Africa's World War: Congo, the Rwandan Genocide, and the Making of a Continental Catastrophe

by Gerard Prunier

  • The old ALIR I in North Kivu: made up of ex-FAR and Interahamwe, it was about four thousand strong. In May 2001 they reentered Rwanda in Gisenyi and Ruhengeri. For the first time the abacengezi did not get much support from the local population, which was tired of the war. This allowed for quick success for the RPA, which killed over two thousand of them and captured eighteen hundred. After this episode, ALIR I was largely a spent force,69 even resorting to attacks on Mayi Mayi groups in North Kivu to steal their weapons.

  • The new ALIR II operated in South Kivu out of Kinshasa-supported bases in Kasaï and northern Katanga. It had over ten thousand men, and although many of the officers were old génocidaires most of the combatants had been recruited after 1997. They were the ones who fought around Pepa, Moba, and Pweto in late 2000.

  • The even newer Forces Démocratiques de Libération du Rwanda (FDLR) had about three thousand men, based at Kamina in Katanga. Still untried in combat, they had been trained by the Zimbabweans and were a small, fully equipped conventional army.70

  The problem was that everybody was playing with loaded dice. Just as the RCD-G was playing Kigali’s game, thus Joseph Kabila, discreetly supported by Zimbabwe, dragged his feet in helping MONUC implement DDRRR because many of the “negative forces” targeted by the disarmament exercise were used behind the scenes to “put pressure” on Kigali.71 The domestic situation was not much more honest, with the ANR constantly harassing and arresting civil society members who thought that the “Republican Pact” signed in Gaborone actually guaranteed them the right of free speech.72 The Addis Ababa meeting in October, which was supposed to reinvigorate the process, was in fact a nonstarter: the facilitation was confused and short of money, there were only eighty delegates, who kept bickering at each other, and Mwenze Kongolo declared he would not even start without the full quorum of 330 delegates. The talks got bogged down in procedural squabbles and the Kinshasa delegation eventually left abruptly.73 But not before the increasingly dynamic Pretoria observer team was able to make a convincing offer of reconvening the conference at a later date in South Africa. On December 4 Facilitator Masire announced that a new meeting would soon be held in Sun City, South Africa.

  The South African breakthrough

  The talks reconvened in Sun City on February 26, 2002. What was going to turn into a big South African diplomatic victory was all the more interesting for coming on the heels of the latest attempt at European conventional diplomacy. Between January 21 and 24, French Foreign Minister Hubert Védrine and his British counterpart, Jack Straw, had traveled to Kinshasa, Kigali, and Kampala in an obsolete display of “great power diplomacy.” The common trip was supposed to show to the world that Paris and London had buried their past differences; the problem was that nobody on the ground was very much interested in the white men’s hectoring. The two gentlemen told Joseph Kabila that he must carry out DDRRR with his Hutu forces; they told Kagame that he should leave the Congo and Museveni that he should control his army. They were met in all three capitals with barely disguised irritation.74 It is worth comparing this half-hearted attempt with the uninspired but doggedly persistent South African efforts.75

  The conference started almost immediately on the wrong foot, with bickering about accreditation which nearly led to an early breakup. Then Kinshasa asked for the expulsion of Ugandan and Rwandese security officers, who were openly consorting with “their” rebel protégés. The RCD-G laughed it off, adding in a rare display of humor that it did not mind the presence of the Angolan and Zimbabwean security men on the other side. Rebels harassed civil society representatives from the occupied areas, accusing them (not without reason) of being pro-Kinshasa. Then there was further wrangling on the accreditation of “journalists,” many of whom were fake. The whole thing was markedly chaotic and, in a strange sort of a way, quite cheerful.76 There was a serious incident when DRC Human Rights Minister Ntumba Luaba declared, “The Congo would explode like Somalia if President Kabila were forced to step down after these talks,” and the RCD-G answered that the whole point of the conference was to get him to resign. Kikaya bin Karubi, the press and communication minister, a new regime heavyweight, replied, “The government is prepared to share power but in no way to step down.” When talk came around to the possibility of elections, there was a clear lack of interest from all the armed parties (Kinshasa included), while the civil society struggled to make room for that agenda. Veteran 1960s politician Antoine Gizenga said that the survivors of Congo’s first Parliament, who were the only ones to have ever been fairly elected, should select the members of the future Assembly. He was politely ignored. In the midst of all this hubbub news suddenly arrived that the RCD-G had stormed the small Lake Tanganyika harbor of Moliro.77 The Kinshasa delegation withdrew. Then it came back a week later and the RCD-G offered to give Moliro to MONUC if the FAC would evacuate the two outposts of Kayaya and Yayama in northern Katanga. Through all this, the South African mediators never lost hope and kept cajoling their charges, reasoning with them and at times discreetly blackmailing them.

