Copyright© 1996 The Johns Hopkins University Press. Allrights reserved. This work may be used, with this header included, fornoncommercialpurposes within asubscribedinstitution.No copies of this work may be distributed electronically outside of thesubscribed institution, in whole or in part, without express writtenpermissionfrom theJHU Press. Thisrevolutionary publishing model dependson mutualtrust between user andpublisher.
SAIS Review 16.1 (1996) 221-240

The Other Lesson of Rwanda: Mediators Sometimes Do More Damage Than Good

Alan J. Kuperman

The 1994 genocide in Rwanda shocked and horrified the world. Inits wake, analysts have concentrated most of their attention on theinternational community's failure to take prompt action in order tostop the killing. This inaction has been attributed variously to theinefficiency of the United Nations decision-making process, the lackof an on-call international reaction force, and the pervasive impact ofracial and cultural biases. Proposed reforms have included establishingan on-call UN reaction force, relying more on individual states andad hoc military coalitions to enforce peace, and increasing theresources devoted to preventive diplomacy. More recently, attentionhas turned to the plight of the hundreds of thousands of refugees whoappear to have worn out their welcome in Zaire.

However, this overwhelming focus on events after the outbreak of violencerisks overlooking perhaps the most fundamental lesson of Rwanda--thatmediators sometimes inadvertently provoke the very tragedies they seekto prevent. Indeed, the events leading up to the tragedy in Rwandademonstrate that mediation remains an inexact science. While conflictmanagement gurus have in recent years emphasized the need for mediatorsto utilize leverage1--such as aid, trade, sanctions andembargoes--the Rwanda experience demonstrates that such leverage isa double-edged sword, equally capable of driving contending parties tothe most extreme measures. [End Page 221]

Ironically, the pre-genocide mediation in Rwanda which led up to the1993 Arusha accords was a virtual textbook case of modern conflictmanagement. The conflict was ripe for settlement because the civilwar had reached a mutually hurting stalemate.2 Mediatorsused leverage, coordinated their activities, and ultimately put asideindividual agendas. The settlement seemingly addressed all of thesocial, economic and political roots and consequences of the conflict,and the international community agreed to guarantee the settlement witha peacekeeping force. Unlike in Angola, the settlement did not undulyrush democratic elections, postponing them until the contending armiescould be integrated under an interim transition government.

But the mediators had a blind spot. They failed to appreciate how muchRwanda's entrenched elite had to lose under political pluralization,and the lengths to which it would go to preserve the status quo. Themediators' application of leverage succeeded in compelling Rwanda'sPresident to sign and begin to implement the Arusha accords, but this verysuccess raised the insecurity of Rwanda's elite to the breaking point. Toprotect the privileges they saw the international community trying towrest from them, extremists proved willing to massacre fellow countrymenwith whom they had been living in relative calm for more than two decades.

That the mediation not only failed to control the conflict but propelledit into such a massive escalation, despite ostensibly propitiousconditions, calls for a reevaluation of prevailing conflict managementtheory. It also underscores the need for potential mediators both toobtain more accurate intelligence and to exercise better judgment inutilizing that intelligence. Most importantly, it demonstrates thatthe international community must be more cautious in applying leverage,especially in cases where it is unwilling to provide enforcement measuresif events go awry.


Rwanda is a densely populated, central African country of just over 7million people (pre-genocide) squeezed into an area the size of Maryland.Its population is comprised mainly of three groups--the tiny minorityTwa, [End Page 222] the larger Tutsi and the overwhelming majorityHutu.3 Even prior to German colonization, the minority Tutsi(primarily cattle owners) dominated over the Hutu (mainly farmers) andthe aboriginal Twa. However, it was Belgium which, after taking controlfrom Germany after World War I, institutionalized Tutsi dominance andsolidified these divisions through the issuance of ethnic identity cardsin 1926.4

These group divisions, which function as quasi-castes, are notstrictly "ethnic" since inter-marriage has long been afact of life. Indeed, as early as Belgium's 1933-34 census, Tutsi wereidentified not by ethnic heritage but rather as those who owned ten ormore cattle. Even official ethnicity has remained somewhat flexible, withwealthy Rwandans able to use bribes to change the affiliation on theiridentity card. Nor is the Hutu group itself monolithic. Rather, a sharpdivide exists between the Hutu from northern Rwanda, and the masses fromthe central, southern and eastern portions of the country--a distinctionthat was to play an important role in the tragedy.5

In 1959, as Rwanda began to push for self-determination, the Hutu andTutsi fought a bloody battle for political control, leaving 10,000dead.6 The Hutu prevailed by the time of independence, in1962, through a movement known as Parmehutu (Party for Hutu Emancipation),based among the majority, non-northern Hutu.7 Hundredsof thousands of Tutsi fled to neighboring Uganda, Tanzania, Burundiand Zaire.

