oordination - reliefweb.int

146
THOMAS J. WATSON JR. INSTITUTE FOR INTERNATIONAL STUDIES Occasional Paper #40 THE DYNAMICS OF COORDINATION MARC SOMMERS

Upload: others

Post on 17-Jan-2022

15 views

Category:

Documents


0 download

TRANSCRIPT

WIIS • 2 Stimson Ave. • Brown University • Providence, RI 02912

TH

E DY

NA

MIC

S OF C

OO

RD

INA

TIO

NO

ccasional Paper #40

THOMAS J. WATSON JR. INSTITUTE FOR INTERNATIONAL STUDIES

Occasional Paper #40

THE DYNAMICS OF COORDINATION

MARC SOMMERS

ATTENTION!!!
For citation purposes, use the page numbers that appear at the bottom of the text pages (which correspond to those on the published version). DO NOT USE the Acrobat page numbers.
ATTENTION!!!
For citation purposes, use the page numbers that appear at the bottom of the text pages (which correspond to those on the published version). DO NOT USE the Acrobat page numbers.

OCCASIONAL PAPERS OF THE WATSON INSTITUTE

#1: North-South/East-West: Establishing a Common Agenda by Harry G. Barnes Jr., YuriA. Krasin, and Enriqué Iglesias#2: The United States, the Soviet Union and Security in the Third World by MichaelClough, Meryl A. Kessler, Robert Levgold, S. Neil MacFarlane, Augustus RichardNorton, Celeste A. Wallander, and Thomas G. Weiss#3: Human Sources of Global Change: Priority Research Initiatives for 1990-1995 by RobertW. Kates, William C. Clark, Vicki Norberg-Bohm, and B. L. Turner II#4: Superpowers and Regional Conflict in a Post-Cold War World: The Caribbean Basin andSouthern Africa by James G. Blight, Jorge I. Dominguez, Fen Osler Hampson, Jo L.Husbands, Gillian Gunn, Lloyd Searwar, Newell M. Stultz, Wayne S. Smith, andThomas G. Weiss#5: Toward Collective Security: Two Views by Sir Brian Urquhart and Robert S.McNamara#6: Brazil and Mexico: Contrasting Models of Media and Democratization by Ilya Adler,Elizabeth Mahan, Joseph Straubhaar, and Nelson H. Vieira; introduction by ThomasE. Skidmore#7: Working Together by Eduard A. Shevardnadze#8: Humanitarianism and War: Learning the Lessons from Recent Armed Conflicts byLarry Minear, Thomas G. Weiss, and Kurt M. Campbell#9: “This Will Not Be Another Vietnam:”George Bush and the Persian Gulf War byRichard A. Melanson#10: Collective Security in a Changing World by Lincoln P. Bloomfield, Tom J. Farer,Leon Gordenker, Ernst B. Haas, John Mackinlay, Oscar Schachter, James S. Sutterlin,and Thomas G. Weiss#11: Competing Gods: Religious Pluralism in Latin America by Gerard Béhague, David J.Hess, Marc Belanger, and Anani Dzidzienyo; introduction by Thomas E. Skidmore#12: Continuity and Change: Women at the Close of the Twentieth Century by ReginaCortina, Eleanor Doumato, Marida Hollos, Prema Kurien, and MarilynRueschemeyer#13: United Nations Coordination of the International Humanitarian Response to the GulfCrisis 1990-1992 by Larry Minear, U. B. P. Chelliah, Jeff Crisp, John Mackinlay, andThomas G. Weiss#14: Humanitarian Challenges in Central America: Learning the Lessons of Recent ArmedConflicts by Cristina Eguizábal, David Lewis, Larry Minear, Peter Sollis, and ThomasG. Weiss#15: United Nations Authority in Cambodia by Jarat Chopra#16: United Nations Financing Problems and the New Generation ofPeacekeeping and Peace Enforcement by Anthony McDermott#17: German Big Business and Europe in the Twentieth Century by Volker R. Berghahn,Reinhard Neebe, and Jeffrey J. Anderson#18: Humanitarian Action in the Former Yugoslavia: The UN’s Role, 1991-1993 by LarryMinear (team leader), Jeffrey Clark, Roberta Cohen, Dennis Gallagher, Iain Guest,and Thomas G. Weiss

#19: Mexico: The Artist is a Woman by Lucretia Giese, Carmen Boullosa, MarjorieAgosín, Sandra Berler, Elena Gascón-Vera, Laura Riesco, and Margo Glantz;edited by Regina Cortina#20: Humanitarian Action and Security in Liberia 1989-1994 by Colin Scott, incollaboration with Larry Minear and Thomas G. Weiss#21: Armed Conflict in Georgia: A Case Study in Humanitarian Action and Peacekeepingby S. Neil MacFarlane, Larry Minear, and Stephen Shenfield#22: The Policies of Mercy: UN Coordination in Afghanistan, Mozambique, and Rwandaby Antonio Donini#23: Haiti Held Hostage: International Responses to the Quest for Nationhood 1986-1996by Robert Maguire (team leader), Edwige Balutansky, Jacques Fomerand, LarryMinear, William G. O’Neill, Thomas G. Weiss, and Sarah Zaidi#24: Imagining a Free Cuba: Carlos Manuel de Céspedes and José Martí by Cathy L.Jrade, George Monteiro, Nelson R. Orringer, Louis A. Pérez, Jr., Ivan A. Schulman,Thomas E. Skidmore, and Wayne S. Smith; edited by José Amor y Vázquez#25: Humanitarian Action and Politics: The Case of Nagorno-Karabakh by S. NeilMacFarlane and Larry Minear#26: War and Humanitarian Action in Chechnya by Greg Hansen and Robert Seely#27: National Self-Determination: Approaches and Case Studies by Galina Starovoitova#28: An Overview and Assessment of 1989-1996 Peace Operations Publications byCindy Collins and Thomas G. Weiss#29: Humanitarian Impacts of Economic Sanctions in Burundi by Eric Hoskins andSamantha Nutt#30: Integration and Disintegration in the Former Soviet Union: Implications forRegional and Global Security by P. Terrence Hopmann, Stephen D. Shenfield, andDominique Arel#31: Toward More Humane and Effective Sanctions Management: Enhancing theCapacity of the United Nations System by Larry Minear, David Cortright, JuliaWagler, George A. Lopez, and Thomas G. Weiss#32: Humanitarian Action in the Caucasus: A Guide for Practitioners by Greg Hansen#33: Relief and Development: The Struggle for Synergy by Ian Smillie#34: A Humanitarian Practitioner's Guide to International Human Rights Law byWilliam G. O'Neill#35: Protecting Human Rights: The Challenge To Humanitarian Organizations byMark Frohardt, Diane Paul, and Larry Minear#36: NATO and Humanitarian Action in the Kosovo Crisis by Larry Minear, Ted vanBaarda, and Marc Sommers#37: Humanitarian Action: Social Science Connections by Stephen C. Lubkemann,Larry Minear, and Thomas G. Weiss#38: When Needs Are Rights: An Overview of UN Efforts to Integrate Human Rights inHumanitarian Action by Karen Kenny#39: Humanitarian Action: A Transatlantic Agenda for Operations and Research byLarry Minear and Thomas G. Weiss#40: The Dynamics of Coordination by Marc Sommers

i

Occasional Paper #40

THE DYNAMICS OF COORDINATION

Marc Sommers

ii

Occasional Papers is a series published by

The Thomas J. Watson Jr.Institute for International StudiesBrown University, Box 19702 Stimson AvenueProvidence, RI 02912

Telephone: (401) 863-2809Fax: (401) 863-1270E-mail: [email protected]://www.brown.edu/Departments/Watson_Institute/

Thomas J. Biersteker, Ph.D., Director

Statements of fact or opinion are solely those of the authors; theirpublication does not imply endorsement by the Thomas J. Watson Jr.Institute for International Studies.

Copyright © 2000 by the Thomas J. Watson Jr. Institute for Interna-tional Studies. All rights reserved under International and PanAmerican Convention. No part of this report may be reproduced ortransmitted in any form or by any other means, electronic or me-chanical, including photocopy, recording, or any information stor-age and retrieval system, without prior written permission from thepublisher. All inquiries should be addressed to Occasional Papers,Thomas J. Watson Jr. Institute for International Studies.

iii

CONTENTS

Foreword................................................................................... v

Executive Summary ................................................................. ix

1. The Coordination Challenge .............................................. 1

2. The Conflicts in Sierra Leone and Rwanda...................... 7

3. Coordination Among Organizations .............................. 21

4. Coordination Among Functions ...................................... 51

5. Coordination Within Programs ....................................... 81

6. Conclusions and Recommendations ............................... 99

Notes ........................................................................................ 113

Appendix I: Acronyms ...................................................... 123Appendix II: Chronology of Major Events ...................... 125Appendix III: Agencies Consulted ..................................... 129Appendix IV: About the Author and the Humanitarianism and War Project ........... 131

iv

v

FOREWORD

In the field of humanitarian action, everyone touts thevalue of coordination. Donors want to ensure the cost-effec-tive use of resources. Political authorities receiving humani-tarian assistance expect coherence among the many actors thatdescend upon them at the drop of an emergency. Aid agenciesthemselves have an interest in an efficient division of labor andfear media exposure of interagency rivalries. The public hashad enough experience with humanitarian circuses to expectthe worst.

Beneath the surface, however, the situation is far morecomplex. Donors complicate the task of coordination by im-posing conditions, earmarking resources, and injecting otherintrusions into aid work. The host authorities have politicalagendas that may work against the coordination they professto want. The agencies themselves have little patience with thecosts of coordination, whether in reduced flexibility or inresources diverted from assistance programs. The public isoften fickle, insisting that contributions be applied to the directalleviation of suffering rather than underwriting the coordina-tion machinery to ensure greater cost-effectiveness.

Coordination involves the systematic use of policy instru-ments to provide effective assistance and protection to vulner-able populations. This definition—quoted in full in Chapter 1of this study—emerged from the Humanitarianism and WarProject’s 1992 review of the crisis in the Persian Gulf. Coordi-nation has been a leitmotiv in subsequent country reviews andthematic studies.

Perhaps our most extensive treatment of the subject was inThe Policies of Mercy: UN Coordination in Afghanistan,Mozambique, and Rwanda, a 1996 review by Antonio Donini.The author, on leave from coordination responsibilities in theUN secretariat, proposed a tripartite typology. Coordinationby command, he suggested, involves leadership authority,reinforced by carrots and/or sticks; coordination by consen-sus uses less authority and requires more persuasion; andcoordination by default has even fewer structures for orches-trating concerted action.

Coordination by command, Donini concluded, is not a

vi

realistic option either for donors or for the United Nations(UN) itself. The most to be hoped for, he said, is coordinationby consensus, in which “UN organizations and various NGOswould at least share information and attempt to avoid dupli-cation.” Yet his three-country comparison demonstrated thevalue—other analysts might say the indispensability—of acommand element. Even a modest amount of resources en-trusted to the designated coordination agent, Donini found,produced significant paybacks in program effectiveness. Inthe absence of more assertive authority, he concluded, re-sponses that “rely solely on personalities, goodwill, and intel-lectual leadership” are likely to disappoint.

Several years and numerous crises later, there is nowgreater support for the command model. A newly releasedU.S. State Department study of American humanitarian policyand practice tells a familiar “tale of poor coordination, mis-sions being duplicated or falling through the cracks, andconfusion inside the administration and the private humani-tarian groups that sometimes cannot be sure with which[government] agency to work.”

At the same time, however, confusion and ambivalencecontinue to characterize many current discussions of coordi-nation. On one occasion in early 2000 at which donor govern-ment officials and international NGOs discussed improvingdisaster coordination, some nongovernmental organizations(NGOs) expressed a preference for the phrase “operationalcooperation” rather than coordination. A wary donor govern-ment official countered by saying that her agency would notfund anything that failed explicitly to embrace the wordcoordination.

