The Roots of the Great Lakes Conflicts: Forty Years of the Same War

Forty years of wars, massacres and exile have made the Great Lakes region one of the most enduring centres of violence in the world. Behind every episode, the same causes return: land that is too scarce, contested citizenship, poorly drawn borders, weakened states, coveted wealth and unpunished crimes. As long as they are not addressed, each peace agreement merely postpones the next conflict.

The Great Lakes conflicts are often told as a series of separate crises: the Rwandan civil war, the Burundian civil war, the 1994 genocide, the two Congo wars, then the CNDP and M23 rebellions. This episode-by-episode reading hides what matters most. Each crisis inherited the causes of the one before, and none was closed by dealing with what had set it alight.

This analysis goes back to the 1980s, the decade in which tensions accumulated since colonisation met a brutal economic crisis and the exhaustion of the regimes in power. It identifies seven root causes, then shows how they became knotted together. It ends with what the Way of “We” proposes as a way out.

Part IA legacy that froze identities

The causes of the 1980s have deeper roots. In Rwanda, the Belgian administration introduced ethnic labels on identity cards in the early 1930s: Hutu, Tutsi or Twa. Fluid social categories became fixed memberships, passed from father to child. The colonisers added a narrative claiming that the Tutsi had come from elsewhere, a narrative that hate propaganda would take up decades later.

In neighbouring Kivu, from 1937 the colonial authorities organised the resettlement of families from Rwanda to the territory of Masisi, to supply labour to the plantations. These arrivals joined Kinyarwanda-speaking populations present in the region long before the borders were drawn. Migration then continued spontaneously, driven by famine and political violence in Rwanda. By the early 1990s, according to the FAO, people of Rwandan origin made up at least 70% of the inhabitants of Masisi.

This dual legacy, frozen identities on one side and populations settled on both sides of a recent border on the other, is not in itself a cause of war. It becomes one when land grows scarce, when citizenship is called into question, and when political leaders turn origin into an instrument of power.

Part IILand: scarce and poorly protected

The Great Lakes region is among the most densely populated rural areas in Africa. On the cultivated land of Masisi, density approaches 300 inhabitants per square kilometre, according to the FAO. For families who live from farming and herding, every plot lost is a matter of survival.

Yet in Zaire, land law weakened those who worked the land. The laws of 1966 and 1973 proclaimed state ownership of the soil and stripped customary rights of their legal value. Traditional chiefs kept the power to allocate land in practice, but elites with connections in the administration could obtain official titles to land occupied for generations. Land thus became the object of a double competition: between communities, and between peasants and the powerful.

An analysis published in Forced Migration Review sums up the mechanism: when Banyarwanda buy land, they are accused of seeking to “infiltrate” local structures. And every return of refugees reopens the question of the land abandoned in flight. In the author's view, successive peace agreements have left these root causes aside.

Part IIICitizenship, granted then withdrawn

The nationality of Kivu's Kinyarwanda-speaking populations is one of the most direct causes of the violence to come. In January 1972, Law 72-002 collectively granted Zairian nationality to people from Ruanda-Urundi settled in Kivu before 1 January 1950. In 1981, a new law annulled this provision and imposed an individual naturalisation procedure. In May 1982, an ordinance declared the nationality certificates issued in 1972 null and void.

Legally, overnight, hundreds of thousands of people born in Zaire became suspect citizens. The 1981 law was little enforced at first, but it planted an idea that would never leave the politics of the east: some Congolese are supposedly less Congolese than others. The distinction between “indigenous” and “non-indigenous” became an electoral argument and, soon, an armed one.

The democratisation of the early 1990s sharpened the tension instead of easing it. During the Sovereign National Conference (1990-1992), the nationality of Kinyarwanda speakers, and therefore their right to vote, returned to the heart of political debate. In March 1993, massacres broke out in Masisi. Within a few months, violence by militias on both sides had killed between 10,000 and 14,000 people and displaced more than 200,000, according to the FAO. By May 1996, local human rights groups put the toll at 70,000 dead and 250,000 displaced. In 1995, the transitional parliament went as far as ordering that all people of Rwandan or Burundian descent, the Banyamulenge included, be sent back to their countries of “origin”.