In the following years, these refugees became a festering regionalproblem, often discriminated against in their new host countries (withthe notable exception of Tanzania). Rwandan officials blocked the returnof most Tutsi, ostensibly on the grounds that the country already sufferedfrom overpopulation. Indeed, Rwanda reportedly has the highest populationdensity in black Africa8 and the highest birthrate (8.3 livebirths per woman) in the world.9 Nevertheless, the Rwandangovernment's rejection of the refugees is best viewed as an outgrowthof continuing ethnic hostility, highlighted by periodic massacres ofthousands of Rwandan Tutsi in 1963, 1967 and 1973. In 1973, Major GeneralJuvenal Habyarimana, a northern Hutu, seized power from the Parmehutu,putting an end to the anti-Tutsi pogroms, but continuing to oppose therefugees' return for nearly two decades. [End Page 223]

Outbreak of Civil War

Recent events in Rwanda can be traced to 1986, when the Ugandan rebelarmy--which included many Rwandan Tutsi refugees--overthrew thegovernment of Uganda. Inevitably, and especially in light of continuingdiscrimination within Uganda, the Rwandan contingent of the rebelsturned their sights on their homeland. Secretly forming the RwandanPatriotic Army (RPA) and a political counterpart, the Rwandan PatrioticFront (RPF), the rebels invaded in October 1990, displacing hundreds ofthousands of Rwandans internally and as refugees.

Third countries initially rushed to take sides. Uganda, despite publicclaims of opposition to the invasion, served as a refuge and arms supplierfor the RPA. Zaire, Belgium, and France openly provided military supportto the Rwandan government. Soon, however, Zaire and Belgium withdrew theirsupport, leaving France and Uganda as the chief third-party protagonists.

An initial Zairian mediation effort achieved a cease-fire in March1991, but Habyarimana immediately renounced it because of alleged RPFviolations. The Rwandan government had little incentive to compromise atthe time, since it had gained the upper hand in the war.10In the terminology of conflict management, the lack of a hurtingstalemate meant that the conflict was not ripe for settlement. Numerousnegotiations over the next year--orchestrated by France and Rwanda'sneighbors--failed to make meaningful progress.

Demographics and Political Dynamics

Before long, however, pressure again mounted on the ruling regime.First, international aid donors and trading partners began utilizing theirleverage to insist on reform. Then, in late 1991, the RPA resurged underthe leadership of Commander Paul Kagame.11 To make sense ofthe eventual genocide, it is necessary to examine how these pressuresimpacted three sectors of Rwanda's population.

Confronted with these dynamics, President Habyarimana initiated nominalsteps toward political pluralization. In 1991, he legalized multiplepolitical parties. Of the 16 parties eventually formed, five dominated:MRND (the President's party), MDR (Parmehutu), PSD (Social Democrat),PDC (Christian Democrat) and PL (Liberal), the latter three beingmulti-ethnic. Simultaneous with this ostensible pluralization, however,Habyarimana established two new institutions to empower extremists andentrench his authority, the CDR (a right-wing political party) and theInterhamwe militia (a paramilitary organization that was to play a majorrole in the genocide).

At the end of 1991, the President made another feeble show ofpluralization, adding a single opposition member to his cabinet. Then,in April 1992, Habyarimana announced a truly multi-party cabinet. Whilea step forward, it was by no means democratic since no elections wereheld, Habyarimana's party maintained an absolute majority in the cabinet(10 of 19 seats), and the President retained all real authority.