Earlier reviews by the Humanitarianism and War Projectprovide the context in which the current study of the dynamicsof coordination is set. As in previous work, we approach thesubject from an operational standpoint. The issue is not whatis said about coordination at agency headquarters and aroundinteragency tables in New York and Geneva. It is, rather, whathappens on the ground in major humanitarian emergencies.This particular study reviews the international responses tothe situation in to Sierra Leone during the years 1994 to 1998and to Rwandan refugees in Ngara, Tanzania between April1994 and December 1996.

vii

Each crisis evolved differently in terms of the political-military context on the ground and in terms of the approachesto coordination that were employed. In comparing and con-trasting developments in each of the two locations, this studyviews the coordination issue from three vantage points: orga-nizational (Chapter 3), functional (Chapter 4), and program-matic (Chapter 5). It seeks to answer questions such as thefollowing:

• Is there a correlation between the degree of coordina-tion authority and the effectiveness of humanitarianactivities?

• What is the relative importance of institutional struc-tures, on the one hand, and leadership by key humani-tarian officials, on the other?

• To what extent does effective coordination requirecompatible approaches among humanitarian organi-zations to neutrality and political engagement?

In keeping with the Project’s inductive methodology, webelieve that there is much to be learned from a close-in reviewof the dynamics of situations as they evolve on the ground.Those dynamics are, in one sense, unique to each situation.Chapter 2 thus examines the conflicts in Sierra Leone andRwanda that created the need for humanitarian action. Ap-pendix II provides a chronology of major events in each of thetwo settings. Maps are also included in the text.

Yet, humanitarian challenges and operational and institu-tional constraints recur in crisis after crisis. As a result, thereare significant lessons with potentially wider implications tobe learned from a rigorous comparative analysis of situationssuch as Sierra Leone and Ngara. Some of these are identifiedin Chapter 6, which provides a number of concluding reflec-tions.

We are pleased that Marc Sommers has brought to bear onthe task at hand his skills as a social scientist (an anthropolo-gist, to be more precise) and as a knowledgeable consultant tohumanitarian organizations. (Biographical information onSommers is provided in Appendix IV.) He conducted the bulkof the more than 100 interviews for this study in 1998 to 1999,

viii

supplementing field visits to both locations with discussionsin Geneva, Brussels, New York, and Washington. The agenciesthat contributed their viewpoints are identified in AppendixIII. We express our thanks to the many officials interviewed,some at great length and on more than a single occasion.

Our colleagues Giles Whitcomb and Thomas G. Weisswere involved several years ago in discussions that helpedframe the parameters of this study. We are also indebted toLaura Sadovnikoff, Margareta Levitsky, Ryoko Saito, and JoySomberg of the Project and Frederick Fullerton of the WatsonInstitute for assistance in the production process. Specialthanks go to the editor of this volume, Mary Lhowe.

The work of the Project is underwritten by financial con-tributors from practitioner organizations, governments, UNagencies, and foundations identified in Appendix IV. We aregrateful for their ongoing support and for the use that theymake of our materials. We welcome comments from them andother users.

This Occasional Paper will be among the final publica-tions by the Humanitarianism and War Project at the WatsonInstitute at Brown University. In its new home at the FeinsteinInternational Famine Center at Tufts University after Septem-ber 1, 2000, the Project will continue its research and publica-tions. Additional copies of this and earlier monographs maybe downloaded directly from our website at www.brown.edu/Departments/Watson_Institute/H_W. Our new website atTufts will be www.hwproject.tufts.edu.

Larry Minear, DirectorHumanitarianism and War ProjectProvidence, Rhode IslandJuly 2000

ix

EXECUTIVE SUMMARY

This study takes a fresh look at coordination dynamics byexploring two situations that were noteworthy both for thedifficult challenges humanitarian organizations faced and fortheir innovative responses. The cases shed light on coordina-tion in general and two contexts in particular: coordination fora country (Sierra Leone in the 1990s) and coordination forrefugee camps (Rwandan refugees in Ngara District, Tanza-nia).

After reviewing the conflicts that precipitated the crisis ineach setting, the study analyzes coordination from three per-spectives. The first explores the nature of relationships amonghumanitarian actors in Sierra Leone and in Ngara District. Thesecond highlights three important coordination functions:strategic planning, resource mobilization, and security infor-mation management. The third examines two examples ofprogram coordination: the controversial repatriation ofRwandan refugees from Tanzania, and education for SierraLeonean and Liberian refugees. Throughout this monograph,the Sierra Leone and Ngara cases are used as optics forviewing how six sets of humanitarian actors—UN agencies,international and local NGOs, the Red Cross movement, do-nor and national governments, military contingents and rep-resentatives of recipient populations—respond to dynamicchallenges in the field.

The study finds examples of significant, replicable inno-vation in both field cases, such as the Committee on Food Aid(CFA) for the Sierra Leonean food sector and coordinateddonor funding to the United Nations High Commissioner forRefugees (UNHCR) in Ngara. It also finds difficult challenges,such as three competing UN entities in Sierra Leone andUNHCR’s evolving relationship with the Tanzanian govern-ment.

The Dynamics of Coordination concludes that donor govern-ments command more power than any other set of actors toexercise a positive influence on coordination. At the sametime, the marginalization of national governments from coor-dination structures is shortsighted because it limits synergiesbetween relief and longer-term development and risks alien-

x

ating national authorities that have been and will be keyplayers. The study also finds that humanitarian officials tendto overstate the role personalities play in bringing abouteffective coordination and to underemphasize the signifi-cance of well-structured institutional relationships and clearlydelineated coordination systems.

Final recommendations highlight the importance of incor-porating a command element into the practice of humanitar-ian coordination and establishing a clear role in coordinationfor the national authorities. It also recommends that UNagencies and key humanitarian donors dramatically improvetheir relationships; that the UN’s Office for the Coordinationof Humanitarian Affairs (OCHA) be expected to exercisecoordination for all in-country responses; that UN and NGOsdevelop a formal coordination arrangement adaptable to spe-cific contexts; that the UN’s Consolidated Appeals Process(CAP) be reformed; that key roles in humanitarian action forlocal NGOs be recognized; and that all major humanitarianactors, and international NGOs in particular, work to improvetheir relations with national government actors. Innovationsto be replicated elsewhere, with specified adjustments, in-clude the collaboration between the European CommunityHumanitarian Office (ECHO) and the U.S. government, re-ductions in the number of NGOs in humanitarian theaters,and structures like the Committee on Food Aid used in SierraLeone.

1

CHAPTER 1

THE COORDINATION CHALLENGE

Former Médecin sans Frontières (MSF) President RonyBrauman has observed, “Humanitarian action is the principalexpression of collective values.”1 If so, then why has it provenso difficult for humanitarian actors to work well together? Inreality, humanitarian action is neither as simple nor as straight-forward as suggested by a signboard posted on an Americanmilitary jeep in Kosovo in 1999—“Savin’ Lives.” Shared val-ues do not necessarily lead towards coordinated responses,much less cooperative relationships.

The challenge of creating and maintaining a coordinatedinternational response to a given complex humanitarianemergency begins with different meanings of the word “co-ordination.” Antonio Donini’s observation that humanitar-ian agency officials in Rwanda during the 1994 crisis had “noclear or common understanding of what ‘coordination’meant” could be applied to any number of humanitariancrises.2 Even in dictionaries, the word coordination hasconflicting meanings. One definition of coordination as anoun posits it as the “act or state of coordinating or beingcoordinated,” based on the verb “to coordinate,” which itconsiders “of the same order or degree” and “equal in rankand importance.” A second definition, however, suggeststhat coordination, in addition to conveying equality amongparts, can also confer a “proper order or relationship” amongthem.3 In times of humanitarian crisis, should coordinationarrangements lean towards horizontal equality, as the firstdefinition suggests, or towards more hierarchical arrange-ments, as the second implies?

Coordination lies at the heart of the international humani-tarian enterprise. When a humanitarian emergency erupts,officials from governments, UN and nongovernmental orga-nizations, along with journalists and other interested parties,descend on the scene, armed with a variety of mission orders,capacities, and institutional agendas. Harnessing their diver-gent energies to create a cohesive system is hardly easy. It mayalso be improvisational, leading to an uncertain result. Many

2

people assume that the UN should assume the mantle ofleadership and set up a harmonious humanitarian responsesystem. But agencies negotiate their relationships afresh at theoutset of each new humanitarian emergency and they may ormay not place the UN at the center. In the view of onehumanitarian coordination expert interviewed for this study,“There is no coordination system for the United Nations. FullStop. None.”

Examining Coordination Through Two Case Studies

This monograph explores the dynamics of coordinationby reviewing two major humanitarian emergencies. One arosein Sierra Leone in the wake of civil war. The case studyexamines field coordination from 1994 through 1998, particu-larly the latter part of this period. The other involved refugeesin Ngara District of western Tanzania who fled genocide andcivil war in Rwanda. Study of the Ngara case extends from theRwandan refugees’ entry into Tanzania in April 1994 untiltheir repatriated return to Rwanda in December 1996.

These two cases represent responses to two differentkinds of humanitarian emergencies, each with a differentmodel of coordination. The Ngara case illustrates coordina-tion of refugees in camps, where the United Nations HighCommissioner for Refugees is typically charged with orches-trating assistance with international and sometimes local NGOsas implementing partners. Host governments usually enterinto tripartite arrangements with UNHCR and the govern-ment of the refugees’ country of origin.

For a country—as distinct from camps—the model ofhumanitarian coordination is more complicated. In this ap-proach, illustrated by the Sierra Leone case, a number of UNagencies may compete for a seat at or near the top of thecoordination structure. While the UN Office for the Coordina-tion of Humanitarian Affairs is charged with performing thefunction specified in its name, the country representative fora second agency, the UN Development Programme (UNDP),is usually expected to become the primary coordinating chiefand to receive the title of Humanitarian Coordinator (HC). Insome situations, a Special Representative for the Secretary-

3

General (SRSG) may also be appointed. The Sierra Leone casecontains all three entities and illuminates how working rela-tionships between them must be negotiated in the field.

As these three UN focal points (and perhaps other UNagencies as well) negotiate relations, UN relationships withinternational NGOs and Red Cross agencies tend to be lessstructured than in refugee camps. In a country, however,NGOs may have already established a presence as develop-ment agencies long before famine or war abruptly shifted theiroperations into the humanitarian relief sphere. Thus, evenwhile it may appear that the “international community” isoperating under the auspices of the UN during a crisis, inpractice, NGO and Red Cross officials who are accustomed toworking independently of the United Nations in pre-crisistimes may not be inclined to fall in step behind UN agenciesonce emergencies arise. This reality usually makes coordinat-ing humanitarian emergencies inside a country in crisis farmore complex than in a refugee camp. It is thus necessary todescribe the Sierra Leone case in more detail.

These two case studies should not be seen as evaluationsof performance. “Sierra Leone” and “Ngara”—here used asshorthand for the crises they represent—are explorations oftwo sets of coordination challenges, dynamics and innova-tions. The frame of reference includes six sets of actors: UNagencies, international (and, where applicable, local) NGOs,the Red Cross movement, donor and national governments,Military contingents, and representatives of recipient popula-tions.

Organizing Themes and Subjects

Two recurrent themes are the impacts of uncertainty andtension on the coordination process. Uncertainties arise fromthe tenuous nature of humanitarian crises and they affect thecontext for coordination work. Both Sierra Leone and Ngaraprovide numerous examples of the impact of unpredictabilityon humanitarian coordination and on how humanitariansrespond to unexpected challenges. Second, a range of internaltensions between humanitarian actors regularly creates testy,adversarial, and sometimes counterproductive relationships

4

that challenge the actors’ ability to coordinate their actions.The primary questions considered here are:

• What is the appropriate coordination relationshipbetween UN bodies charged with coordination; be-tween UN bodies and NGOs; between donors andNGOs, UN agencies and the local government; be-tween international humanitarian agencies and in-digenous actors, including government authorities,NGOs, and recipients; and between military and civil-ian actors?