Part IVPoorly drawn borders

The DRC shares its borders with nine countries: the Republic of Congo, the Central African Republic, South Sudan, Uganda, Rwanda, Burundi, Tanzania, Zambia and Angola. It is party to 22 border treaties, all predating its independence. These lines were set between 1885 and 1915 by agreements between colonial powers: with Portugal in 1885 for the mouth of the river, with Britain in 1894 for the Zambian border, with Germany by the convention of 11 August 1910 for Kivu, Lake Kivu and the Ruzizi, and with Britain in 1915 for Lake Albert. They were often drawn on imprecise maps, across lakes, rivers and volcanoes the negotiators knew poorly.

In 1964, in Cairo, the Organisation of African Unity proclaimed the inviolability of borders inherited from colonisation. Yet half a century later, less than a third of Africa's borders are precisely delimited: states committed to respecting lines that no one had marked out. The African Union Border Programme, launched in 2007, has since encouraged their delimitation and demarcation, but implementation remains slow.

The DRC has experienced this along several of its borders:

• Uganda. The Lake Albert boundary was set by an Anglo-Belgian agreement in 1915, when the region's geography was poorly understood. When oil was discovered there, Rukwanzi island became a flashpoint: in 2007, clashes between the two armies caused deaths on both sides, and a British oil company employee was killed. Further south, the Semliki River, which serves as the border, shifts its course and moves the boundary between the two countries with it.

• Zambia. An 1894 treaty between the British and King Leopold II is read differently by each side. In March 2020, Zambian troops took up positions in several localities of Tanganyika, including Kalubamba and Moliro, before the SADC, to which Kinshasa had appealed, ruled in the DRC's favour. As recently as September 2026, a joint technical commission meeting in Pweto was working to demarcate 32 kilometres of border rivers and noted the destruction of historic boundary markers.

• Angola. The DRC has only a narrow coastline, at Muanda, squeezed between Angola and the Angolan enclave of Cabinda. The disagreement over the oil-rich continental shelf led to an agreement on a Common Interest Zone on 30 July 2007, a Congolese law on maritime spaces in 2009, and an implementation protocol in July 2023. On land, the occupation of villages in Kahemba territory from 2007 adds to it.

• Rwanda. The border rests on the 1910 convention and on 22 markers laid in 1911. In April 2015, a joint commission began relocating them by GPS around Goma. Above all, this border lies at the heart of mutual accusations of support for armed groups: it was to verify them that the Joint Verification Mechanism of the International Conference on the Great Lakes Region was launched in Goma on 14 September 2012.

The borders with Burundi, Tanzania, the Central African Republic and South Sudan are not subject to any major known boundary dispute; they are nonetheless crossed by armed groups, refugees and trafficking.

These disputes are not the main cause of the wars in the east. But they all follow the same mechanism: where the border is vague and the subsoil rich, each state is tempted to push its positions forward, and border populations pay the price of the uncertainty.

Part VThe 1980s: exhausted states

The 1980s were a decade of decay, and of war in Uganda. In Zaire, Marshal Mobutu's regime sank into corruption and bankruptcy: between independence in 1960 and the end of his rule in 1997, the country's gross domestic product fell by some 65%. The state no longer provided security, justice or regular pay for its soldiers. In the east, far from Kinshasa, it became one actor among many.

In Rwanda, the economy relied heavily on coffee. The collapse of coffee prices in the late 1980s sharply reduced the income of the state and the elites, forced budget cuts, and fuelled competition for power and foreign aid. The country also carried a question unresolved for thirty years: that of its refugees. According to the United Nations, some 480,000 Rwandans, most of them Tutsi, were living in exile in neighbouring countries by the end of the 1980s. President Habyarimana's regime refused their mass return, arguing that the country lacked land.

In Burundi, the cycle of violence was already old. The 1972 killings, which first targeted Tutsi and then, during the army's crackdown, overwhelmingly Hutu, left more than 100,000 dead according to estimates. In 1988, the violence in Ntega and Marangara, in the north of the country, killed between 5,000 and 50,000 people depending on the source. Fear of the other group shaped the army, the parties and family memory.