As the rebels continued to make progress, Habyarimana also agreed tothe first serious mediation efforts in Arusha, Tanzania, in July 1992.Led by the host nation, the mediation also included all key regionaland international interests: France, Belgium, the United States, Uganda,Zaire, Burundi, the United Nations, and the Organization of African Unity(OAU). The talks made quick progress, reaching a cease-fire in July(monitored by the OAU), an agreement on legal authority in August, and apower-sharing protocol two months later. In October, however, Habyarimanabacked away from the negotiating process, disavowing the commitmentsof his ministers and belittling the agreements as "scraps of paper."Rwanda's relatively powerless Prime Minister, a Parmehutu, denouncedthe President for speaking with "two tongues."14

After continued government backpedaling, the cease-fire broke down inFebruary 1993, and the RPA again made quick military advances-- doublingthe territory under its control, reaching within 25 kilometers of Kigaliand displacing internally up to a million Rwandans--once more [End Page 226] forcing Habyarimana to the peace table.15 Undera new cease-fire the following month, the RPA agreed to withdraw to itspre-February positions, with OAU peacekeepers installed as a bufferbetween the contending forces. At the request of Uganda and Rwanda,the United Nations also authorized a peacekeeping force (UNOMUR) tomonitor the border between the two nations. Finally, on August 4,1993, the contending parties reached a historic peace agreement atArusha, laying out a comprehensive roadmap for a transition to genuinemultiparty democracy.

Arusha: A Textbook Mediation

The theory of modern conflict management asserts that conflicts arenot ripe for settlement until all parties are suffering and havelost hope of prevailing through escalation--a situation known as amutually hurting stalemate. Mediators, therefore, are to be selectedfor their ability to bring about this ripe moment by wielding economicand diplomatic carrots and sticks--known as leverage. Mediators shouldnot have self-interested objectives other than resolving the conflict,since these hidden motivations could distract them from the primarymission. Finally, to be successful, a settlement should address not onlythe root causes of the conflict, but also complications arising from theconflict itself. The Arusha process embodied each of these principles:

Foot-Dragging, International Pressure and Genocide

Within a month after the Arusha accords, disturbing signals emerged.A member of the President's family established RTLM (Radio Mille Collines)and began broadcasting virulently anti-Tutsi propaganda and incitementto violence. The 37-day target came and went without any sign of atransitional government. RPA troops refused to enter the capital untilthe departure of French troops, who in turn refused to depart until thearrival of the UN force. New target dates for the transitional governmentwere repeatedly established and postponed. The President insisted that thetransitional government be expanded to include small parties, includingthe extremist CDR, and quibbled over lists of ministers, further delayingthe process. Political violence escalated, culminating in the Februaryassassinations of the heads of the PSD and CDR parties.

Only the most limited progress was made toward implementation of theArusha accords. UNAMIR forces finally arrived in late December 1993,enabling French troops to leave the country, which in turn permitted[End Page 229] RPF politicians to finally enter Kigali, a stepessential for establishing the transitional government. But progressthen halted.

In the spring of 1994, the international community again turned up theheat on Habyarimana, applying unprecedented diplomatic pressure to forcehim to fulfill the agreements (which their earlier pressure had compelledhim to sign in the first place). On April 3, a group of ambassadors fromkey western nations met with the President, insisting that he immediatelyinstall the transitional institutions. Three days later, regional heads ofstate and the OAU Secretary General cornered Habyarimana in the Tanzaniancapital, Dar es Salaam, to convey the same message.

Under this staunch international pressure, Habyarimana finally appeared torelent. Following the meeting, Germany's ambassador stated, "I personallyexpect the establishment of the institutions in the course of thisweek." At the April 6 summit, according to Rwanda's Finance MinisterMarc Rugenera, "there was a great deal of international pressure toproceed with implementation of the agreements. The Akazu [the rulingnorthern-Hutu elite] realized the game was up."22

Later that day, on the trip home, the plane carrying Habyarimana andthe president of Burundi crashed. Press reports, while not definitive,suggest the plane was shot down by Hutu extremists unhappy with thecompromises accepted by Habyarimana. The genocide began the same day.

Mediators' Contribution to the Tragedy

There is little doubt that the proximate cause of the genocide in Rwandawas that President Habyarimana signed and began to implement an agreementthat threatened the privileged position of powerful extremists in hiscountry. While ethnic hatred had long existed in Rwanda, and small-scaleethnic violence revived immediately following the 1990 invasion, it wasnot until the Arusha accords of August 1993 that extremists took thefinal steps necessary to implement genocide. Only in September 1993,for example, did Radio Mille Collines begin broadcasting its message ofethnic hatred used to incite and coordinate the violence.