• To what extent does successful humanitarian coordi-nation depend on local governments relinquishingsome of their sovereignty?

• What is the appropriate balance between meetinghumanitarian needs and supporting reconstructionor development initiatives?

• Can humanitarian actors committed to neutrality findways of working with other humanitarian organiza-tions that have embraced the political objectives ofone set of belligerents?

• How should coordination arrangements balance theneed for strong leadership against the desirability offirm institutional structure?

• Can coordination innovations from these two settingsbe replicated elsewhere?

Coordination is a multidimensional activity that takesplace among a variety of actors at multiple levels across arange of activities. It is a dynamic process, responding tochanging political, military, and humanitarian circumstanceson the ground. The conflicts that generated the need forhumanitarian action in Sierra Leone and Ngara, are reviewedin Chapter 2. Coordination in its sense of orchestrating rela-tionships among organizations is the focus of Chapter 3.Coordination as the elaboration of a division of labor amongfunctions is reviewed in Chapter 4, with strategic planning,resource issues, and information management singled out forspecial attention. Coordination of activities at a more pro-grammatic level is highlighted in Chapter 5. The final chapteridentifies conclusions and offers several recommendations.

5

Methods and a Definition

At the heart of the research methodology were interviewscarried out with approximately 100 officials with first-handknowledge of humanitarian activities in Sierra Leone andNgara. The interviews covered many subjects, from the UN’sConsolidated Appeals Process and the Committee on FoodAid in Sierra Leone to UNHCR’s dominance over the coordi-nation process and involvement in the refoulement effort inNgara. A recurring theme of interviews, however, was that thetask of coordination was as personal and idiosyncratic as itwas technical and demanding. Indeed, the role of personalitiesfigured prominently. Although the roles that key individualsplayed will be examined here, attention will also be paid to theinterplay between such individuals and coordination struc-tures or systems.

The concept of coordination that will serve as the refer-ence point is as follows:

Coordination is the systematic use of policy instrumentsto deliver humanitarian assistance in a cohesive and effectivemanner. Such instruments include: (1) strategic planning; (2)gathering data and managing information; (3) mobilizingresources and assuring accountability; (4) orchestrating a func-tional division of labor in the field; (5) negotiating and main-taining a serviceable framework with host political authori-ties; and (6) providing leadership.4

This definition informs the structure and analysis of thisstudy, with two caveats. First, while successful humanitariancoordination requires that all involved feel it is necessary,often they do not. In both Sierra Leone and Ngara, somehumanitarian actors viewed coordinated action as restrictiveand even counter to fundamental humanitarian principles. Intheir opinion, the most effective humanitarian response maynot necessarily have been a coordinated one.

Second, using policy instruments in a systematic fashionin the field may represent a distant goal rather than a presentreality. Coordination is a messy, dynamic, and evolving pro-cess; the crises that created the humanitarian emergencies inthe first place ensure that this will be true. Strategic planningmay be directed far more at reacting to changing events than

6

developing plans in advance of them. Whether one ultimatelyfaults the individuals or agencies involved or the circum-stances, coordination efforts often center on developing im-mediate, short-term solutions to difficult, perplexing, andrapidly evolving problems.

7

CHAPTER 2

THE CONFLICTS IN SIERRA LEONE AND RWANDA

An examination of coordination in contexts as different asan entire war-torn country and a single district containingrefugee camps might lend the impression of two situationswith little in common. But Sierra Leone and Ngara sharedmany important similarities. In both, the international coordi-nation structures were headed by UN bodies that developedboth positive and problematic relationships with interna-tional NGOs. Relationships between international agenciesand national government bodies also appeared deceptivelyeasy in both cases, but they eventually presented considerableproblems for coordinated humanitarian action.

Moreover, in both cases national and international actorsalike had difficulty in dealing with nongovernmental authori-ties, namely the rebel Revolutionary United Front (RUF) inSierra Leone and the Rwandan refugee leaders in Ngara. InSierra Leone, inability to relate to the RUF with any regularityexcept for one brief period during the war made humanitarianwork in areas the RUF controlled virtually impossible and thethreat of RUF attacks on aid convoys an everyday possibility.Rwandan refugee leaders in Ngara were readily accessibleand frequently helpful to humanitarian agencies. But on cer-tain issues, particularly relating to repatriation, they blockedprogress and eventually foreshortened the Rwandans’ stay inNgara.

Another important similarity between the Sierra Leoneand Ngara cases was the regional impact of nearby wars. Civilwars in Liberia and Burundi directly influenced the coordina-tion challenges in Sierra Leone and in Rwandan refugee campsin Ngara, respectively. Contingency planning and strategicanalyses had to incorporate the impact of instability andnegative influences from troubles just over several borders,heightening the difficulties of coordination.

In many other ways, the two conflicts described here wereas different as the coordinated responses that aid agenciesengineered for them. Sierra Leoneans suffered from nearly adecade of killing, amputations, displacement, and child ex-

8

ploitation. The war began years before aid agencies declareda humanitarian emergency in the country. The Rwandanexperience of conflict centered on a sudden genocide of ex-traordinary proportions, which dramatically increased theneed for humanitarian aid. This dissimilarity between theconflicts was reflected in the international response each re-ceived. For much of the 1990s Sierra Leone was probablyAfrica’s most overlooked war, while Rwanda’s 1994 genocideimmediately attracted widespread attention.

There were other differences, too. RUF banditry and sa-dism usually made it easy for Sierra Leoneans to disassociatethemselves from the war’s combatants. But the refugee campsfor Rwandans in Ngara contained fugitives from justice—theorganizers and perpetrators of genocidal acts. Sierra Leone’swar was not ethnically based, while Rwanda’s conflict was.Neither side in Sierra Leone’s conflict employed many landmines, which were used regularly in Rwanda. Finally, some ofthe fiercest fighting in Sierra Leone’s war has taken place in thevicinity of its many diamond mines, but Rwanda’s war did notfeature vying for control of natural resources (the countrylacks resources).

What follows is a brief description of the two conflicts,with particular attention to aspects that influenced coordina-tion efforts. While the Sierra Leonean conflict is reviewed fromits beginning in 1991 to the present, the horizons of the Rwandanconflict are set more narrowly, ranging from the outset ofgenocide in April 1994 until the December 1996, when nearlyall Rwandan refugees in Ngara were forcibly repatriated. Theaccount presented here does not include events in Rwandaand the Democratic Republic of Congo that happened after theRwandans had returned from Tanzania.

Sierra Leone: A Creeping Catastrophe

Sierra Leone’s tragic strife contains all the elements of ariveting war novel. Since its beginning in 1991, the coups,corruption, and terror have involved mercenaries from SouthAfrica and Ukraine, Gurkhas from Nepal, diamond smug-glers, militias with magical charms and a reputation for canni-balism, a regional fighting force, secret societies, and thou-sands of child soldiers and captives.

9

10

Novelesque elements notwithstanding, war in Sierra Leonehas been as real as it has been devastating. Eight years of civilwar triggered a sequence of political upheavals, including twomilitary coups; the euphoria of democracy’s return in 1996with the election of Ahmed Tejan Kabbah as president; widelyhailed peace accords; a third military coup in 1997 carried outby dissident military soldiers called the Armed Forces Revo-lutionary Council (AFRC), which was joined by the RUF; asecond failed peace accord in 1997; Nigeria’s rout of the RUF-AFRC junta in Freetown in early 1998; Kabbah’s return to thepresidency; a second attack on Freetown in January, 1999; anda third peace treaty in June 1999 and its unraveling early in2000. Long years of uncertainty mixed with spasms of suddenviolence greatly hampered the international community’sability to coordinate humanitarian relief. The conflict left morethan 50,000 Sierra Leoneans dead and at least one milliondisplaced. The number of refugees in surrounding countries(but mostly in Guinea) has risen to 500,000.1

Counter to perceptions of the RUF as sadistic hooligans,2

over the course of the war the Revolutionary United Frontbecame, quite simply, one of the best guerrilla outfits in theworld. Consistently underestimated, the RUF’s brilliant anddiabolical leader, Foday Sankoh, together with commanderssuch as Sam “Mosquito” Bokarie, developed an astonishinglysuccessful military strategy. It featured children between theages of 7 and 14,3 whom they captured and then used asfighters, forced laborers, and concubines. Its well-deservedreputation for amputating limbs was not a sign of madness butrather part of a calculated strategy to depopulate vast areas ofinland Sierra Leone.4 This obviously prevented the RUF frombuilding a popular base—a central criticism of the RUF—butthe secretive group lacked a well-known political platform inthe first place. Its effectiveness rested on its ability to com-mand political power by wreaking havoc on civilians virtu-ally at will, to cultivate a youth-oriented ideology within itsranks, and to control some of the world’s richest diamondmines. The RUF accomplished this with a fighting force ofperhaps 5,000 combatants for much of the war, most of whomwere young and profoundly traumatized children.5

The implications of RUF tactics for humanitarian work

11

have been considerable. For much of the war, security re-mained a major concern for civilians and aid personnel alike.No agency was able to negotiate with the RUF to gain ongoingaccess to displaced populations behind RUF lines. RUF attackscreated huge displaced populations in cities and in refugeecamps in Guinea, with RUF infiltrators in their midst.

If international actors in Sierra Leone were largely unableto negotiate humanitarian space with the RUF, the distancemany maintained with government officials was a matter ofchoice. Aid agencies frequently created coordination arrange-ments that kept national government actors on the periphery.While some officials were concerned about aligning them-selves too closely with one side in the civil war, an overridingconcern for many was more pragmatic. With few exceptions,they considered Sierra Leonean government officials corruptand lacking the requisite training and skills. Granting themtoo much control over coordination mechanisms would thusslow the delivery of emergency services. Cooperation withgovernment counterparts might pay dividends, but activecollaboration with the authorities, many agencies and donorsbelieved, had to be minimized.

The corruption of governance in Sierra Leone has deeproots. Since the country gained independence in 1961, itsexcellent and abundant diamonds have bankrolled a wealthyand nepotistic political elite and separated it from one of theworld’s poorest populations. A quarter century of politicaldominance and economic mismanagement by President SiakaStevens laid the groundwork for a declining economy. Thegross domestic product (GDP) growth rate fell an average of3.7 percent between 1965 and 1973, to an average of 1.8 percentbetween 1974 and 1984, to -3.6 percent in 1995.6 In 1996 theUNDP Human Development Report ranked Sierra Leone lastamong 174 countries in per capita gross national product(GNP), life expectancy at birth, and adult literacy in 174countries. Even before the economy’s collapse following theMay 1997 coup, the Economist Intelligence Unit predicted that“the [Sierra Leonean] economy’s overwhelming reliance ondonor aid was set to continue well into the 21st century.”7

Sierra Leone’s long-running war can be traced back notonly to the government’s legacy of corruption amidst extreme

12

poverty—a gap that the RUF has tried to exploit among poorSierra Leoneans— but also to a desire for revenge by Liberia’smost famous militia leader and current president, CharlesTaylor. The war began in Sierra Leone during the term ofStevens’s handpicked successor, Major-General Joseph SaiduMomoh. A handful of Sierra Leonean political exiles andeconomic refugees who had settled in Liberia began to receivedirect support from Charles Taylor, then the political leader ofthe National Patriotic Front of Liberia (NPFL), which hadinvaded Liberia from Cote d’Ivoire in late 1989. By the middleof 1990, Liberia had become a humanitarian catastrophe, withmuch of the population displaced, nearly 400,000 of whomlived in refugee camps in surrounding countries. That sameyear, Sierra Leone President Momoh decided to allow theEconomic Community of West African States MonitoringGroup (ECOMOG) forces, the West African multilateral force,to use Sierra Leone as a staging post.