Neighbouring Uganda was at war. In February 1981, after the disputed elections of December 1980 that returned Milton Obote to power, Yoweri Museveni took to the bush; in June 1981, his guerrilla merged with that of former president Yusuf Lule to form the National Resistance Movement (NRM) and its armed wing, the NRA. In an analysis published in 1986 in the journal Politique africaine, the historian Gérard Prunier described a government army that was “practically unpaid and undisciplined”, whose counter-insurgency turned in 1982-1983 into “the outright massacre of civilian populations”. In the “Luwero triangle” alone, north of Kampala, he estimated that there had “no doubt [been] more than 200,000 deaths in four years”. For the war as a whole, estimates range from 100,000 to 500,000 dead.

Rwandan refugees living in Uganda were both victims and actors of this war. In 1982, and above all in October in Ankole, they were driven from their land together with Ugandan Banyarwanda: some 40,000 people were pushed back into Rwanda, which took in only some of them. Thousands remained stateless along the border for years, and many joined the rebellion. Among Museveni's first fighters were young Rwandan refugees, including Fred Rwigyema and Paul Kagame. When the NRA took Kampala in late January 1986, several of them became officers in the new Ugandan army: Rwigyema was appointed deputy minister of defence, and Kagame acting head of military intelligence.

Prunier saw in the NRM's victory “an extremely original experiment”: having fought a prolonged armed struggle “not against colonial troops but against an African government” was, he wrote, “a radical novelty on the continent”. He then asked whether this was “a unique phenomenon, specific to the evolution of Ugandan politics”, or whether this change “heralds others on the regional political stage”. What followed answered him. For the first time, thousands of Rwandan refugees had military experience, allies in power and an organisation. In December 1987, in Kampala, their association renamed itself the Rwandan Patriotic Front. (Quotations from Prunier are our translation from the French.)

Weakened states, a refugee population with no prospect of return but now battle-hardened, an economy in crisis: by the end of the 1980s, every ingredient of a regional conflagration was in place.

Part VIThe regional spiral (1990-1994)

On 1 October 1990, the Rwandan Patriotic Front (RPF), made up largely of the children of Tutsi refugees who had served in the Ugandan army, entered Rwanda from Uganda. The civil war that followed was suspended in August 1993 by the Arusha Accords, which provided for power-sharing and the merger of the two armies.

Two months later, on 21 October 1993, Burundian army officers assassinated Melchior Ndadaye, the first democratically elected Hutu president, a few months after his election in June. The violence that followed, massacres of Tutsi and then army reprisals against Hutu, killed between 50,000 and 100,000 people within a year and drove many Burundians into exile. The Burundian civil war would last until 2005 and claim some 300,000 lives; the Arusha Agreement for Burundi, signed in August 2000, would set the framework for ending it.

On 6 April 1994, the plane carrying the presidents of Rwanda and Burundi was shot down over Kigali. Within hours, the genocide against the Tutsi began. In about a hundred days, until the RPF's military victory in early July, estimates range from 500,000 to more than one million dead, the vast majority Tutsi, as well as Hutu who opposed the killings; the United Nations uses the figure of more than one million. According to the United Nations, 150,000 to 250,000 women were raped.

The defeat of the government army then triggered an exodus on an unprecedented scale: some 1.4 million people crossed into Zaire, according to the United Nations. Among them were civilians, but also soldiers of the former Rwandan army and Interahamwe militiamen who had carried out the genocide. They settled in the Kivu camps, rearmed and launched raids into Rwanda: in the first half of 1996, these attacks killed around a hundred people a month. The genocidal forces were thus regrouping in the heart of Kivu.