The question, then, is why did Habyarimana sign and begin to implementan accord that so threatened a powerful and violent segment [End Page 230] of his population? In part, certainly, it was a responseto the military threat posed by the rebels. Indeed, as demonstratedafter the genocide, the RPA was strong enough to defeat the Rwandanarmy.13

But according to Rwandan and western officials, it was continuinginternational pressure that ultimately drove Habyarimana to agree toimplement the accords. While the threat of a renewed rebel offensivewas also a factor, this threat was exacerbated by the removal of Frenchtroops in December 1993. Thus, the international community increasedboth economic and military pressure on Habyarimana, leaving him littlechoice but to sign and implement the accords.

Even the United Nations attempted to use its limited leverage to ensureimplementation. In October 1993, the UNAMIR authorizing resolutionwarned that its deployment beyond 90 days would be contingent on adetermination that "substantive progress has been made towards theimplementation of the Arusha Peace Agreement,"24 and theSecurity Council repeated similar warnings in February and March of1994.25 While this language was intended to compel the partiesby the specter of renewed hostilities, it was also a frank confession ofthe Security Council's distaste for a repeat of its Bosnia and Somaliaexperiences. Unfortunately, the implicit threat may have backfired byalerting the extremists to the fragility of the international militarycommitment.

While international pressure clearly hastened the onset of the genocideby artificially increasing the threat to the extremists, it is possiblethat the violence would have occurred even without internationalpressure. Indeed, so long as the extremists believed they could advancetheir interests more effectively through genocide than power-sharing, theywere likely to lash out if threatened sufficiently. The RPA might havecontinued to advance militarily, threatening the extremists and spurringa genocide in any case. But events could have unfolded differently. Themilitary struggle might have stalemated, defusing the threat. The Rwandangovernment might have pursued a more gradual political reform that didnot immediately threaten the elite, and thereby proved acceptable to theextremists. Most importantly, the genocide was not inevitable. [End Page 231]

Why the Mediators Failed

In order to prevent future such mishaps, it would be useful tounderstand the causes of the mediators' failure in Rwanda, which fallinto two possible categories: inadequate intelligence gathering and/orfaulty judgment. Since the mediators knew that the UN was in no moodfor another peace enforcement operation, they should not have appliedpressure sufficient to provoke a violent outburst from Hutu extremists.However, their information may have been insufficient to alert them tosuch impending risks.

It is important to keep in mind the mixed signals being sent by RwandanPresident Habyarimana. By 1992, Habyarimana faced pressure from at leastfour directions: militarily from the RPA, economically from internationaldonors, politically from the Parmehutu, and internally from extremistsin his own group. His apparent strategy was to maintain the statusquo for as long as possible, doing or saying whatever was necessary torelieve immediate pressure, even if it directly contradicted previouscommitments. For example, to appease the RPA and the internationalcommunity, he repeatedly pledged to implement multi-party democracy, butinevitably renounced these promises to satisfy extremists. Similarly,while he negotiated and signed an elaborate plan to share power withthe Tutsi, he simultaneously established a military and communicationsinfrastructure to annihilate them.

Due to his premature death, it is impossible to know Habyarimana's truenature and ultimate intentions--that is, whether he was an extremistor a reluctant moderate. If an extremist, then his commitments duringthe Arusha process were but a cynical effort to buy sufficient time toorganize and implement a "final solution" for the Tutsi. If a moderate,his repeated backpedaling from those commitments was a desperate, andultimately failed, attempt to mollify extremists in his government whilereluctantly steering Rwanda toward a pluralist future.

As summarized in the following chart, identifying the mediators'specific shortcomings is impossible until the riddle of Habyarimanaand his interaction with the Hutu extremists can be deciphered. [End Page 232]

If the mediators appreciated the strength and desperation of theextremists (cases one, three, five and seven) and/or thought thePresident an extremist (cases three, four, five and six), but insistedon pressuring Rwanda to accept the accords, then they made an error injudgment. They should not have used leverage to impose the accords onextremists, knowing the international community was unwilling to enforcepeace in case the strategy backfired.