While Momoh evidently sought to play the role of peace-maker in Liberia’s conflict, Sierra Leone’s involvement inLiberia’s war turned out to be Momoh’s undoing. 8 ECOMOGtroops stopped Taylor’s NPFL fighters from overtakingMonrovia in 1990. Part of Taylor’s response was to encourageand supply the RUF’s entrance into Sierra Leone in 1991.9 InApril 1992, Momoh was overthrown in a military coup, fled toGuinea, and was replaced by Captain Valentine Strasser, whoat age 26 became the world’s youngest head of state.

Although Sierra Leone’s civil war did not shift from apolitical rebellion to a resource war until 1993, interviews withSierra Leonean refugees from the RUF’s initial entry point inthe country—the peninsula-shaped district of Kailahun—sug-gest that the political component of the RUF’s fight did not lasteven that long. Disgusted by the pervasive poverty and cor-ruption in the country, many refugee parents reported howthey had allowed their eldest sons to remain behind in KailahunDistrict while they fled the fighting. The boys were instructedto listen to the RUF’s village presentations and report back totheir parents. But refugees’ descriptions of what took placematch the reports of RUF tactics by Paul Richards. 10 Insteadof leading a “people’s movement for national recovery,” theRUF captured youths and forcibly conscripted them into their

13

fighting force. 11 This early experience with the RUF markedthe beginning of a war that many Sierra Leoneans eventuallyviewed as pointless and bewildering.12

The war was also difficult to understand for the interna-tional community. Were RUF fighters mere bandits or a trueopposition force? If bandits, would aid agencies be relieved ofthe need to avoid choosing sides between the RUF and anelected government? But if the RUF were an opposition force,it was a difficult one to decipher. Its leaders were reclusive, theorganization decentralized. Direct connections between“Mosquito’s” troops in Kailahun; bush camps allegedly headedby Liberians; and Foday Sankoh, who spent much of the warin forced exile, were hard to make out.

The uncertainty and danger that would plague humani-tarian action in Sierra Leone throughout the war was furthercomplicated by the advent of a new kind of actor: the sobel.Sierra Leoneans coined this word to describe national armysoldiers who impersonated rebels. As soldiers, they protectedcivilians, but at night they looted them, and then, perhaps,amputated a civilian’s arm to make it seem that the RUF hadcarried out the atrocity. It was thus never clear just who hadcarried out a particular village raid. Were they actual RUFrebels or sobel impersonators? “Who are the RUF anyway?”said a senior UN humanitarian official charged with coordina-tion tasks. “Can you trust them when they make promises?”

There was still another troubling aspect of the conflict.“Why has there not been a concerted battle since nearly theoutset of the war, just small skirmishes?” a veteran aid officialasked in 1998. Whispers of collusion between government andRUF troops that had begun to circulate in the mid-1990s wereconfirmed by the May 1997 coup staged by mutinous nationalarmy soldiers calling themselves the Armed Forces Revolu-tionary Council. The AFRC had received support from theRUF, which, for the first time, left the forests and enteredFreetown. Even after the two groups were chased back into theforests by the Nigerian-led ECOMOG force in January andFebruary 1998, talk of collusion persisted. Some observers inthe aid community found it difficult to believe that the RUFand AFRC could have survived the ECOMOG attack. Oneconcluded that “ECOMOG [had] allowed the AFRC/RUF

14

convoy to escape into the forest” during their retreat fromFreetown.

The central thread tying possibly colluding national armyand ECOMOG troops, mercenaries, civil defense units, multi-national companies, and regional leaders such as Liberia’sTaylor and Burkina Faso’s Compaore to the RUF was theinsurgents’ control over diamond mines.13 Together with thebrutality, the lure of diamonds and the enormous wealth andpower they represented helped make the violence seem, asone Sierra Leonean author put it, more like “civil chaos” thanoutright war.14 However repugnant, the RUF strategy ofextreme child exploitation, terror, brutality, depopulation,collusion, and diamonds has proved extraordinarily success-ful.

In the view of one observer, the terms of the third peacetreaty that Foday Sankoh’s RUF signed with President Kabbah’sSierra Leonean government “amounted to a victory for Mr.Sankoh.”15 In addition to receiving amnesty for RUF andAFRC leaders against prosecution for crimes against human-ity—a provision denounced by international human rightsgroups—Sankoh received a government position equivalentto vice president and also became the head of a new commis-sion on mineral resources and national reconstruction. Even inthe tenuous post-war period, the RUF’s connections to thediamond trade appears to have been preserved, the govern-ment of Ahmed Tejan Kabbah remains weak, there is no viablenational army and instability and violence continue to regu-larly hamper aid operations. Despite local and internationalcondemnation of their tactics and a concerted military re-sponse to their aggression, the RUF continues to dominate theSierra Leonean landscape.

Rwanda: Refugees and Killers in Camps

The ferocity and staggering efficiency of the Rwandangenocide has spawned a cottage industry of publications onthe darkest chapter of central African violence.16 In retrospect,the horror of the human slaughter overshadowed the interna-tional community’s numbed response to the 100-day geno-cide. Unlike the videotape of the fallen American soldier

15

16

dragged through the streets of Mogadishu, Somalia, whichwas broadcast endlessly on news shows across the globe,Rwanda’s tragedy could not be captured by a single dreadfulimage. On April 6, 1994, when the violence began, only twointernational journalists were present in Rwanda. In newspa-pers and on television, evidence of extraordinary atrocitiesand massive brutality in Rwanda mixed with widespreadindifference in the UN and many governments. Murders anddismemberment took place in full view of UN contingents andmedia cameras. Corpses were piled high in churches or dumpedinto rivers. As the days and weeks passed, the spectacle ofhorror seemed only to get worse.

On April 28, three weeks into the genocide, an uncompre-hending world watched nearly 250,000 Rwandans cross theRusumo Bridge and enter Tanzania. The size of the movementon a single day was unprecedented. The media attention paidto this stunning scene was the first of what a UNHCR officiallater called “a set of unusual circumstances” that soon ben-efited the humanitarian coordination effort. The press was inNgara en route from covering elections in South Africa, theofficial continued. “So when they heard of a quarter millionpeople crossing in a day, they came.” In their wake wereplaneloads of donor government officials, eager to support thehumanitarian response in an unusually coordinated fashion.

The Rwandan refugees who arrived in Tanzania duringthe initial influx were put on a hill near the border calledBenaco, a site that UNHCR had previously identified for about60,000 Burundian refugees. The sheer size of the populationpresented problems of unprecedented magnitude. Routinetasks such as digging latrines, providing potable water, andsupervising firewood collection became major operations.The logistics required to deliver large quantities of food andsupplies to Ngara District, a particularly remote section of EastAfrica, were complicated and difficult.

The attention paid to Rwandan refugees in Tanzaniaturned out to be short-lived. In July, 850,000 Rwandans en-tered Goma, Zaire, in what has been called "the largest andmost sudden population movement in modern history.”17

This second Rwandan refugee influx radically changed howNgara was viewed. From mid-1994 until late 1996, the more

17

than half million Rwandan refugees in Tanzania were over-shadowed by the presence of a million-plus Rwandan refu-gees in Zaire. The refugees in Tanzania eventually became akind of enormous backwater. While Goma attracted ongoinginternational attention, Ngara’s refugee camps, well-coordi-nated and orderly by comparison, did not.

Tracing the origins and experiences of the Rwandans wholater occupied the refugee camps near Ngara helps shed lighton the perplexing challenges they presented. Most came fromByumba Prefecture in Rwanda’s northeast and Kibungo Pre-fecture in the southeast. The refugees from Byumba wereamong the first Rwandans internally displaced by the civilwar that began in their prefecture in October 1990, three anda half years prior to the genocide. By early 1992, more than300,000 Rwandans, most of them from Byumba, were in campsfor internally displaced persons (IDPs).18 The civil war periodleading up to the 1994 genocide continually drove the IDPssouthwards. Some later reported to have lived in as many aseight IDP camps before becoming refugees.19

By the time the genocide began, many of the IDPs fromByumba were residing in Kibungo Prefecture, a major site forkillings during the genocide’s early stages: “The vast majoritywho had fled to Tanzania from Kibungo préfecture (300,000)were Hutu, and they were not fleeing massacres, as theirleaders tried to pretend, but on the contrary they were thepeople who had just killed between 25,000 and 50,000 Tutsi ineastern Rwanda and were fleeing to escape what they feltwould be the vengeance of the advancing RPF [RwandanPatriotic Front] forces.20

The population that crossed into Ngara District was thusquite unlike refugees that aid and Tanzanian governmentofficials were accustomed to working with. They were plaguedby extreme fear and trauma. They were violent. At night,murder, gang warfare and occasionally even militia trainingtook place while girls and women avoided going to latrines forfear of being raped. The population contained genocidaires,Rwandans who had orchestrated or carried out the killings,including members of the notorious Interahamwe militia. Whileofficials debated the likely percentages of guilty killers andinnocent refugees, the exact ratio remained unknown. Some

18

Rwandans, swept up in the tumult and terror of genocide andflight, were faced with “incredibly complex moral and socialsituations”21 that led one observer to label them “innocentmurderers.”22 Still, Prunier argued that “the main agents of thegenocide were ordinary peasants,”23 though it should also beremembered that much of the Rwandan refugee populationwas comprised of children who had witnessed killings but didnot commit them.

Refugees were also, on the whole, deeply suspicious offoreigners. They viewed Tanzanian authorities as allies of theRPF leaders who had assumed power in Rwanda followingthe genocide. Refugees, and those from Byumba in particular,also believed that the international emphasis on the genocideoverlooked the tragedies that had befallen them in the civilwar period. During interviews in 1996, Rwandan refugeesexpressed their conviction that UNHCR and other foreignofficials viewed them as guilty of genocide rather than victimsof war.24

The refugees’ feelings of vulnerability, fear, and isolationproved unshakable. UNHCR’s attempts at voluntary repatria-tion, a campaign that began soon after the refugees arrived inTanzania and extended until just before their return late in1996, met with scant success while the agency’s credibilityamong refugees steadily declined. Indeed, relations betweenaid agency officials and refugees often seemed adversarial. Inthe eyes of Tanzanian authorities, the refugees graduallybecame unwelcome guests. They had entered a forested areain 1994 and within a matter of months left it nearly bereft oftrees. The Tanzanians increasingly viewed the refugees as asevere security risk as well. Over time, they began to assertcontrol by limiting refugee movements, patrolling the campsand the surrounding areas, and, by the end of 1996, driving therefugees back into Rwanda.

The emerging distance between most refugees and boththe Tanzanian authorities and the international humanitariancommunity created a vacuum that Rwandan refugee leaderseagerly filled. Veteran politicians who arrived in Tanzaniawith significant proportions of their constituencies, the refu-gee leaders directly supported coordination systems thatUNHCR and its NGO partners devised. Working with these

19

leaders, however, proved to be a double-edged sword. Theirability to organize the refugee majority greatly facilitatedlarge-scale humanitarian activities such as food distributions.But there was a reason for the remarkable refugee obedienceto their leaders’ wishes. Many if not all of the major refugeeleaders were suspected mass murderers during the genocide.Unlike many of the refugee leaders in Zaire, who arrived inmilitary uniform with arms and thousands of identifiablehenchmen, fugitives from justice in the Tanzanian campsthreatened bona fide Rwandan refugees in a more clandestinefashion. Threats of retribution against any who may havewished to identify the killers in their midst were less direct andless public, and the perpetrators of violence were more diffi-cult to trace. Still, in Tanzania as well as Zaire the refugeeleaders were fearsome; the leaders that aid agencies eventu-ally worked with in both countries did not represent refugeecommunities as much as rule over them. Thus did “the im-perative of feeding Rwandan refugees [conflict] with the ne-cessity of bringing to justice those among them who had beenthe architects and perpetrators of genocide.”25

The perplexing moral and logistical challenges thatRwandan refugee leaders presented to the humanitarian com-munity dramatized the coordination challenges faced in Si-erra Leone and Ngara. In both cases, aid agencies, with the UNat the forefront, were supposed to mount and coordinatemajor activities. Their expertise in specific tasks and theirability to attract funds from donor governments allowed themto maintain their roles. Their commanding presence also lim-ited the contributions of national governments andmarginalized input from local NGOs. But the rules of thehumanitarian game could change at any time. Entities such asthe Rwandan refugee leaders, the Tanzanian authorities, or, inSierra Leone, the RUF held the power to alter the dynamics ofhumanitarian activities. At various times they all invoked thatpower. While in charge of coordinating humanitarian activi-ties, the roles of the internationals were largely reactive toevents carried out by others.