Part VIIFrom the camps to the Congo wars (1996-2003)

In October 1996, a rebel coalition, the Alliance of Democratic Forces for the Liberation of Congo (AFDL), led by Laurent-Désiré Kabila and backed mainly by Rwanda, Uganda and Burundi, and later by Angola, attacked the camps and then marched on Kinshasa. Mobutu fled the capital on 16 May 1997. The UN Mapping Report documents massacres of Rwandan Hutu refugees and Congolese civilians by the AFDL and the Rwandan army during this campaign, and raises the question of whether some of these crimes could be classified as genocide. The Rwandan government strongly contested these findings. The first Congo war settled the question of the camps by force, but none of the root causes: not land, not nationality, not border security.

In August 1998, the break between Kabila and his former allies opened the second Congo war. Nine African states took part, for or against Kinshasa, and several rebellions divided the east and north of the country between them. The Lusaka (July 1999), Sun City (April 2002) and Pretoria (July 2002) agreements led to a transitional government in 2003.

The human toll remains disputed. The International Rescue Committee estimated excess mortality at 5.4 million people between 1998 and 2007, mostly from hunger and disease. Other researchers, such as the Human Security Report Project, challenge the method: for 2001-2007, they reduce the IRC's estimate of 2.8 million to around 860,000. Either way, it ranks among the deadliest wars in contemporary Africa.

Part VIIIWealth that funds the war

Eastern Congo is one of the world's richest regions in minerals: gold, tin, tantalum (coltan), tungsten, and further south cobalt and copper. During the second war, these resources stopped being a mere stake and became the fuel of the conflict. In April 2001, a United Nations panel of experts documented the illegal exploitation of diamonds, cobalt, coltan and gold by armies and rebel groups, and recommended an immediate embargo on trade in these resources with Burundi, Rwanda and Uganda. It treated Zimbabwe, an ally of Kinshasa whose companies had obtained concessions, as a separate case, and noted that, unlike Rwanda and Uganda, Angola and Namibia, Kinshasa's other allies, had not profited economically from the conflict. Relations between Kinshasa and its Angolan ally have nonetheless been marked by territorial disputes. According to a declaration by Congolese members of parliament, from January 2007 Angolan forces occupied eleven villages in Kahemba territory, a diamond-rich area in the south-west of the DRC, driving out their inhabitants; Luanda also contests the line of the maritime border, off which lie major oil fields. Some Congolese see this as the price of Angola's military help during the war, though no official document establishes it.

This war economy outlived the official peace. When an armed group controls a mine, a toll or a road, it no longer needs to win the war to prosper: it only needs the war to go on. This is one reason for their proliferation. In April 2023, the Disarmament, Demobilisation, Community Recovery and Stabilisation Programme (P-DDRCS) recorded 266 armed groups in South Kivu, North Kivu, Ituri, Tanganyika and Maniema: 252 local groups and 14 foreign ones. Many are local militias that present themselves as self-defence forces, and that also live off taxing the population.

Part IXTwo armed groups from elsewhere: the FDLR and the ADF

Among the foreign armed groups recorded in eastern DRC, two have been entrenched for more than twenty-five years and weigh on the whole region: the Democratic Forces for the Liberation of Rwanda (FDLR) and the Allied Democratic Forces (ADF), of Ugandan origin. Their tolls, established by the United Nations and human rights organisations, remain partial: in these remote areas, many victims are never recorded.

The FDLR since 2000

The FDLR was officially founded in Lubumbashi in 2000, from the forces that had fled Rwanda in 1994 and whose ranks included perpetrators of the genocide. In the aftermath of 1994, these forces numbered, according to estimates, between 20,000 and 50,000 fighters; by 2014, the United Nations counted only about 1,500, in North and South Kivu.

Their presence has gone hand in hand with repeated violence against Congolese civilians. In July 2005, they killed some fifty inhabitants of the village of Mamba. In 2009, during the military operations launched against them, Human Rights Watch established that they deliberately killed at least 701 civilians between late January and September; the worst massacre, at Busurungi on 9 and 10 May 2009, left at least 96 dead. From 30 July to 2 August 2010, a coalition of the FDLR, the Mai-Mai Sheka militia and deserters from the Congolese army raped at least 387 people in 13 villages of Walikale territory, according to a United Nations investigation. In 2012, the group committed at least 282 murders of civilians.

