ddungu after talking to residents in the area, “created much speculation among people in the area that the uprising had been planned well in advance.”52

The more ranchers were pushed in a corner in the physical confrontation in Mawogola, the more they went on a political offensive. Critics claimed that squatter violence was instigated, and that “instigators” came from the NRA.53 They accused the government of a double favoritism, siding with both those belonging to the president’s ethnicity and their nonindigenous cousins from across the border, the Banyarwanda.54 In response, those favoring a pro-squatter legislation claimed that the tension in Parliament reflected “nothing but a class struggle,” and that the government’s proposal was an attempt to rectify a historically perpetrated “social injustice.”55

Ranchers claimed that governmental power—particularly military power—was bent on favoring nonindigenous Banyarwanda squatters. When an influential group in Parliament echoed the view, it made for the most heated moments of the parliamentary debate. The pro-rancher lobby in Parliament was joined by others determined that the nonindigenous be excluded from entitlement in the new era. Together, they concentrated their fire on Banyarwanda refugees as the core beneficiaries of ranch restructuring. Calling on the House to condemn “destructive or barbarous acts that took place in Mawogola county,” the women’s representative for Mbarara, Miria Matembe, claimed: “I have been talking to people, they are saying, Mr. Chairman, that these acts were committed by the NRM at the order of officers in this government and, Mr. Chairman, that your government is using this opportunity to bring in foreigners, pushing people in their land.” And then she added: “Rumors are saying that your Minister seated there, is behind the whole thing. And they are saying that this is an attempt to bring in foreigners, and you know, Mr. Chairman, the question of citizenship is a crucial matter in this country.” Muruli Mukasa, the representative for Nakasongola County in Luwero District, in the heartland of the Luwero Triangle where the NRA had fought the guerrilla war, wanted to know who would be the beneficiaries of ranch restructuring, the Rwandese or the Baruli of Nakasongola: “Mr. Chairman, the issue of citizenship has been highlighted. It has been really a very crucial one particularly in Nakasongola because most of the squatters there, at least, their citizenship is a bit ambiguous. Most of them happen to be from Rwanda and Burundi! So, people are saying okay, if there is this restructuring, it is fine but who are going to be the immediate beneficiaries; the Rwandese and Barundi!” The minister talked of “strong rumors,” mainly in Kyaka County, in Kabarole District—based on observations that “numbers within the refugee settlement have reduced from 30,000 to less than 10,000 now”—“that some refugees were undergoing military training.” The member for Mawogola County, Mr. Kasiaja, also alleged that squatters in Mawogola were being trained militarily.56 Later, many an observer would wonder whether these may indeed have been part of RPF ranks that crossed the border on 1 October 1990.

There had been wild allegations outside Parliament, repeated by some within, alleging that the purpose of the newly begun daily bus service from Kigali to Kampala was really to bring Rwandese to live in settlements in Uganda, something the minister of state for defense was compelled to deny in Parliament.57 Toward the end of the debate, Miria Matembe returned to press home the key point: “Mr. Chairman, we are not complaining of Ugandans to be given land; we are complaining of non-Ugandans—Banyarwanda—those who came as refugees to own land in this country. This is what we are concerned with.”58 She was joined by several others.59

Museveni dismissed these allegations as the work of “opportunists,” those “always around,” always “looking for unprincipled ways of gaining the support.” At the same time, he acknowledged that “they are heavily represented in all leadership of Uganda.”60Though the government denied every charge that it was being partial to nonindigenous Banyarwanda, the opposition had clearly succeeded in placing not only the question of ethnicity, but particularly that of indigeneity, at the center of the political debate. In the process, it had also managed to erode the political cohesion of the government. The clearest evidence of this was that the special three-day parliamentary session convened to discuss the question of the ranches turned into the most heated debate since the NRM had assumed power, pitting the president against his own minister of state for defense. David Tinyefuza, the minister of state for defense, supported squatter action unequivocally and consistently. “There was bound to be an explosion,” he argued. “These fellows can organise themselves. They are now more organised because with the fundamental changes in Uganda, the squatter is also the squatter of new mentality. The squatter is no longer the old squatter. He is a new squatter, enlightened, a bit more powerful even because now he has a representation even here which he never had. He is a bit more powerful; a bit armed here and there. (Laughter) He is a different squatter. So, this is the crux of the matter.”61