21

CHAPTER 3

COORDINATION AMONG ORGANIZATIONS

The starting point for discussing the dynamics of coordi-nation in Sierra Leone and Ngara involves relations amonghumanitarian agencies. Particularly during emergencies, aidofficials tend to equate coordination with productive interper-sonal relationships: that is, whether or not agency officials getalong. If they do, the thinking goes, coordination flourishes; ifthey do not, it flounders.

Relationships between agency officials, however, are onlyone element in coordination, although an important one.Conflicting organizational mandates and varying readings ofpolitical context are other factors that influence institutionalrelations. A review of the nature and evolution of institutionalrelationships over time in this chapter sets the stage for analy-sis of the functional and programmatic dimensions of coordi-nation in the following two chapters.

Sierra Leone: Coordination for a Country

This section examines the structuring of relationshipsamong three UN entities charged with coordinating the inter-national humanitarian response in Sierra Leone, as well astheir relations with international NGOs and national govern-ment bodies. It then reviews a particularly illuminating eventthat exposed and personalized sharp philosophical differ-ences among UN and NGO agencies. It closes with a look athow donors managed to maintain good relations among them-selves and coordinate assistance across a divided communityof humanitarian agencies.

Three Structural Problems

Humanitarian coordination in Sierra Leone illuminateshow personality clashes among key actors can obscure under-lying structural difficulties and discourage creativity andinnovation. By late 1998, when field research was carried outin Sierra Leone for this study, the head of UNDP in that

22

country, Elizabeth Lwanga, had become a focal point for whatwas good or bad about humanitarian coordination there,depending on viewpoint. The controversy that surroundedher pointed to a weakness in the overall coordination schemein general and three structural problems in particular. Thesewere: UNDP’s inherently problematic position as humanitar-ian coordinator, a logjam at the top of the UN’s coordinationstructure, and contradictory approaches regarding how inter-national and national actors should work together.

UNDP’s Mandate. Though the government’s role as thetitular head of coordination was not insignificant, the powerand ability to orchestrate international humanitarian coordi-nation lay with the United Nations. UN agencies had thenecessary resources, international recognition and experi-ence. Many international participants also believed that itwould have been inappropriate for the national governmentto coordinate humanitarian action because it was a partisanplayer in the civil war that had caused the emergency. Butbeyond this debate over principles lay other problems for theSierra Leonean government. War threatened its sovereignty,undermined its domestic economic and political resourcebase, and kept relationships with donor governments subjectto negotiation. The authorities had little alternative but toaccept the operational UN hegemony. Indeed, when the RUFand AFRC twice invaded Freetown, the authorities in Freetownwere understandably more concerned with their own survivalthan with coordinating relief.

As the RUF slowly infiltrated forests in portions of easternand northern Sierra Leone during the early stages of the war,international responses to changes in the countryside varied.A split arose in 1993 when many non-UN agencies with long-standing presence in the country such as the United StatesAgency for International Development (USAID) and CatholicRelief Services (CRS) began to shift from development toemergency work. At the same time, the UNDP resident repre-sentative (ResRep), who served as the resident coordinator(RC) of all UN activities there, was reluctant to label thesituation a humanitarian emergency. Speaking in 1998, aUNDP official conceded that “the emergency started before itwas called an emergency” but cited rebel attacks on Kabala in

23

November 1994 as “the signal for the [UN’s] emergency re-sponse.” The UN awarded the UNDP ResRep a second title:humanitarian coordinator.1 This decision sowed the seeds ofdivision and conflict.

UNDP is a development agency with an Emergency Re-sponse Division (ERD). Unlike most aid agencies, it assumes adevelopment-oriented approach even during conflicts and inwar zones. “All our coordination has been around strengthen-ing the government” of Sierra Leone, Ms. Lwanga explained inOctober 1998. War did not disrupt UNDP policies that hadbeen developed during peacetime. “The only way to deal with[the government] is to work with them,” she continued, add-ing that as ResRep, her role “is to harmonize what all UNagencies do operationally” and to facilitate their cooperationwith the host political authorities. For UNDP to assume theHC role was appropriate, she concluded, simply because, asthe ranking UN aid official on the ground, “the residentcoordinator should be the humanitarian coordinator.”

Although such declarations attracted criticism from otheraid actors, Ms. Lwanga’s position was entirely consistent withher agency’s mission and goals. A senior ERD official inUNDP’s New York headquarters explained the logic of havingUNDP’s ResRep in a country at war serve as both the RC andthe HC. “The two [titles] are linked up in one person to builda bridge between humanitarian and development work,” heexplained. Wearing both hats is important because “humani-tarian agencies are not committed to linking with the develop-ment-oriented” agencies. The official saw no structural con-tradiction; he believed the two positions were complemen-tary.2

In UNDP’s view, war and humanitarian emergencies donot destroy development activities: they only upstage them. Ifdevelopment processes are endangered during emergencies,then UNDP should support development initiatives that are“curative.” “Being mandated to deal with the entire develop-ment process,” explains one UNDP document, “UNDP’s pri-mary concern has been to ensure, as the conflict unfolds, thatbridges are built between the humanitarian operations andfuture sustainable human development assistance.”3

While UNDP views itself as an advocate for development

24

and a supporter and enabler of government institutions, itsmission statement also explains that the agency is “politicallyneutral” and “impartial.” Given that Sierra Leone’s govern-ment was directly involved in the civil war, the two roles—government supporter and neutral development agent—seemcontradictory, if not schizophrenic. But UNDP officials con-sidered their approach during periods of war and peace inSierra Leone as both consistent and appropriate. During civilstrife as in peacetime, one UNDP official in Sierra Leone noted,“humanitarian workers have to deal with the local [govern-ment] authority in order to serve the people.” Deliveringhumanitarian relief requires close relations with the govern-ment and careful distance from opposition groups, the officialcontinued, because “you can’t negotiate with [the RUF rebels]and so alienate the government.”

Few if any other international agencies in Sierra Leoneagreed with UNDP’s approach. A high-level official of an-other UN agency in Freetown offered a common critique.“UNDP is mandated to help the government with its develop-mental objectives,” he explained, “so it’s incumbent uponUNDP to side with the government.” Although UN agenciessuch as “the United Nations Children's Fund (UNICEF) andthe World Food Programme (WFP) know how to providehumanitarian assistance without taking sides” in civil con-flicts, “UNDP’s mandate precludes it.” While the neutrality ofUNICEF, WFP, and other UN organizations at the level of aidoperations is surely debatable, the official argued that, unlikea number of UN agencies with considerable experience inhumanitarian emergencies, UNDP “is a complete novice.” Asa result, UNDP’s role as humanitarian coordinator was “thebiggest obstacle in the UN to addressing crises” effectively asa system.4

UNDP did not take such criticism sitting down. “Some[international] agencies don’t give a damn about Sierra Leone,”a senior UNDP official in Sierra Leone charged, because “theyjust distribute food without caring about the political aspect ofthe crisis.” In many respects, the humanitarian effort for SierraLeone was severely weakened because key coordination fig-ures had daggers drawn against each other. A controversialepisode in nearby Guinea that illuminated the seriousness ofthis divide will be described shortly.

25

The UN Logjam. UN coordination responsibilities in SierraLeone gradually spread across three separate bodies. At theoutset of the emergency, UNDP led the coordination effort. In1996, a second agency arrived, the Humanitarian AssistanceCoordination Unit, or HACU, the local office of the UnitedNations’ Office for the Coordination of Humanitarian Affairs(or OCHA, known before January 1998 as DHA, the Depart-ment of Humanitarian Affairs). HACU’s entrance divided UNcoordination into two levels. While Ms. Lwanga remained UNhumanitarian coordinator, HACU became informal facilitatorof a coordinated humanitarian response that extended wellbeyond the UN family of agencies.

Though it was generally accepted that HACU wouldcoordinate activities under UNDP’s guidance, perceptions ofthe relationship between these two agencies differed. UNDPofficials saw their relationship with other humanitarian agen-cies in hierarchical terms. Reflecting her view of UNDP as theinternational agency in charge of humanitarian coordination,Ms. Lwanga explained that “HACU is the executing body forcarrying out coordination” that UNDP, as humanitarian coor-dinator, supervises. But a HACU official described his agency’srole in horizontal terms, as a “balance between the UN agen-cies and the others.” Compared to UNDP’s primary focus oncoordination between the host government and UN agencies,he saw HACU’s mission as much broader and more inclusive.For HACU, the “coordination players” included not just UNand host government agencies but “other parties to the con-flict: local and international NGOs, community-based organi-zations such as farmers’ and women’s groups, tribal and othertraditional leaders, and religious leaders.”

Emblematic of conflicting conceptions of coordinationwas the fact that UNDP, while delegating coordination forhumanitarian activities to HACU, also carried out its owndevelopment activities. Asked to explain why UNDP’s devel-opment work was not coordinated with humanitarian activi-ties, a HACU official responded, with considerable frustra-tion, “because UNDP does development” and HACU doesnot. The implication was that UNDP viewed its dual roles indevelopment and humanitarian work as not in conflict witheach other. UNDP development activities, however, under-

26

mined its credibility as humanitarian coordinator. Though itsdevelopment and humanitarian work may have seemed anappropriate mix of “curative development” and humanitar-ian action to UNDP, the approach seemed intentionally divi-sive to many other international actors. In their eyes, govern-ment-oriented UNDP development work was directly at oddswith its humanitarian coordination responsibilities, whichshould have emanated from the principle of impartiality. Ms.Lwanga’s leadership style as HC, together with UNDP’s dis-tinct approach to humanitarian response, combined to isolateher and her agency from many if not most other internationalofficials, NGOs in particular, while allowing her to maintainclose relations with Sierra Leonean authorities.

The emergence of a third UN entry atop the coordinationstructure, the UN Observer Mission in Sierra Leone(UNOMSIL), complicated coordination further still. Estab-lished by the UN Security Council on July 13, 1998, UNOMSIL’soffice was located not with UNDP and HACU in UN House inFreetown’s city center, but instead on a hillside overlookingthe town. To many internationals, this symbolized the sepa-rate position that this new UN body assumed in the coordina-tion structure.

UNOMSIL was originally headed by Francis Okello, theUN Special Representative of the Secretary General for SierraLeone. Its staff had a large military component but also in-cluded human rights monitors from the UN High Commis-sioner for Human Rights. While UNOMSIL was charged withmonitoring the military and security situation and reportingon human rights violations, its coordination role was lessclear. A senior UNOMSIL officer explained that the entranceof UNOMSIL into Sierra Leone did not upset the existingcoordination structure: it simply assumed the position at thetop of it. “Once a UN SRSG arrives,” the officer explained, “heautomatically becomes the UN family head. He coordinatestheir activities without violating the mandate of other UNagencies.” The SRSG’s family head role was seen to comple-ment and not interfere with existing UN activities because“the SRSG doesn’t get involved with local issues.”

