How the RPF Won

On 2 October 1990, on the second day of the war, the Rwandan Patriotic Front's commander was shot dead. Within a week two of the three remaining senior commanders were also dead. The force had crossed the border with light weapons and no secure rear base, against an army receiving French officers, French training and French equipment within days. By the end of October it had been pushed back into the mountains and was, by most external assessments, finished.

Three years and nine months later it held the country.

This chapter is about how, and the answer is not mainly tactical. It is about what happens when a political organisation is built by people who have no alternative, because the world has spent thirty years treating their statelessness as a settled administrative fact.

The generation that had nowhere to go back to

Chapter 24 ended with between 120,000 and 300,000 Tutsi driven out of Rwanda between 1959 and 1964, and more in 1973. What is less often stated is what happened to them afterwards: nothing. Successive Rwandan governments refused their return on the grounds that the country was full. Habyarimana repeated the position for two decades, in a metaphor he was fond of: Rwanda was a glass filled to the brim, into which not another drop could be poured. The international community accepted it. The UNHCR's own doctrine at the time held that refugee crises ended in one of three ways, repatriation, local integration or resettlement, and for the Rwandan caseload none of the three was ever seriously pursued.

So a population grew up in camps and in exile in Uganda, Burundi, Tanzania and Zaire: by the late 1980s perhaps half a million people, many of them born outside Rwanda and holding no citizenship anywhere. Statelessness is not an abstraction; it is a daily administrative condition. It decides whether you can own land, sit an examination, take a job, cross a district line. In Uganda the refugees were repeatedly the target of local political mobilisation, blamed by turns for the crimes of the Amin years and for the pressures on land in the south-west. In October 1982 the Obote government drove tens of thousands of Banyarwanda off their land in Ankole; party youth wings burned houses and seized cattle; Rwanda admitted a few thousand and closed the border; the rest were left stranded for months in a strip of no-man's-land between two states, each insisting they belonged to the other.¹

That is the formative fact. This was a generation for whom there was no ordinary life available and no legal route home, and which had been told, in effect, that its non-existence was permanent. The lesson its members drew is the one that runs under everything that follows: citizenship is not a sentiment but a protection, and nobody restores it for you.

The apprenticeship

Many of them made their own route. When Yoweri Museveni launched his bush war against the Obote government in February 1981, with a few dozen fighters and a handful of rifles, Rwandan exiles were in the original band, and thousands more joined as the war and the persecution of the Banyarwanda fed each other: for a young refugee man in Uganda after 1982, a disciplined rebel army was in some respects the safest available address. Fred Rwigyema rose to become deputy minister of defence and deputy army commander in the victorious Ugandan government; Paul Kagame became head of NRA military intelligence. By the National Resistance Army's capture of Kampala in January 1986, Rwandans made up a substantial fraction of the force, by Prunier's estimate some three thousand of perhaps fourteen thousand fighters.²

They had therefore acquired, at somebody else's expense, exactly what an insurgency requires: five years of successful guerrilla warfare, staff experience, an intelligence apparatus, and a working model of a politically organised army, one that fed itself without looting and treated the civilian population as the terrain that mattered. It is not incidental that the NRA's code of conduct reappears, almost clause for clause, in the force this chapter describes.

Meanwhile the political organisation formed in parallel. The Rwandese Alliance for National Unity, founded in 1979 in the aftermath of yet another round of anti-refugee violence, was renamed the Rwandan Patriotic Front at a congress in 1987. Its programme was not, on paper, ethnic: return of refugees, national unity, an end to the system of ethnic classification, an end to corruption, democracy. Its membership by then included Hutu opponents of the regime as well as Tutsi exiles, a fact the regime's propaganda would spend four years denying.

By 1989–90 Ugandan domestic politics had turned sharply against the Rwandan presence in the army: parliamentary restrictions on Banyarwanda land rights, agitation against refugee officers, and finally the removal of Rwigyema and Kagame from their posts. Museveni owed the refugees a debt and could not politically afford to keep paying it. The window closed, and the RPF moved.

October 1990: near-annihilation

The invasion of 1 October 1990 was, in its mechanics, a desertion. Over the preceding days, soldiers of Rwandan origin slipped away from NRA units across Uganda with their weapons, their uniforms and a quantity of vehicles, and converged on the border post at Kagitumba: somewhere between two and four thousand men, crossing into a country most of them had never seen.

It went wrong immediately. Rwigyema was killed on 2 October, on a hill a few kilometres inside the border. The official account is a stray bullet; a persistent alternative account, which Prunier records without endorsing, holds that he was shot in a quarrel over strategy with his own subordinate commanders. The circumstances have never been definitively established, and the deaths within days of two more senior commanders, Peter Bayingana and Chris Bunyenyezi, are themselves the subject of conflicting accounts.³ What is not disputed is the result: a force that had lost its leadership in its first week, advancing down open roads against an army that was being reinforced by the hour.