In a change of tack and direction, President Museveni pinned individual responsibility on his minister of state for defense: “I can give you information because I am also the Minister for Defence and I hold a number of other hats. What the rumours are saying is that Tinyefuza who comes from that area is the one who has been instigating these people…. But Tinyefuza I have heard, because there were three boys whom he put there, I have their names, they are army officers and I heard that they were involved in instigating squatters—these boys, the army officers, I ordered that they must leave that area when I heard about it.”62 In recognition of squatter links with many in the army, the battalion was withdrawn and all armed personnel forbidden from staying at the ranches. President Museveni confronted his defense minister: “Would you like to tell the House whether you were inciting the people to fight one another or what?” And then, in an obvious attempt to belittle Major-General Tinyefuza’s involvement as that of an individual and distinguish it from that of the state: “But Tinyefuza is not even a military authority in the area. Who made him a military authority there? I am not aware. All I am aware is that Tinyefuza is a Minister here in Kampala and he lives in that area. But in terms of command, he is not in charge of the area. So, how would he be a military authority?”63 The opposition claims that government policy on ranches had been turned into an instrument of refugee Banyarwanda interests—backed up by governmental, and particularly military, power—seemed to have succeeded in driving a wedge between the political and military leaderships.

Museveni dismissed those who had raised the Banyarwanda issue as “opportunists,” but the concessions he made on ranch restructuring had a lasting—and negative—impact on the question of citizenship. In repudiating the political legacy of the guerrilla war, it confirmed the colonial inheritance, yet again. Opening the second day of debate with the observation that while the country had dealt with the citizenship status of resident “Europeans,” he insisted that the question of citizenship “is not yet resolved, especially in relation to the Africans who came from neighbouring territories.”64 Though he began the day on a Pan-Africanist note, drawing a basic distinction between two kinds of foreigners65—those from outside Africa and those from within—he ended the day confirming the colonial legacy: redrawing the line between Ugandans and foreigners, not simply all foreigners, but particularly foreigners from neighboring areas—nonindigenous Africans. The real outcome of the debate was to strengthen the link between citizenship and indigeneity, and specifically once again to brand the Rwandese as non-Ugandan. Not only the original refugees, but also their offspring were now distinguished from the Ugandan Banyarwanda as nonindigenous.66 The president took the lead, even against the advice of his attorney general. When a member moved a clause—that “only Ugandan citizens shall be beneficiaries of these ranching schemes”—the attorney general objected that the clause would violate the spirit of Pan-Africanism.67 But President Museveni disagreed. The clarification, he said, was necessary not only “to undermine this campaign which has been going on but also to make a new beginning.” He insisted on a new beginning: “This is our first opportunity—we have never had any opportunity to say anything on this land. In 1900 it was the British and some chiefs; then later on it was all sorts of corrupt people. Now that we have got a chance to say something, why not to say it? … (Applause).”68 When another member asked what would happen to Ugandan squatters in the refugee settlements, President Museveni declared that he “would be in favour of the Ugandan cattle squatters in the settlement camps of the refugees be removed to come to these ranches—these lands” and further, that “the Banyarwanda who may be in the ranches be removed to the settlements.” That would leave only one question unresolved, added the president: “What do you do with the Banyarwanda who acquired properties outside the ranching scheme?” With no doubt in anyone’s mind that the government intended to turn its back on the legacy of the guerrilla war and that entitlements would yet again be linked to descent, the attorney general obliged and moved Clause 8, which he had just opposed: “Mr. Chairman, Clause 8, I beg to move that it reads as follows: ‘That only Ugandan citizens shall qualify to benefit from the restructured ranch schemes.’ ”69 For the Banyarwanda refugees, the die had been cast.