By the fall of 1998, however, relations between UNDP andUNOMSIL had grown difficult. Lines of authority were not

27

clear. A UN official in Sierra Leone explained that this was tobe expected, since the two groups derived their authority fromdifferent UN bodies. “The SRSG is endorsed by the UN Secu-rity Council,” the official explained, “while the UNDP’s resi-dent representative is endorsed by the UN General Assem-bly.” UN policy, the official continued, positions the SRSG asthe highest authority, while the ResRep/HC is charged withcarrying out specific coordination tasks in consultation withthe SRSG.

In the view of many NGOs, the UNDP/UNOMSIL clashseemed increasingly irrelevant. As one veteran NGO officialsaid, “The coordination structures have developed in spite ofthe Resident Representative/Humanitarian Coordinator, andwhen UNOMSIL came along, the coordination structure wasalready in place.” UNDP’s HC, the official continued, “hadtried to prevent the NGOs’ participation in coordination”planning and had lost. Indeed, several NGOs noted that theyeventually ignored UNDP and worked with HACU to de-velop a more inclusive coordination strategy even as UNDPand UNOMSIL, whom they perceived as pro-government anddetached from the larger humanitarian arena, followed aseparate course.

In sum, three institutional UN players, all of them viewingthemselves as coordination leaders but perceived differentlyby other humanitarian actors, complicated but did not entirelyundercut the exercise of coordination on the ground. By 1998,most international NGOs and International Committee of theRed Cross (ICRC) officials seemed to view UNDP, UNOMSIL,and the Sierra Leonean government essentially as a singleunit. Such an alliance, whether real or perceived, clearlyundermined the credibility of UNOMSIL and UNDP as coor-dination agents. Opinions regarding HACU were more di-vided, but even those suspicious of its coordination role foundits officials accessible. HACU may never have been the UN’sanointed coordination leader, but its officials were the mosteffective coordinators. While holding the UN’s lowest coordi-nation rung, HACU worked with a diverse range of humani-tarian actors and, in the eyes of most observers, usually suc-ceeded.

The National Government’s Role. In Sierra Leone, coordina-

28

tion structures often incorporated the official whom interna-tional agency officials preferred to deal with rather than theranking official from the appropriate government ministry.Officials defended this approach by arguing that governmentcounterparts were frequently inept or corrupt, or both. Butignoring formal government coordination relationships inpursuit of solutions to immediate crises may have created newand lasting problems.

“The ‘good’ people [in the government] are earmarked aspoints of entry,” said one NGO official. “Then everybody goesto [them] with everything, and soon they can’t do anythingproperly. It’s a potential vicious cycle, and [soon] their [gov-ernment] peers will call them a ‘donor baby’—beholden toforeign interests [and] not a true patriot.” By selecting whichgovernment officials to work with, international agencies senta message to the Freetown authorities about power relations.They defended this in the name of expediency—they weretrying to get things done fast to save lives. From thegovernment’s perspective, however, the message was differ-ent: international agencies, not the government, were callingthe shots in their country.

What role in coordination should the government play ina country that is at war and wracked by a humanitarianemergency? In Sierra Leone, the question was never resolved.Unlike failed states such as Somalia in the early 1990s, the civilconflict in Sierra Leone weakened but never destroyed thestate. Even when the AFRC and RUF overwhelmed Freetownand sent President Kabbah into exile, his government retainedinternational recognition and remained a player in domesticpolitics. For many international NGOs and donors in SierraLeone, the tension between preserving the mantle of neutral-ity to facilitate access to all victims of conflict and establishingsome sort of working relationship with the government proveddifficult to resolve. This tension became a persistent theme incoordination politics there. It frustrated efforts to find com-mon ways of proceeding and underscored the need to actingin concerted fashion.

International agencies with a greater attachment to theprinciples of neutrality and impartiality perceived UNDP’sprimary strength throughout the crisis years—its mainte-

29

nance of good relations with the Sierra Leonean govern-ment—as a weakness. This perception was particularly strongon the part of NGOs, many of whom antagonized thegovernment’s central body for humanitarian coordination,the National Commission for Reconstruction, Resettlementand Rehabilitation (NCRRR). Specially created and chargedwith coordinating assistance to the line ministries (e.g., health,education), the NCRRR sought to increase control over theinternational coordination structure by establishing itself asHACU’s counterpart.

NCRRR was supposed to receive findings from all coordi-nation groups and advise the government’s Inter MinisterialCommittee on humanitarian policy.5 In reality, however, itscoordination role was less clear and its relations with othergovernment actors frequently strained. A UN official notedthat even though the NCRRR may have had its own budget todispense to line ministries, these ministries opposed its role aschief coordinator of the government’s humanitarian work. Asa result, the Inter Ministerial Committee, which the NCRRRchaired, rarely met. The official explained that “the NCRRRneeds support to get the Inter Ministerial Committee going,”and added that UNDP was among the barriers, for it “focuseson the [line ministries] to support them against the NCRRR.”

This last comment illuminated both the complexity of thegovernment’s involvement and its largely reactive role.UNDP’s apparent alignment with the line ministries con-trasted with HACU’s relationship with the NCRRR. At thesame time, many international NGOs and donors kept theirdistance from all but a few government figures. And whileHACU seemed able to maintain relations with more humani-tarian actors than any other agency—a notable and importantstrength—a Sierra Leonean with experience working for boththe UN and the government offered a persuasive explanationfor its edge over the NCRRR. HACU, he said, took the lead incoordination and NCRRR did not because HACU had “moreclout” and far better relations with the powerful, well-orga-nized, and well-funded NGOs.

NCRRR officials reserved a great deal of criticism forNGOs. In their view, NGOs trampled on their authority andgovernment sovereignty. As one official commented, “The

30

international NGOs go wherever they want [and] we can’tcoordinate them. This is a burning issue with us. Sometimes Ifind NGOs just putting up health centers or digging a well andthat is it. Some NGOs send us monthly reports but don’t saywhat money they have left. We don’t know their future plans.But the problem of everything is that NGOs go to the peoplethey choose in the government. The NGOs do what theywant.”

Other government officials also viewed NGOs as threatsto their sovereignty. They were contemptuous of them forworking outside of the government and focusing on humani-tarian activities instead of tackling, as one NCRRR official saidin late 1998, “what’s needed now—development work.” Oneinterview with an official concluded with his assessment that“maybe we have to do like the Rwandans and Ethiopians” andexpel some NGOs from the country. Significantly, even thoughdonor agencies such as the European Community Humanitar-ian Office and USAID’s Office of Foreign Disaster Assistance(OFDA) maintained their distance from government officials,they were spared the criticism directed at NGOs, perhapsbecause the government continued to seek access to theirfunds.

Relationships between international and national officialsilluminated the impact of selectivity on coordination. TheNGO emphasis on efficient service delivery enraged manygovernment officials while donors, in turn, funded such groupsbecause of their ability to “get things done.” As ECHO andOFDA officials explained, “We [jointly] fund directly to NGOs”because they are “the big guys, the professionals.” This atti-tude among donors may have contributed to governmentresentment. Meanwhile, the government and local NGOsreceived comparatively little outside funding during the waryears.

The collective message from the international agencieswas that they controlled the coordination cards and attachedhigher priority to addressing human need than to respectingsovereignty. The issue, of course, is built into humanitarianaction in settings of civil strife. But the question of whether theagencies would be able to retain their coordination cardswould subsequently surface in Ngara. There, too, successful

31

humanitarian coordination eventually clashed with aspects ofnational sovereignty.

A Watershed Event

Politicization often characterizes the application of sanc-tions and crossborder assistance during civil wars, and SierraLeone was no exception. After President Kabbah’s govern-ment fled into exile following the rebel attack on Freetown inMay 1997, the humanitarian community split into two groups.All UN agencies and many NGOs set up operational bases inConakry, Guinea, as did the exiled Kabbah administration.The International Committee of the Red Cross and severalEuropean NGOs, by contrast, continued to base their opera-tions inside Sierra Leone. One effort to coordinate activitiesbetween the two groups arose following the formation of theHumanitarian Exemptions Committee. The story of thiscommittee’s work illuminates the difficulties of maintaining aunified coordination structure when key actors are deeplydivided against each other.

The divide within the humanitarian community came tobe seen as a dispute between a high-profile group, includingHC Lwanga, SRSG Okello, President Kabbah, and PeterPenfold, British High Commissioner for Sierra Leone on onehand, and the ICRC-European NGO group on the other. Theoperational reality was more complex. A number of AmericanNGOs such as World Vision and Catholic Relief Services (CRS)set up bases in Conakry while continuing to manage opera-tions inside Sierra Leone with local staff, with some of theirstaff eventually commuting to Freetown from Conakry.

ECHO field officers conducted regular field visits intoSierra Leone while their OFDA counterparts, and all other U.S.government staff, remained in Guinea. Even some UN agen-cies such as UNICEF managed to maintain a measure ofoperational activity inside Sierra Leone, even though UNinternational staff were restricted from entering the country.Members of this intermediate group, which comprised themajority of the international aid community, straddled thegrowing separation between the two polarized groups.

From the outset, members of the two main groups per-

32

ceived in radically different ways the basic issue of danger tohumanitarian operations. Members of the ICRC/EuropeanNGO contingent based inside Sierra Leone emphasized thehigh level of safety in rural areas, reporting that their ability todeliver humanitarian assistance to areas outside the capitalwas better than at any other time since the outset of theemergency in 1993-1994. They claimed that this was due to theshift to the capital of RUF forces, which had previously pa-trolled inland forests. International staff from NGOs whowere based in Conakry but visited Sierra Leone generallysupported this view.

Many in Conakry, however, had a different perspective.UN agency officials, some press reports, the ousted SierraLeonean regime, officials of the Economic Community of WestAfrican States (ECOWAS) and its Monitoring Group(ECOMOG), and others maintained that Sierra Leone was fartoo dangerous for relief operations. They focused their atten-tion not on rural areas but on Freetown, where the juntaregime was based. Many also supported the internationalsanctions against the junta rulers.

Eventually, the ousted president and his supporters be-came hostile to the European NGOs and the ICRC, which hadnever left Sierra Leone. In their view, delivery of aid inside thecountry supported the junta. Yet Kabbah and his supportersdid not condemn all agencies delivering aid inside SierraLeone, singling out only those who retained their countryheadquarters inside Sierra Leone (ICRC and European NGOs)as “Junta NGOs.” NGOs such as CRS, World Vision, andCARE were spared such criticism.

Politics overwhelmed humanitarianism in Conakry. Afterthe AFRC-RUF rulers and the Kabbah government signed theConakry Accords in the fall of 1997, ECOWAS was to super-vise a staged, six-month peace plan. Although ECOWAS hadinvoked sanctions against the junta during this period (withUN support ), the peace plan also called for “a mechanism... tofacilitate the flows of humanitarian assistance” into SierraLeone.6 The Humanitarian Exemptions Committee becamethis mechanism.

The committee consisted of representatives of UN agen-cies and international NGOs, along with observers from

33

ECOWAS and the governments of Guinea and Sierra Leone.Its task was to review requests that certain humanitarian itemsbe exempted from the sanctions levied by ECOWAS and tofacilitate their distribution within Sierra Leone. ECOWAS,HACU, and the SRSG were considered the major players onthe committee, along with Ms. Lwanga, the UNDP ResRepand HC, who served as committee chair. Ms. Lwanga laterreported that the committee approved all four exemptionrequests that it received, a widely accepted fact. The disputethat emerged focused on those who prevented humanitarianrelief from crossing into Sierra Leone.

Some UN officials blamed ECOMOG. It had impededdeliveries, one stated. “ECOMOG placed monitors at theborder to clear the goods, but they had few people to get itdone.” HACU’s effort to expedite border clearance by payingthe monitors’ salaries had no effect. Truck convoys loadedwith food at the border never crossed into Sierra Leone. To itsmany detractors, the committee’s failure to coordinate reliefwas intentional and politically motivated. In their view, block-ing aid to the country, then under AFRC-RUF domination,supported President Kabbah’s belief that “rice was being usedin Sierra Leone as a ‘weapon of war.’”7 Food delivery wassupposed to wait until Kabbah returned, even if it meant thatcivilians would starve to death in the meantime, as untoldthousands of Sierra Leoneans apparently did.