  Then suddenly, two days before the talks were going to go into recess, Bemba’s representative, Olivier Kamitatu, signed a power-sharing agreement with Kinshasa. This was obviously done in agreement with Kampala, leaving Rwanda out in the cold. The RCD-G was incensed and Azarias Ruberwa declared, “The Congo will not be reunified and there will be no peace.”78 Mwenze Kongolo accused him of “not negotiating freely,” that is, of remaining hostage to Kigali’s decisions.79 Kigali desperately tried to put together an anti-Agreement coalition, but, apart from four inconsequential organizations,80 all it could come up with to join with RCD-G was Tshisekedi’s old UDPS. The veteran oppositionist agreed to create the Alliance pour la Sauvegarde du Dialogue Inter-Congolais (ASD) and accused Kabila of “high treason.”81 But the ASD soon discredited itself by its shrill tone and unrealistic proposals.82

  Meanwhile, the first dividends of the first real diplomatic breakthrough since 1998 started to appear when Equateur began to open up to travel and commerce with the government-held areas. Boat traffic returned to the Congo and Ubangi Rivers, and after a while air travel restarted on the Kinshasa-Lisala-Basankusu-Bumba-Kinshasa route.83 There was such palpable relief among the civilian population that the RCD-G found it very difficult to enforce its policy on noncooperation, even among its own troops. At 4 a.m. on May 14 part of the Kisangani RCD-G garrison mutinied against their officers and their Rwandese minders.84 The mutineers’ radio call to “throw the Rwandese out of the Congo” was immediately popular, but the uprising was poorly organized;85 by 8 a.m. loyalists were back in control of the radio and the public buildings and by 4:30 p.m. Antonov An-12 transports arrived from Goma, bringing 216 RPA and RCD-G troop reinforcements. Since the RCD-G unit under Félix Mumbere, who had spearheaded the uprising, was mostly made up of ex-FAZ, all ex-FAZ prisoners were shot, most of them at the bridge over the T-shopp River so that their bodies could be dumped into the water.

  Repression went on during the next four days, with particular attention to the Mangobo suburb, where the civilian population had demonstrated in support of the mutiny. The killings were organized by the notorious commander Gabriel Amisi86 a.k.a. “Tango Four,” who was at the time the RCD-G chief of logistics. By May 17 the RCD-G was firmly back in control of the city. But its forces there showed serious signs of weakening as men started to desert at night, stealing canoes and paddling downstream to the now peaceful MLC-held territory. For once MONUC departed from its careful neutrality and published a report squarely placing the blame for the bloodbath on the RCD-G and estimating the number of deaths at 150, later increased to 183.87 The RCD-G was incensed. It asked for MONUC head Amos Namanga Ngongi’s recall and later banned him from traveling to the areas it controlled. The official motive given for this was rather paradoxical, since Adolphe Onusumba accused the UN official of supporting the April 17 Sun City Peace Agreement, something he would be reasonably expected to do. On May 31 the RCD-G expelled from its zone the UN
Human Rights director Luc Hattenbruck and two days later added two MONUC functionaries to the expulsion order. This drew condemnation by the UN Security Council on June 5, and Secretary-General Annan appointed two special representatives to disentangle the situation, former Senegalese prime minister Mustafa Niasse and former Eritrean diplomat Haile Menkerios. These were shrewd appointments: Niasse was a tough and experienced politician and Menkerios had been President Issayas Afeworki’s “controller” for the AFDL in 1996–1997 and had an intimate knowledge of the Congolese power structure.