In July 2012, the International Criminal Court issued an arrest warrant for Sylvestre Mudacumura, the FDLR's military commander, for war crimes committed in Kivu in 2009 and 2010; he was killed by the Congolese army in September 2019 without ever being tried. In October 2019, an attack at Kinigi, in northern Rwanda, killed 14 people; the Rwandan authorities attributed it to the FDLR, though their claims could not be independently verified.

The FDLR are not alone to blame for this violence. During the same 2009 operations, Human Rights Watch documented the deliberate killing of at least 732 civilians by the Congolese army and its allies, including 143 Rwandan Hutu refugees. And for twenty-five years the FDLR's presence has served as a justification for Rwandan interventions in Congo: its neutralisation heads the commitments of the 2025 Washington agreement.

The ADF since 1995

The ADF was born in Uganda in 1995 from a merger of several rebel movements, including the remnants of the National Army for the Liberation of Uganda (NALU). Its first major attacks struck western Uganda. In November 1996, the attack on Bwera and Mpondwe killed some fifty people and forced 25,000 to flee. On 8 June 1998, at the Kichwamba Technical Institute, 80 students were killed and 80 others abducted. Pushed back by the Ugandan army, the ADF moved most of its activities to the DRC in the early 2000s and settled in the forests of Beni territory.

From October 2014, massacres of extreme brutality, attributed mainly to the ADF, were committed there, often with machetes. Between October 2014 and August 2016, more than 700 civilians were killed in the Beni area alone; at Rwangoma, on 14 August 2016, at least 64 bodies were recovered. In 2019, the Islamic State recognised the group's pledge of allegiance, and it became part of its “Central Africa Province”.

The violence then intensified and spread to Ituri. According to MONUSCO, the ADF killed around 1,300 civilians in 2021. Amnesty International recorded more than 1,000 deaths, mostly civilians, in 2023, and at least 1,022 in 2025 in more than 200 attacks. In June 2023, in Uganda, an attack on a school in Mpondwe killed 42 people. In July 2025, more than 40 worshippers were killed in a church in Komanda, in Ituri; in September 2025, more than 60 people were massacred at Ntoyo. In all, the ADF has killed several thousand civilians since 2014.

These two groups illustrate the Knot of Destinies in its most tragic form: born of a war next door, they made eastern Congo their refuge, and their victims are above all Congolese farmers who have nothing to do with the conflicts that gave rise to them.

Part XImpunity, and peace through integration

On 1 October 2010, the UN Office of the High Commissioner for Human Rights published the Mapping Report. It documents 617 serious violent incidents committed in the DRC between March 1993 and June 2003. The alleged perpetrators belong to Congolese forces and armed groups, but also to the armies and armed groups of Rwanda, Uganda, Burundi and Angola. Sixteen years later, most of these crimes have still not been tried.

This impunity has taken a particular form: peace through integration. At the end of each rebellion, fighters are folded into the national army, often with their ranks and chains of command intact. The National Congress for the Defence of the People (CNDP), launched in 2006, was integrated into the Congolese forces in March 2009. Three years later, some of its former officers mutinied and founded the M23, which took Goma on 20 November 2012 before withdrawing in early December.

On 24 February 2013, eleven countries of the region signed the Peace, Security and Cooperation Framework in Addis Ababa, later joined by Kenya and Sudan. They committed in particular to stop supporting armed groups in neighbouring countries and to reform their security sectors. But without a binding mechanism or justice for past crimes, these commitments did not prevent the return of war.

Part XIThe current crisis: the same causes, on a larger scale

In late 2021, the M23 took up arms again in North Kivu. In 2022, the UN Group of Experts reported “solid evidence” of Rwandan army interventions on Congolese territory alongside the M23. The same report stated that Congolese officers had supplied weapons and ammunition to a coalition of armed groups. The Rwandan government declined at the time to comment on what it called an unpublished and unvalidated report.

On 27 January 2025, the M23 seized Goma, then Bukavu on 16 February. In its 2025 report, the Group of Experts concluded that Rwanda exercises “command and control” over the M23 and estimated that some 6,000 Rwandan soldiers were deployed in Kivu at one point in the conflict. Kigali rejects these findings: it says it is acting against the persistent threat of the FDLR, an armed group whose ranks include perpetrators of the 1994 genocide. UN experts have themselves documented support by Congolese officers to armed groups, including the FDLR.