The committee’s failure had a lasting impact on coordina-tion efforts. Late in 1998 in Freetown, a year after it had beenformed, the Humanitarian Exemptions Committee continuedto arouse passionate arguments over what had happened andwho was behind it. One of many aid officials highly critical ofthe collective stance of the SRSG, President Kabbah, and Ms.Lwanga on this issue argued that “their strategy worked:holding back humanitarian aid helped Kabbah get back intoSierra Leone.” But in the process, the UN’s lead role on theexemptions committee had seriously damaged its credibilityby positioning the world body as “blocking its own [humani-tarian] operations [and] preventing humanitarian assistancefrom crossing the border.” Although Ms. Lwanga and othersdisputed this version of events, UNDP’s and the SRSG’ssupport of the Kabbah regime, both in Conakry and following

34

Kabbah’s return to Freetown in early 1998, confirmed suspi-cions that political objectives outweighed humanitarian goals.

On the opposing side were the ICRC and many NGOs whoconsidered the delivery of humanitarian aid to all civiliansaffected by the war their overriding goal. “We try not so muchto be neutral,” an NGO official explained, “but to be impartial.How can you reach the people on the other side? By being asimpartial as possible.” An ICRC official stated that his orga-nization sought neutrality, which he defined as “a position tobe in support of the suffering.” Although supporting thesuffering behind RUF lines became virtually impossible oncewar was started again in 1998, the ICRC official maintained,together with officials from European NGOs, that “if one[humanitarian] group is seen as taking sides with the govern-ment, then you don’t have a chance of getting to the sufferingin rebel-held territories.” This perspective interpreted thegovernment-friendly policies of UNDP and other like-mindedagencies as creating the impression that all humanitarian aidwas partisan and anti-RUF, thus blocking efforts to aid civil-ians behind RUF lines.

While the ICRC and some NGOs strained to be seen asneutral and/or impartial, other humanitarian agencies didnot. This second group’s main objective was supporting ademocratically elected government against the RUF, whichcommitted human rights atrocities. As one NGO official ob-served, the issue is “black and white. Some [RUF soldiers] arecommitting human rights abuses and some [ECOMOG andgovernment soldiers] aren’t. It’s a democratically elected gov-ernment vs. human rights abusers.” A UN official wentfurther still, expressing the view that “the refusal [of otherhumanitarian actors] to take sides, to acknowledge who’sright and who’s wrong, will perpetuate the crisis.” ICRC andEuropean NGO officials responded by arguing that soldierson the government side had also committed atrocities, albeit ata far lower rate than the RUF. In their view, their allegiance toa basic humanitarian principle—neutrality—was being mis-understood as sympathy for the RUF cause because the deliv-ery of humanitarian assistance had become so politicized.

Although politicization often characterizes the applica-tion of sanctions and crossborder assistance during civil wars,

35

proponents on both sides of this chasm shared one commonbelief: that the other side was making their work much moredifficult. The Humanitarian Exemptions Committee imbro-glio had brought the debate about democracy vs. impartial-ity/neutrality into the open and damaged interagency rela-tionships irreparably. Despite this, however, coordination atthe sector level proceeded. The reasons for this seemed clear:the most politicized humanitarian agencies—UNDP andUNOMSIL—were generally not directly involved in sector-level activities, and aiding civilians behind RUF lines re-mained more of a goal than a reality. As a result, even thoughsharp philosophical differences remained within the humani-tarian community, aid to civilians still continued.

Donor Perspectives

In the midst of corrosive disagreements involving virtu-ally the entire international humanitarian community inSierra Leone, the two largest donors—OFDA and ECHO—managed to continue coordinating most of their work witheach other. A number of important elements contributed totheir success in doing so.8 First, both agencies shared the viewthat “both sides will get better results if we coordinate well.”ECHO-OFDA coordination succeeded, they believed, “be-cause we make it work.” Second, their preference for keepingOFDA and ECHO coordination links informal was helpful.“Coordination works,” in Sierra Leone, an OFDA officialobserved, because “nobody makes a big deal about it.”

A third factor was field experience. Field representativesfrom OFDA and ECHO came to their posts with professionalexperience in the region and enjoyed good relations withsupervisors in their respective agencies. Fourth, with col-leagues in the resident European Union and U.S. governmentdelegations in West Africa, they carried out collaborativefield assessments which often resulted in joint funding for agiven NGO. Fifth, ECHO and OFDA representatives regu-larly shared information on NGOs and the general situationin Sierra Leone to prevent duplication and to enhance theirknowledge about NGO operations and program effective-ness. Successful collaboration between OFDA and ECHO

36

also benefited from the fact that their respective donor gov-ernments had generally compatible political policies towardsSierra Leone, which was, in the words of an ECHO official, “aforgotten little place” where OFDA and ECHO representa-tives could “do what we like” without much headquartersinterference.

That the two largest humanitarian donors could worktogether in a cohesive fashion amid such discord was remark-able. OFDA and ECHO officials were generally sympathetic toboth the human rights and humanitarian principles that di-vided the community and managed to maintain relativelygood relations with nearly all of the humanitarian actors. Atthe same time, their assessments of host government manage-ment capacities were generally low and their strong supportfor NGOs exacerbated the NGO conflict with the Sierra Leoneanauthorities. Both were particularly critical of what they viewedas the UN’s coordination failure, which they blamed onUNOMSIL and UNDP. Yet OCHA and ECHO differed ontheir support of UN agencies. While ECHO was reluctant tofund them and could be highly critical of their operations,OFDA funded some UN agencies, notably HACU, while avoid-ing others altogether.

Should Sierra Leone’s civil war have been perceived as anemerging democracy vs. an unpopular opposition guilty ofmassive human rights abuses? If anything can be learned fromthe Sierra Leone case, it is that although humanitarian prin-ciples of neutrality and impartiality suffered amid thepoliticization, humanitarian activities proceeded nonetheless.The urge to provide humanitarian relief regularly upstagedsupport for democracy. Western humanitarian donor agen-cies were among those reluctant to involve the government,even a democratically elected one perceived as fighting againstragtag sadists. As one donor official commented, the reasonhad nothing to do with principles. The central issue was apragmatic one: timing. “Development is about investing forthe future,” the official explained. “Humanitarian aid is aboutsaving lives now.” Thus, the official concluded, “Supportingdevelopment for tomorrow doesn’t make sense when peopleare dying today.”

37

Ngara: Coordination for Refugee Camps

After a quarter million Rwandan refugees entered NgaraDistrict in Tanzania on April 28, 1994, Ngara became an instantmedia and humanitarian sensation. With the world eyeing thehumanitarian response in one of East Africa’s remotest re-gions, the international humanitarian community presum-ably would have struggled to respond with the alacrity andefficiency the situation required. Also facilitating a promptresponse was the fact that the Burundian refugee crisis, whichhad taken place near Ngara, had only recently subsided. Manyaid agencies and supplies were still in the area, and a numberof key actors—international donors and NGOs, the Tanzanianauthorities, and UNHCR—were familiar with each other andprepared to work out a humanitarian response quickly. Thecoordination framework that emerged was what Donini de-scribed as “coordination by command.” It was characterizedby “strong leadership... accompanied by some sort of author-ity, whether carrot or stick.”9 UNOMSIL and especially UNDPhad attempted this sort of approach in Sierra Leone with, atbest, uneven results. But UNHCR’s authoritative role in Ngaraproved far more successful. An analysis of the dynamics ofthat role is the subject of this section.

The Burundian Precursor

Humanitarian officials present in Ngara, Tanzania, fol-lowing the sudden influx of Rwandans into Tanzania haddistinct advantages over their counterparts present for theRwandan influx into Zaire two months later. The Rwandanpopulation in Tanzania was smaller, healthier, and less milita-rized. Tanzanian troops successfully disarmed many of theRwandans. Tanzanian officials also developed good relationswith aid agency officials, something that could not be said ofagency relations with the Zairian government. Even the earthitself was superior: drilling for water or digging latrines wasrelatively easy in Tanzania, but not in the hard volcanic rockof Goma, Zaire.

An often overlooked advantage, however, was the expe-rience gained from a disastrous response to an emergency that

38

had just taken place there. Fleeing their own civil conflict, asmany as 500,000 or more Burundians had crossed into Tanza-nia in 1993 and early 1994 and spontaneously settled in 50 sitesalong the forested, remote Burundi-Tanzania border.10 Arriv-ing during the rainy season, the Burundians were almostimmediately beset by dysentery, measles, and malaria. Theirsituation soon grew even worse. Within weeks, “a refugeepopulation, which had arrived in reasonable health, experi-enced a famine.”11 Though no accurate figures exist, campsurveys found that the combination of disease and malnutri-tion had created morbidity rates “40 times higher than ex-pected.”12 Thousands perished in the forests. One UNHCRofficial lamented that “We lost about 50,000 Burundian chil-dren. This is a shame.”

The humanitarian community’s response was also unfor-tunate. Although the Joint Evaluation of Emergency Assis-tance to Rwanda (JEEAR), a comprehensive study, identifiedthe tendency “for staff of one agency to attribute blame for aproblem to the other agency”13 in the Central Africa regiongenerally, it reported that “the worst example encountered ofsuch blaming-the-other concerned the very high rates of mor-tality experienced by Burundian refugees in Tanzania in late1993.” In this case, “whilst WFP personnel saw the principalcause of mortality being dysentery resulting from UNHCR’stardiness in the provision of adequate water and sanitation,UNHCR personnel saw inadequate supplies of food by WFPas the principal cause.”14

The JEEAR report also noted that “the relative significanceof the logistical difficulties, inadequate resources and manage-ment factors in contributing to the increased mortality [of theBurundian refugees] were difficult to gauge.”15 But every offi-cial interviewed about the Burundian refugee crisis admittedthat the humanitarian response was, as one NGO official presentrecalled, “a screw-up.” No agency seemed to have performedparticularly well, including UNHCR, which received low marksfor the exercise of its coordination role. A UNHCR officialadmitted that there was not good coordination but added thatthere was also no press interest and no donor money. All thiswould change immediately after the Rwandan influx, whichthe official considered “a CNN emergency.”

39

The impact of limited donor funding for the relief effortunexpectedly hastened the end of the Burundian refugeecrisis. WFP had initiated a new “Immediate Response Ac-count” to process donations for the Burundians quickly butthe donor assistance remained “slow and inadequate.”16 AnNGO official explained that, as a result, “WFP didn’t havemoney to make local purchases, while any food arriving fromthe U.S. and Europe would not reach the refugees for weeks ormonths.” Facing starvation in Tanzania and with the situationin their homeland having improved somewhat, all but 60,000of the refugees returned to Burundi by early 1994.

During this period, UNHCR dispatched an EmergencyResponse Team (ERT) to Ngara District, Tanzania. Anticipat-ing a second Burundi refugee influx, its mission was to iden-tify a viable refugee camp site in the area. Eventually it decidedon a hill called Benaco. Equipped with an artificial lake, itwould be presumably far enough from the Burundian borderto stabilize a frightened incoming population. Benaco, as itturned out, would instead host a huge Rwandan refugeepopulation, and its proximity to the Rwandan border wouldcreate serious security problems for international and Tanza-nian officials there. At the same time, the Burundians’ misfor-tune in Tanzania provided a boost to the Rwandans. With thehumanitarian community already in the area and a refugeecamp site established, the foundation for an effective responseto the Rwandan emergency was already in place.