If the mediators did not appreciate the strength and nature of theextremists (cases two, four, six and eight) or thought the president amoderate when he was actually an extremist (cases seven and eight), thenthey suffered from a failure of intelligence gathering. This error isperhaps more excusable, since even many Rwandans failed to appreciatefully the extremist threat. As Finance Minister Marc Rugenera laterconceded:

I did not think that those who wanted war constituted a significantthreat. At most I thought they would continue assassinating individual[End Page 233] political opponents. But never, never did I think,even for a second, that they would mount the operation whose results wesee today.16

The Roads Not Taken

In the aftermath of the genocide in Rwanda, critics have cited severalpreventive measures that might have averted the tragedy. While manyof their arguments are superficially compelling, most do not survivecloser scrutiny.


One of the most important lessons to emerge from the field ofinternational conflict management is the ability of mediator leverage tocompel contending parties to sign and implement negotiated settlements.The Rwanda experience provides an equally important caveat: mediatorsmust exercise extreme caution in the use of such leverage, especially insituations where the proposed settlement threatens a powerful interestgroup and where the international community is unwilling to enforceimplementation.

Overly broad conclusions, however, should be avoided. Indeed, it wouldonly magnify the tragedy of Rwanda if the international community tookthe lesson that leverage never should be used to compel a negotiatedoutcome. Rather, the proper lessons are to place a premium on intelligencegathering, to respect the power of leverage, and if in doubt to err onthe side of caution.

In the Rwanda case, the mediators could not have known that theirpressure on President Habyarimana and the extremists to implement theArusha accords would trigger a genocide. Yet, they should not have beensurprised. A wild animal, when cornered, should be expected to attack.

Genocide, while it can never be predicted with certitude, is always apossibility. Mediators, therefore, have a duty to obtain the very bestintelligence about the intentions and capabilities of all parties toa conflict. If there is any chance that imposing a settlement throughthe use of leverage will trigger massive violence, mediators shouldrefrain from applying such pressure unless they or another third partyare willing and able to enforce the peace.

If the international community is unwilling to provide the troops andresources for the risky business of peace enforcement, as currentlyappears to be the case, it may also have to refrain from utilizing itsleverage to compel negotiated settlements. In other words, the price ofavoiding another Rwanda may be that some protracted conflicts drag ona little longer. While this may prove frustrating to statesmen, and totelevision-viewing publics exposed to persistent images of violence,it is the only prudent course. [End Page 237] For if Rwandateaches nothing else about conflict management it is that, sometimes,doing the wrong thing is far worse than doing nothing at all.

Alan Kuperman is a MA candidate in Conflict Management and the 1995-1996Paul H. Nitze Fellow at the Johns Hopkins School of Advanced InternationalStudies (SAIS).


1.See Saadia Touval, "The Superpowers as Mediators," in JacobBercovitch and Jeffrey Z. Rubin, eds., Mediation in InternationalRelations (New York: St. Martin's Press, 1992); I. William Zartmanand Saadia Touval, "Mediation: The Role of Third-Party Diplomacy andInformal Peacemaking," in Sheryl J. Brown and Kimber M. Schraub, eds.,Resolving Third World Conflict (United States Institute of Peace,1992); and Jacob Bercovitch, J. Theodore Anagnoson and Donnette L. Wille,"Some Conceptual Issues and Empirical Trends in the Study of SuccessfulMediation in International Relations," Journal of Peace Research,Volume 28, Number 1 (February 1991), p. 13.

2.For an explanation of these terms, see for example,I. William Zartman, Ripe for Resolution: Conflict and Interventionin Africa (New York: Oxford University Press, 1985).

3.Pre-genocide estimates of the Hutu proportion of thepopulation are from 85 to 90 percent, Tutsi from nine to 14 percent,and Twa about one percent.

4.The identity cards were purportedly the only ones issued onthe continent outside South Africa. See "Unrest Could Ignite WidespreadViolence," Africa News Service, October 22, 1990; and AlexShoumatoff, "Rwanda's Aristocratic Guerrillas," The New York TimesMagazine, December 13, 1992.

5.African Rights, Rwanda--Death, Despair and Defiance(London: African Rights, 1994), pp. 11-13.

6.For a detailed account, see Richard F. Nyrop et al.,Area Handbook for Rwanda (Washington: U.S. Government PrintingOffice, 1969).