Human Rights Watch documents serious crimes committed by all parties: the M23, the Congolese army and the Wazalendo militias fighting alongside it, and the ADF and CODECO in Ituri. In September 2025, the DRC counted 5.7 million internally displaced people, and 122,500 people had fled to neighbouring countries.

Diplomacy has intensified. On 27 June 2025, the DRC and Rwanda signed a peace agreement in Washington providing for the neutralisation of the FDLR, demanded by Kigali, the withdrawal of Rwandan troops and the lifting of their “defensive measures”, demanded by Kinshasa, as well as the return of refugees and a regional economic integration framework; it was complemented in December 2025 by the Washington Accords. In parallel, a negotiating framework between Kinshasa and the M23 was set up in Doha in November 2025. But in September 2026, according to Security Council Report, only two of the eight protocols planned in Doha had been signed, and fighting continues in both Kivus.

Each of the causes described above can be found in this crisis. Land and refugee returns; nationality, and the rhetoric that brands some Congolese as foreigners; a state absent from whole swathes of its territory; mines that fund the fighters; crimes never tried; and neighbours who see in each other's chaos either a threat or an opportunity.

Part XIIWhat these causes have in common

At first sight, these seven causes belong to different fields: land law, nationality law, the drawing of borders, the economy, security, justice. Yet they share the same logic: each rests on the idea that one party's gain requires the other's loss. My neighbour's plot is a plot I do not have. A citizen declared a foreigner is one voter fewer for the opponent. A mine held by an armed group is a mine lost to the state. One country's security is built against its neighbour's.

This is exactly what the Manifesto of the Way of “We” calls the Illusion of the “I”, transposed to the scale of communities and nations. The Great Lakes offer its costliest demonstration: forty years of strategies built on excluding the other have brought lasting security to none of the region's peoples.

Part XIIIWhat the Way of “We” proposes

The doctrine of We-ism starts from an observation that geography imposes: in the Great Lakes region, peoples are not separate blocs locked in a zero-sum game, but strands of the same rope. This is what the doctrine calls the Knot of Destinies. It draws three rules of action from it.

1. The indivisibility of peace

Peace can no longer be defined as the absence of war on one's own territory while the neighbouring territory burns. Displaced people, weapons and armed groups cross borders. Supporting an armed group in a neighbour's territory, or tolerating at home a group that threatens it, amounts to preparing one's own insecurity. This rule echoes the non-interference commitment made in Addis Ababa in 2013; it turns it into a conviction rather than a constraint.

2. Co-development or shared shipwreck

The resources of eastern Congo can no longer be spoils of war. The doctrine proposes replacing the wall with the network: infrastructure, energy and mineral supply chains designed together, traceable and profitable to the people who live on that land. A precedent exists. In 2014, Burundi, the DRC and Rwanda signed the convention establishing the Lake Kivu and Rusizi River Basin Authority (ABAKIR), tasked with coordinating the management of a resource that no one can exploit alone; its ratification is still under way. Europe did the same in 1951 with coal and steel, until then the fuel of its wars.

3. The Realism of Forgiveness

No lasting peace will be built on forgetting. The Realism of Forgiveness is not an amnesty: it replaces neither the investigations nor the trials demanded by the crimes documented in the Mapping Report and in recent reports. It concerns relations between communities, where criminal justice stops. It refuses to make resentment the sole fuel of future political choices.

In practice, these principles call for finally treating causes rather than symptoms. That means a clear and final nationality law, ending the suspicion that weighs on Congolese born in Congo. It also means land registries that protect farmers, whether indigenous or settled from elsewhere, and traceable mineral supply chains whose revenues serve local communities. It means justice for mass crimes, and local Preventive Mediation capable of defusing local conflicts before they become wars.

None of this is simple. But forty years of war have shown the price of the other path. In the Great Lakes more than anywhere, no one saves themselves alone.