  Rwanda’s isolation had developed as predicted. Bizima Karaha implicitly acknowledged this on June 7, when he reiterated his demand for Amos Namanga Ngongi’s recall but declared at the same time that the RCD-G would “keep collaborating with MONUC.” Kigali became frantic and went back to its leitmotiv of terrible Interahamwe danger in the Congo, claiming that there were forty thousand génocidaires in the DRC. When challenged by MONUC’s own figure of around fifteen thousand, Patrick Mazimpaka replied with aplomb, “This concerns only the Kivus and North Katanga. MONUC does not have access to Kinshasa government military camps where these 40,000 are.”88 But on June 27 Belgian Socialist MP Dirk van der Maelen demanded sanctions against all nonsignatory parties to the conflict “to turn the April 17th Agreement from a first step into a comprehensive deal,” and the talks restarted in Pretoria on July 18 with Kigali in a difficult position. It was now getting increasingly bad press coverage while “Li’l Joseph” was stealing all the headlines.89 Rwandese diplomacy then operated a quick about-turn, and on July 30 the RCD-G signed the power-sharing agreement.90 At first most observers doubted the Agreement’s feasibility.

  1. With 45 days for the national dialogue, 76 days for disarmament, and 166 for general foreign troop withdrawal, the Agreement seemed to suffer from the same hurried optimism that had defeated Lusaka in 1999.

  2. Nobody agreed on “negative forces” strength, estimates varying from a high of 50,000 (Rwanda) to a low of 12,000 (MONUC).

  3. The war was no longer two-sided but multisided, and all the nonstate actors (the various Mayi Mayi groups, Interahamwe, the Burundian rebels, several ethnic militias such as those of the Ituri and Masunzu’s Banyamulenge)91 were absent from Pretoria.

  4. There was no hard figure for RPA deployment in the Congo, estimates varying from 20,000 to 40,000 men.

  5. Who would monitor? MONUC had barely 2,500 men on the ground and was not known for its efficiency.

  6. Finally, who would pay? The Economic Commission, probably the only serious leftover from the Lusaka process, had earlier computed a need of about $3.5 billion in fresh money for the DRC to restart its economy, with 50 percent coming from the donors and 50 percent from national sources.92 Neither of the two seemed either willing or able to mobilize the cash.

  And yet somehow this behemoth ultimately got the show on the road. Why did Pretoria succeed while Lusaka had failed so abysmally? There are several reasons, none of which is fully convincing but which together add up to a coherent array:

  • Everybody was tired, which was not the case in 1999. Hopes for a clear military victory on either side had melted away.

  • As mentioned earlier, several of the initial reasons for intervention had either vanished (this was the case for Angola)93 or failed (this was the case for Zimbabwe and Uganda).

  • Unintended domestic consequences of the war had begun to make some of the involved governments suffer (Uganda and especially Zimbabwe).

  • The war had degenerated into a confused melee, particularly in the east, where the foreign actors seemed to be losing control over increasingly unmanageable alliance shifts.

  • The UN had finally recuperated some marginal efficiency, particularly with the publication of its reports on the looting of national resources.

  • The United States had changed its diplomatic stance from partiality toward one camp to neutrality.

  • South African diplomacy, for all its shortcomings, was simply more pugnacious and better informed than that of the largely amorphous “international community.”

  • Last but not least, Laurent-Désiré Kabila was dead and replaced by a man who belonged to the contemporary world.

  This, as usual, left Rwanda as the odd man out. The country whose internal crisis had triggered the massive process of what had been for a while a nearly continentwide war remained in a special situation. Traumatized by the genocide, playing on lingering foreign guilt due to the international community’s former neglect, under massive internal pressure to reconcile a dictatorial minority government and a “guilty majority” population, steeped in Congolese affairs in a radical way that had no parallel among the other foreign actors, Rwanda remained faced with dangerous and visceral unfinished business.

  The bumpy road toward a transitional government

  When it became clear that some kind of peace was going to come out of the Pretoria process, all the parties started to scramble in various directions to take care of their special interests, hoping to fit them within the general framework before it was too late. On August 7 the World Bank announced that it would loan $454 million to the DRC through its International Development Association branch, thus starting the process of economic normalization.94 The same day South Africa decided to commit fifteen hundred men to the peace monitoring. The next day Zimbabwe declared that it would pull out its remaining troops.95 Museveni announced his umpteenth “complete withdrawal,” but for the first time he gave clear details about the forces involved.96 Then former génocidaire Gen. Augustin Bizimungu was suddenly arrested and sent to the ICTR in Rwanda.97 And on August 16 Museveni signed a complete withdrawal treaty, with no conditions attached. The fact that he chose to do so in Luanda rather than in Kinshasa was a last tribute to the realities of the conflict: that it was a war fought among foreigners on Congolese territory for reasons of their own.