France, Belgium and Zaire deployed. Opération Noroît, whose four-year career chapter 25 examined from the Paris end, began within days. Zaire sent units of Mobutu's presidential guard, who looted and raped so energetically that Habyarimana asked them to leave within weeks. And in Kigali, on the night of 4–5 October, the government staged an attack on its own capital: hours of gunfire by its own troops, blamed on RPF infiltrators, and used as the pretext for arresting thousands of Tutsi and opposition figures (estimates run from eight to thirteen thousand), many of whom were held for months without charge. The pattern of the coming war was set in its first week: military response abroad, racial punishment at home.

Kagame was in the United States at the time, on a course at the Army Command and General Staff College at Fort Leavenworth. He flew back, took command of a broken force, and made the decision that saved it: rather than fight for ground it could not hold, the RPF withdrew north-west into the Virunga volcanoes, high, cold, forested terrain above 3,000 metres where a conventional army could not easily follow.

What followed over the winter of 1990–91 is the hinge of the campaign. In appalling conditions, with men dying of cold and hunger, the force was rebuilt: reorganised into a conventional guerrilla structure, retrained, resupplied through covert channels, and (critically) reinforced by a wave of recruits from the diaspora, including students and professionals who left universities and jobs abroad to join. A strict code of conduct was imposed, and enforced with executions, because the political strategy required that the population not experience the force as bandits.⁴

1991–1993: the war

The reconstituted force announced itself on 23 January 1991 with a raid on Ruhengeri, the regime's northern stronghold: it took the town in a morning, held it for a day, emptied the prison and withdrew. Among the freed prisoners was Théoneste Lizinde, once Habyarimana's own security chief, jailed after falling out with the president; he left with the RPF. The operation was designed as much for signal as for effect, and the signal was received: the force that had been destroyed in October was capable of taking a garrison town at will.

The pattern of the next two years was: RPF offensives that took territory in the north, followed by negotiations, followed by government-organised massacres of Tutsi civilians elsewhere in the country (Kibilira, the Bagogwe country, Bugesera), followed by further offensives. Through 1992 the RPF pushed into Byumba prefecture and held a widening strip of the north, behind a front the government army, for all its growth, could not close.

The February 1993 offensive deserves its own paragraph, because it cut both ways. Launched after the massacres in the north-west, it broke the government line in days, roughly doubled the territory under RPF control, and reached within about thirty kilometres of Kigali before halting, a decision taken under heavy diplomatic pressure and in the face of renewed French deployment. Militarily it settled the question of which army would win a fight to the finish. Politically it was expensive: something approaching a million people fled south ahead of the advance, roughly a seventh of the country, into miserable camps ringing the capital. The displaced became the extremist parties' most fertile recruiting ground, and the propagandists of Hutu Power were handed exactly the image their theory required, an invading Tutsi army driving Hutu from their land.⁵ The offensive that proved the RPF could take Kigali also fed the machinery that would make taking Kigali necessary.

The Arusha Accords, signed on 4 August 1993 after a year of negotiation, registered that military reality on paper. Their terms were strikingly favourable to the RPF, which out-negotiated a divided government delegation: a broad-based transitional government in which the presidency was reduced to a largely ceremonial office; the return of refugees as of right, ending the thirty-year fiction that the glass was full; and the integration of the two armies, with the RPF to provide forty per cent of the troops and fifty per cent of the officer corps, terms that implied demobilisation for tens of thousands of government soldiers with no other livelihood. The CDR, the racialist party to the president's right, was excluded from the transitional institutions at the RPF's insistence; whether exclusion hardened men who were already planning extermination, or merely declined to give them a veto over the peace, is still argued about, and this chronicle notes the argument without pretending it is settled.⁶ Enforcement of the whole edifice was assigned to UNAMIR, a UN force authorised in October 1993 with a mandate and a budget calibrated to a peace both parties were assumed to want. For the hard core around the president, as chapter 25 described, Arusha was not a settlement but a death warrant, and they answered it accordingly.

One provision matters most for what follows: the Accords stationed a 600-strong RPF battalion in Kigali to protect the Front's incoming politicians, quartered in the CND, the parliament building, under UN supervision. That battalion was in the capital when the killing began.

April–July 1994

When the genocide started, the RPF did three things simultaneously, and it is worth being clear that they were in tension with each other.

It broke out of Kigali: on the evening of 7 April, with the roadblocks up and the lists in use, the 600-man battalion fought its way out of the parliament compound and held ground in the city until the columns from the north reached it.

It drove south: a general offensive from the north, moving fast, aiming to take the country rather than to negotiate. The design was encirclement rather than frontal assault: down the eastern side of the country, then wheeling west and south, cutting the capital off and leaving the government's forces to choose between fighting an army and continuing the extermination. The RPF refused repeated calls for a ceasefire, on the stated ground that a ceasefire would freeze the lines while the killing continued behind them; whatever else may be said about that position, the killing was in fact continuing behind them. Kigali fell on 4 July; Gisenyi and the north-west by 17 July; a unilateral ceasefire was declared on 18 July.