The debate triggered by the squatter uprising refueled the demand that the nonindigenous—particularly the Banyarwanda—be excluded from citizen entitlement in the postguerrilla political order. An earlier parliamentary decision had called for noncitizens to be identified and dismissed from the army. The shifting fortunes of Banyarwanda in the NRA were reflected in shifts in the fortune of the most respected of their members, Major General Fred Rwigyema. In the very first year of NRM power, he had been transferred from the powerful position of deputy commander of the army to the more ceremonial position of deputy minister of defense. In 1988, he was removed from even this ceremonial position by order of the president as chairman of the High Command.70 It was a decision that would have far-reaching consequences. With every passing year, the search for noncitizens in the army moved from the rank to the file, from the pinnacles of power to those below, literally turning into a witch-hunt, and was extended to other organs of the state. The consequence of the squatter uprising of August 1990 was to brand Banyarwanda cattle-herders as refugees, not citizens. For a brief period, this episode underlined the common predicament of two otherwise distinct refugee groups: the refugee guerrilla leadership which was being hounded out of the new state, and the refugee squatters who were ruled out as possible beneficiaries in the redistribution of ranching land possessed by the state.71 The immediate consequence of the squatter uprising was to swing the balance of opinion, among both refugee commoners and refugee leaders, decisively against naturalization in the countries of their residence and tilt it in favor of an armed return to Rwanda.

That armed return was the RPF invasion of October 1990. When he wrote to Human Rights Watch to explain the dynamic that led to the invasion, Uganda’s new ambassador to the United States, Katenta-Apuli, pinpointed “two fundamental decisions” as relevant: the prohibition against nonindigenous Banyarwanda owning land or holding state positions. “It is believed,” he concluded, “that the combination of these two fundamental decisions convinced Rwandese refugees that they did not have a bright future in Uganda and precipitated (both) the mass desertion from the NRA” and the decision to “invade Rwanda to regain their rights in their country of origin.”72 The logic behind the invasion was reinforced by the Uganda government, which declared on the day of the invasion that all Rwandese who had left the NRA to attack Rwanda would be considered deserters under the army’s Operational Code of Conduct. “That means,” the ambassador clarified, “on conviction by a court martial, they would be punishable by death. This is no incentive for them to cross back into Uganda.”

To most Ugandans, the ambassador’s claim would seem incredible in retrospect. After all, many in the RPF, particularly in the leadership, did cross back into Uganda at several points during the war, and the RPF enjoyed active support from the Ugandan state. After all, the Banyarwanda in the RPF were no strangers to Ugandan society or the Ugandan state. Some observers even thought of the Rwanda Patriotic Army (RPA) as functioning like an army within an army and the RPF like a state within the state. This point of view stressed that the RPA had already been organized inside the NRA as a separate command answerable to Rwigyema.73 It is this command structure that was said to have been activated at the time of the invasion, as part of the foreign-funded demobilization exercise within the NRA. The London-based Economist Intelligence Review quoted a British East Africa expert: “They demobilized by crossing the border in completely equipped units, taking their insignia off their shoulders as they crossed.”74 Years later, President Museveni told his fellow regional heads of state meeting in Harare that, while the Banyarwanda in the NRA had informed him in advance “of their intention to organise to regain their rights in Rwanda,” they had launched the invasion “without prior consultation.” Significantly, he continued, even though “faced with [a] fait accompali situation by our Rwandan brothers,” Uganda decided “to help the Rwandese Patriotic Front (RPF), materially, so that they are not defeated because that would have been detrimental to the Tutsi people of Rwanda and would not have been good for Uganda’s stability.”75 It was as candid an admission of complicity as any head of state could have made.

But a qualification needs to be added to this admission: the precondition for official Ugandan support was that there be no return. RPA officers may cross into Uganda as into a rear base, as many times as necessary, but there would be no return to Uganda, no possibility that any of them could define Uganda as home. When the RPF crossed the Uganda-Rwanda border in October 1990, this did not only constitute an armed invasion of Rwanda; it was also an armed repatriation of refugees from Uganda. The condition for Ugandan support, the bottom line, was that the RPF continue to push into Rwanda. The refugee soldiers who formed the core of the RPF—who had been nearly 4,000 of the roughly 14,000 NRA who took Kampala in 1986 and were probably another thousand-plus in 1990—found themselves between the Rwandan devil and the Ugandan deep blue sea. True, the raw material for the refugee crisis that led to the 1990 invasion was the outcome of post-1959 developments in Rwanda, but the crisis itself was very much Ugandan in the making. The dynamic that led to the 1990 invasion was born of the first crisis of the NRA in power. It was a crisis that split its guerrilla leadership and cadre along lines of indigeneity. But it was also a crisis that the NRA leadership, both those who stayed within Uganda and those who crossed the border into Rwanda with the RPF, tried to turn into an opportunity. It was a gamble whose cost would be difficult to tell, even with hindsight.

If you find an error please notify us in the comments. Thank you!