  On August 28 Kinshasa’s new spokesman Kikaya bin Karubi announced the obvious: that the next step would be to turn the power-sharing agreement into a proper transitional government. Everybody was concerned about how Kigali would react to these developments, and these reactions were confused and ambiguous, as could be expected. Rwanda was constrained by trying to follow two contradictory policies: the first one was to satisfy the international community, without whose financial support it could not keep functioning; the second was to try to preserve its stake in the control of the eastern Congo, which had become a domestic necessity. These two conflicting aims dictated a policy of stealth and proxy operations; DRC Minister-Delegate for Defense Awan Irung soon announced that the RPA had dispatched about a thousand troops to Kasuo, thirty kilometers from Lubero, with the purpose of blocking ex-FAR surrenders.98 By then the relationship of the Kigali regime with the former génocidaires had become increasingly complicated. While in public the RPF government kept selling the position of concern for its security, which had been its official line since the beginning of the conflict in August 1998, on the ground the situation by then encompassed everything from continued fighting to outright cooperation, with all the shades of relations in between (toleration, alliance with some groups to attack others, support for “real” or “false” génocidaires used as a pretext for continued occupation, sharing of mineral resources with some ALIR units, and so on). Then there was the new hostility toward Uganda, which often led the RPA to behave as if the main enemy was not Kinshasa anymore but Kampala.99 A third dimension of Kigali’s policies was the forced repatriation to the eastern Congo of Congolese Tutsi who had been refugees in Rwanda, at times since 1994. Of the 31,923 Congolese Tutsi then staying at the Kiziba and Gihemba camps in northern Rwanda, about 9,500 were forcibly bussed to Masisi before protests from UNHCR and local missionaries stopped this dangerous exercise in October.100

  On September 16 the Americans organized a tripartite meeting between Presidents Bush, Kabila, and Kagame at the UN in New York to reiterate the U.S. resolve to support the peace process. President Kagame then announced a pullout of hi
s forces “within one week.”101 To everyone’s surprise, the forces really did pull out, though not within the week but within a month. On October 7, 2002, Kigali announced that it had completed its evacuation, bringing 20,941 men back to Rwanda out of a figure of 23,760 in the Congo. The 2,819 missing were said to be either sick or gone AWOL. Since the total figure had been announced only at the time of the withdrawal, there were immediate doubts about whether the pullout was indeed complete. All the more so because, as soon as the withdrawal had taken place, rumors of foul play began to develop.102 For example, only 590 Rwandese soldiers had been evacuated from Kindu, though there had been about 2,000 in the garrison there; between October 16 and 20 about 250 RPA troops reentered the Congo near Bukavu and immediately headed for Walungu. In addition, many RPA soldiers were said to have simply changed their uniform and “become RCD-G.”103 In a way this was almost unavoidable since Mayi Mayi forces, which had not been party to the Pretoria Agreement, were now swarming all-over the place, trying to take advantage of whatever pullout had actually been carried out by the Rwandese forces or their allies. Kindu fell to them but was later retaken by the RCD-G, which carried out a systematic massacre of the people who had “collaborated” with the Mayi Mayi.104 Mayi Mayi “General” David Padiri Bulenda took Shabunda on the heels of the retreating RCD-G on October 4 and other commanders occupied Fizi, Baraka, Minembwe, Mwenga, Ubwari, and Walungu. Kabalo and Kindu were still precariously held by the RCD-G, but they were surrounded by Mayi Mayi. Then, on October 13, after two days of intense fighting, Gen. Lukole Madoa Doa105 took Uvira and immediately announced his intention of marching on to take Bukavu. The whole of South Kivu looked like it was going to fall into the hands of pro-Kinshasa forces without Kinshasa actually taking part in the fighting. The Rwandese High Command then called an RCD-G general meeting in Kigali to deal with the situation. This seemed like a partial illustration of the remark made slightly earlier by seasoned DRC observer Pierre Bigras on his website: “Everything is now happening as if Uganda, Rwanda and Zimbabwe had agreed not to fight each other directly any more but to exercise control over the different parts of the Congolese territory through the agency of the Congolese parties they are sponsoring.”106