And it ran rescue operations: detachments were sent to extract survivors from known concentrations of refugees, thousands of them sheltering at the Amahoro stadium under nominal UN protection, others moved through negotiated exchanges from the Hôtel des Mille Collines and other besieged sites across the lines, and columns of people were walked north through the front to safety. Tens of thousands of people survived because of these operations. They also slowed the main advance, which is a real strategic cost and one that some of the RPF's own commanders argued about at the time.⁷

The plain fact, which no serious account disputes: the genocide was stopped by the RPF. Not by the United Nations, whose Security Council voted on 21 April, two weeks into the killing, to cut its force from about 2,500 men to 270, and whose replacement force of 5,500, authorised in May, had barely begun to arrive when the war ended; not by the United States, which was avoiding the word; not by France, whose intervention came in the final month and in the area held by the perpetrators. It was stopped by a rebel army of stateless people, against a state with a larger army and a foreign patron.

One more position needs recording, because it is often misread. The RPF opposed Opération Turquoise, and said publicly that French troops entering its zone would be treated as hostile. Given chapter 25, the opposition requires no elaborate explanation: the state proposing a humanitarian intervention in June was the patron of the army being defeated, and the "safe humanitarian zone" it established did in fact become the corridor through which the genocidal government, army and militia moved intact into Zaire. That outcome, and what grew from it, is chapter 27's subject.

Why it won

• Cohesion born of having no alternative. There was no negotiated future in which these people got to go home as ordinary citizens. Every previous settlement had left them outside. That produces a kind of persistence that better-supplied armies rarely have.
• Professional competence. The Ugandan apprenticeship gave the RPF a staff, an intelligence service, and commanders who had won a war before. The Rwandan army, by contrast, tripled in size in three years and was full of hastily recruited, poorly paid, badly led men.
• Discipline as strategy. A force that did not loot could operate among a population that had been told it was coming to kill them.
• The enemy destroyed itself. The genocide diverted the government's forces, its transport, its command attention and its ammunition into murdering civilians rather than fighting an army. It also annihilated the regime's international standing at the moment it most needed support. The genocide was, among everything else it was, a catastrophic military decision.

What was inherited on 18 July 1994

A country with something approaching a million dead and unburied. Two million people outside its borders, including the entire apparatus of the previous state. No functioning currency, banking system, water supply, electricity grid, courts or police. Roughly a third of the population displaced internally. The national treasury physically removed. Hundreds of thousands of orphans. A judiciary of which very few members were still alive or in the country. And a population in which the perpetrators and the survivors would, within a few years, be living on the same hills.

What was built on that is chapter 28. What the defeated took with them across the border is chapter 27.

Footnotes

1. On the refugee condition and the 1982 expulsions, Mahmood Mamdani, When Victims Become Killers: Colonialism, Nativism, and the Genocide in Rwanda (Princeton: Princeton University Press, 2001), in the chapter on the citizenship crisis of the Banyarwanda in Uganda; Gérard Prunier, The Rwanda Crisis: History of a Genocide (New York: Columbia University Press, 1995), ch. 2. ↩
2. Prunier, The Rwanda Crisis, ch. 2, for the NRA years and the estimate of Banyarwanda strength at the fall of Kampala; Stephen Kinzer, A Thousand Hills: Rwanda's Rebirth and the Man Who Dreamed It (Hoboken: Wiley, 2008), on Rwigyema, Kagame and the Ugandan apprenticeship. ↩
3. The competing accounts of Rwigyema's death are set out in Prunier, The Rwanda Crisis, ch. 3, which reports the internal-quarrel version without endorsing it; see also Kinzer, A Thousand Hills. ↩
4. Prunier, The Rwanda Crisis, chs. 3–4; Colin M. Waugh, Paul Kagame and Rwanda (Jefferson: McFarland, 2004); Philip Gourevitch, We Wish to Inform You That Tomorrow We Will Be Killed with Our Families (New York: Farrar, Straus and Giroux, 1998). For a critical account of the RPF's internal politics, Filip Reyntjens, Political Governance in Post-Genocide Rwanda (Cambridge: Cambridge University Press, 2013). ↩
5. On the February 1993 offensive and the displacement it caused, Prunier, The Rwanda Crisis, ch. 5; Alison Des Forges, Leave None to Tell the Story: Genocide in Rwanda (New York: Human Rights Watch, 1999), on the massacres that preceded it and the radicalisation that followed. ↩
6. On the Arusha negotiations, their terms and the CDR question, André Guichaoua, From War to Genocide: Criminal Politics in Rwanda, 1990–1994 (Madison: University of Wisconsin Press, 2015); Prunier, The Rwanda Crisis, on the accords and the army-integration protocol. ↩
7. Roméo Dallaire, Shake Hands with the Devil (Toronto: Random House Canada, 2003), for the situation in Kigali, the Amahoro stadium and the exchanges across the lines; Des Forges, Leave None to Tell the Story, on the course of the RPF campaign. ↩
8. On the Gersony episode, Des Forges, Leave None to Tell the Story, pp. 726–731, which discusses both the findings and the circumstances of their suppression, and the same work's documentation of RPF killings during the campaign. ↩