7.Randall Fegley, Rwanda (Oxford: Clio Press Ltd.,1993), World Bibliographical Series, Volume 154 , p. xxiii.

8.Fegley, p. xvii.



11.Catherine Watson, "War and Waiting," Africa Report,November-December 1992, p. 55.

12.Filip Reyntjens, with revisions by editor, "Rwanda--RecentHistory," in Africa South of the Sahara (London: EuropaPublications Limited, 1995), 24th Edition, p. 740; and African Rights,pp.13-14.

13.Watson, p.55.

14.BBC Summary of World Broadcasts, Radio Muhabura,translated from French, November 25, 1992.

15.Charles Onyango-Obbo, "Rebellion Adds Momentum to RwandaReform," Africa News Service, April 26, 1993; and David Orr,"Rwanda: Peace of a Sort Heralds End to Years of Misery," ReuterTextline Observer, August 1, 1993.

16.Richard Boucher, US Department of State Spokesman, March9, 1993.

17."Beyond the Rhetoric," Human Rights Watch, Washington DC,June 1993.

18.African Rights, pp. 82-83.

19."Beyond the Rhetoric."

20.Major Rick Orth, The Four Variables of PreventiveDiplomacy: Application in the Rwanda Case, paper presented at the14th Annual Africa Conference, Johns Hopkins University School of AdvancedInternational Studies, Washington, DC, April 7, 1995, p.19.

21.The UN charter actually contains no mention of"peace-keeping," a functional requirement that arose followingestablishment of the United Nations. Nevertheless, UN peace-keepingoperations are considered to be authorized-- whether or not citedexplicitly in the Security Council authorizing resolution-- underChapter VI, which deals with consensual means of settling disputes, suchas mediation and conciliation. By contrast, when the Security Councilauthorizes military force in the absence of the consent of the parties,generally referred to as "peace enforcement," it acts--again, eitherexplicitly or implicitly--under Chapter VII of the charter, "Actionwith Respect to Threats to the Peace, Breaches of the Peace and Actsof Aggression."

22.Both quoted in African Rights, p. 87.

23.Admittedly this defeat came after the army lost thebacking of French troops and was subject to an arms embargo, but manyobservers discount these factors especially in light of the rapidity ofthe RPA triumph.

24.UN Security Council Resolution 872, S/RES/872 (1993),October 5, 1993.

25.African Rights, pp. 84-85.

26.African Rights, pp. 82-83.

27.Orth, p. 30 and footnote 41. While arms embargoes are acommon knee-jerk response to international conflicts, they often proveproblematic even where partially effective. In Bosnia, for example, thearms embargo is widely acknowledged to have handicapped government forcesand thereby facilitated the genocide of Muslims.

28.It can even be argued that the French troops, in retrospect,played a moderating role by mitigating the insecurity of the extremists.

29.Janet Fleischman, Human Rights Watch, presentation atthe 14th Annual Africa Conference, Johns Hopkins University School ofAdvanced International Studies, Washington, DC, April 7, 1995.

30.Anthony Marley, US Department of State, presentation atthe 14th Annual Africa Conference, Johns Hopkins University School ofAdvanced International Studies, Washington, DC, April 7, 1995.

31.UN Secretary General Boutros Boutros-Ghali, "Report of theSecretary-General on the United Nations Assistance Mission for Rwanda,"S/26927, December 30, 1993, pp. 9-10.

32.Article 2.7 of the UN Charter, in Frederic L. Kirgis,Jr., International Organizations in Their Legal Settings: SelectedDocuments (St. Paul, MN: West Publishing Co., 1993), p. 12.

33.See James B. Steinberg, "International Involvement inthe Yugoslavia Conflict," in Lori Fisler Damrosch, ed., EnforcingRestraint (New York: Council on Foreign Relations Press, 1993),pp. 49-55; and Laura W. Reed and Carl Kaysen, eds., Emerging Norms ofJustified Intervention (American Academy of Arts and Sciences, 1993).

34.BBC Summary of World Broadcasts, Zaire TV, Kinshasa,translated from French, September 12, 1993.

35.Quoted in Orth, p. 15.

[Project Muse][Search Page][Journals][Journal Directory][Table of Contents][Top]