Coordinating the Arrivals

The chaos that followed the movement of 10,000 Rwandansinto Tanzania every hour for a full day in late April, 1994,created what one NGO official called “a zoo.” But five factorsput coordination off to a good start. The first was luck. AUNHCR report later noted that “the response [to the crisis]was accelerated because UNHCR was not only on the ground,but had been present at the border as the influx started.”17

There were also supplies on hand to distribute. “The stocksordered for the Burundian [emergency]—food, computersand so on” which did not arrive in time to aid Burundianrefugees, a UNHCR official recalled, “were [instead] used for

40

Rwandans.” Also, UNHCR’s ERT team immediately beganworking with seasoned field professionals from donor andNGO agencies, holdovers from the Burundian emergency.This led one UNHCR official to highlight the connectionbetween coordination and competence: “At the end of the daythe results are a matter of who is on the ground.”18

Good fortune also came in the form of a developmentproject that happened to be in the area when the Rwandansarrived. “The biggest [immediate] problem was sanitation,”said an NGO official afterward. “UNHCR hired Cogifar, anItalian engineering company stationed in the area to buildroads, to build a trench around the camp to prevent fecalmatter from spreading to the lake. This was credited forpreventing a cholera epidemic.”

The second factor was the development of good relationsamong agency officials present at the time of the Rwandaninflux. A UNHCR Mission Report from Ngara, which charac-terized agency relations generally as “excellent,” attributedmuch of this to the agencies’ shared desire to shed territorialconcerns and work together, saying, “The size of the problemfacing everyone on the ground led to an agreement that wewould work as a group instead of strictly within agencyboundaries to ensure that resources were utilized fully andthat coverage in all life saving areas was as great as possible.”19

Esprit de corps emerged during these early days. An NGOofficial later observed that “NGOs were everywhere” andNgara swarmed with intrusive international reporters. In thissituation, UNHCR took charge. Some NGO officials calledUNHCR officials in Ngara “authoritarian,” and there wasindeed strong individual leadership at UNHCR’s helm duringthat time. But most NGOs were generally forgiving of anyexcesses, particularly at the outset.

The third factor was donor willingness—first from ECHOand eventually from the U.S. government—to test a newfunding mechanism that funneled money through UNHCR.For ECHO, this constituted a new kind of relationship withUNHCR, eliciting high marks from observers and from theJEEAR and World Disasters reports.20 A related innovationwas UNHCR-Ngara’s decision to process NGO grants in thefield. This approach was uncommon in refugee situations,

41

where donors tended to fund NGO implementing agenciesdirectly, with the processing of grants completed in homeoffices. These innovations greatly streamlined decision-mak-ing procedures, sped up the funding process and enhancedUNHCR’s control over field coordination. These two innova-tions will be explored in the following chapter.

A fourth factor advanced ECHO’s relationship withUNHCR even further. ECHO seconded a European Uniondelegate to UNHCR’s staff. Although this unusual situationappeared to work fairly well, it troubled UNHCR. Other aidagencies treated the ECHO official more as a donor than asUNHCR staff, creating a situation that was, in UNHCR’s view,“neither clear nor completely comfortable.”21

The fifth factor was the Tanzanian government’s willing-ness to allow UNHCR to approve each NGO that would workin the camps. In retrospect, this factor may have been mostimportant in creating good coordination and generally satisfac-tory interagency relationships. It allowed UNHCR to establish,in consultation with accredited NGOs, an orderly, cohesive,and well-coordinated system. UNHCR accepted twelve NGOsas partners in the Ngara camps and rejected 40. This was laudednot only by UNHCR but by NGOs and others as well:

The Great Lakes scramble left some askingwhether agencies will ever get their act to-gether. The exception was Tanzania, whereteamwork took the place of the agency over-lap and competition in Zaire, Burundi andRwanda. ECHO and USAID put all fundingthrough UNHCR, which also had full author-ity from the Tanzanian government to ap-prove agencies...successful agencies wereunanimous that cooperation had never beenbetter. Coordination triumphed. Donors weredelighted.22

The authors of the JEEAR report concurred:

The limited number of agencies and person-nel involved in the operation engendered an

42

unusually collaborative approach betweenUNHCR and the NGOs and there is almostuniversal agreement that this was one of thekey factors contributing to the highly-effec-tive initial response in Ngara.23

However, NGOs that were not allowed to participate inNgara, not surprisingly, resented UNHCR’s directive ap-proach.24

The fifth and final factor was generally good workingrelationships with refugee leaders. This aspect of humanitar-ian coordination in Ngara would eventually become contro-versial and a direct constraint on refugee repatriation becausesome of those involved—exactly how many was never clear—were former leaders of the genocidal operation in Rwanda. Asthe emergency wore on, this state of affairs would have anenduring impact on the political and security situation in theregion.

In the short term, however, the leaders’ relationships withaid agency officials yielded results. Their ability to organizerefugees according to commune affiliation in Rwanda allowedfor the orderly distribution of food and for a particularlysuccessful component of the regional unaccompanied childprogram spearheaded by the ICRC. In short, regardless ofunsavory past activities and recurrent security problems inthe camps, the refugee leaders proved to be helpful facilitatorsof specific humanitarian tasks. They behaved better than theirmuch more aggressive colleagues in Rwandan refugee com-munities in Zaire, and this enhanced the ways that aid officialsviewed them. Indeed, the comparison made the Rwandanrefugee leaders in Ngara seem a fairly reasonable and produc-tive group.

Establishing Relationships

UNHCR’s assertive approach to coordination appears tohave had no precedent. UNHCR decided whether an NGOcould work in the refugee camps and became its primaryfunding source. In a significant innovation, UNHCR alsomade NGOs solely responsible for different sectors, including

43

transportation (the Tanganyika Christian Refugee Service, orTCRS) and water (Oxfam UK, later Oxfam Great Britain orOxfam GB). Unlike in Sierra Leone, where NGO relationshipswith donors and other agencies working in their sector (suchas health or food distribution) were equally if not more impor-tant than relationships with the chief coordinating bodies(HACU, UNDP and UNOMSIL), UNHCR mattered most inNgara. If Sierra Leone’s coordination structure was complexand marred by conflicts, Ngara’s structure was streamlinedand marked by clarity.

It was also authoritarian. NGOs in Sierra Leone coulddecide whether or not to deal with UNDP, but no such optionsexisted in Ngara. Ultimately, UNHCR, as one NGO officialobserved, laid down the law that NGOs had to follow. UNHCRadmirers in the NGO community considered its approach anenlightened autocracy. “It was the greatest coordination I’dever seen,” one NGO official recalled. “A bit dictatorial, but Ithink it worked.” UNHCR’s descriptions of the heady days inNgara viewed its assumption of coordination responsibilitiesas successful and transparent. “In general, the atmospherewas one of sharing,” said one UNHCR official present at theonset of the emergency. UNHCR believed that “nothing [was]really confidential,” which meant that UNHCR’s partners inNgara had “open access to UNHCR’s files.” A UNHCRcolleague in Ngara concurred: “For good coordination, wehave to be transparent.”

UNHCR’s early successes were credited primarily to itssub-office head in Ngara, Maureen Connelly. Praised or criti-cized, the fingerprints of “Maureen” on humanitarian coordi-nation were everywhere. Decisions that she made in theUNHCR tent at the outset of the crisis, with a handful of UNand NGO officials in attendance, proved lasting. According tothose present, she listened and then made decisions. For some,her dominance over coordination (in tandem with the sub-office deputy head, Jacques Franquin) was a weakness as wellas a strength. One UN official believed that she and her staffrelied too heavily on personal relationships with other agen-cies. UNHCR’s initial coordination arrangements, the officialremarked, “didn’t leave a structure. Personalities are the greaseto the wheels, but without a [coordination] structure, you just

44

have grease.” As noted in the following chapter, this commentoversimplified the situation. Connelly and her staff set up acoordination framework that continued after they were re-placed. Even so, some officials with other agencies in Ngarafelt that successful coordination relied too heavily on gettingalong with Connelly and her staff. It was not something everyUN and NGO official succeeded at.

One of the agencies that found the approach a strain wasthe Tanganyika Christian Refugee Service. A mainstay inrefugee operations for three decades in Tanzania, TCRS wasby far the most experienced NGO on the scene. But the feudthat began during the Burundian emergency in 1993-1994 andpersisted across the entire Rwandan refugee period (one TCRSofficial called relations with UNHCR a “cold war”) centeredon UNHCR’s belief that, as one UNHCR official contended,TCRS “doesn’t do emergencies [well]. It does long-term settle-ment work.” TCRS, of course, strongly disagreed. “If UNHCRbranded us as not fit for emergencies and not capable foranything but settlement work,” a TCRS official asked, “thenwhy did they so often depend on us” in Ngara? For TCRS, itwas as if UNHCR was setting the agency up to fail by givingit ever-increasing work responsibilities and then taking it totask when shortcomings arose.

Though a number of NGOs felt that UNHCR practicedfavoritism with some NGOs while distancing itself from oth-ers, TCRS officials felt this most keenly. They believed that“UNHCR did not want to see TCRS anywhere near Ngara”even while UNHCR loaded “more and more responsibilitieson TCRS.” A UNHCR official admitted that TCRS didn’t likethe operational role it eventually assumed but continued toinsist that TCRS’ problems stemmed from its lack of experi-ence in emergencies. It was not, the official insisted, anythingpersonal. In the end, TCRS and UNHCR did not mend their riftuntil after the Rwandan crisis had passed. Despite conflictsbetween the two agencies, every TCRS official interviewed forthis report credited UNHCR with developing a “model coor-dination structure” in Ngara.

The charge that UNHCR favored some partners overothers was difficult to confirm, and the major implication ofthis dispute on coordination issues stemmed from TCRS’

45

uniqueness. As the Tanzanian office of the Lutheran WorldFederation, it had an unusually high number of Tanzanianstaff, including those in management positions. One expatri-ate TCRS official explained that “Our one strength is our localstaff.” This set it apart from other NGOs, which were staffedlargely by expatriates at managerial levels and with moreRwandan refugees than Tanzanian nationals in lower staffpositions. Pressure from the Tanzanian government encour-aging aid agencies in Ngara to hire more Tanzanians thus didnot apply to TCRS. As TCRS tells it, however, UNHCR “triedto get TCRS to hire more expatriates” to replace some of theTanzanians in high-level positions, offering to increase itsbudget if they did so. TCRS refused. This allegation was hardto confirm: the disagreement took place after Connelly had leftNgara and the UNHCR officials involved were difficult toidentify and track down. However, the presence of a numberof senior Tanzanians on one NGO staff and few on othersraises the question of whether internationals working in hu-manitarian agencies simply prefer to deal with other expatri-ates. A more serious question underlies this and is applicableto the Sierra Leone case as well: to what degree are interna-tional actors—NGOs, UN agencies and donors alike—com-mitted to building local capacities and institutions?

Post-Crisis Problems

By 1996, after Connelly, Franquin, and other UNHCRofficers had left Ngara and new administrators were in place,the coordination dynamics changed. For officials from the twomajor funders of the humanitarian response in Ngara—theU.S. State Department’s Bureau of Population, Refugees andMigration (BPRM) and ECHO—the nature of the change waspredictable. They had come to expect a drop-off in the qualityof UNHCR personnel between those sent to major crises andthose who replaced them after an emergency had been stabi-lized. An ECHO official, too, believed that UNHCR’s “emer-gency response team is good, but the post-emergency person-nel is not as good.” A U.S. official agreed: “When the [UNHCR]emergency teams leave, the regular UNHCR teams [that re-place them] do not have the same quality.” But not all viewed

46

the changes in the same light. One NGO official commentedthat “given a choice, many NGOs preferred the [later UNHCR]period as they felt better recognized.”

Another change in dynamics arose from relationships thatUNHCR cultivated in the camps with the Rwandan refugeepolitical leadership, most of them commune leaders. OneUNHCR official who regularly met with them in Ngara hailedthis connection as “one of [UNHCR’s] successes. We werecoordinating well with them.” The general feeling seemed tobe that, regardless of the leaders’ contemptible pasts, coordi-nating services with them yielded results. A UNHCR fieldofficer in Ngara said that the “commune leaders are able tosettle problems” and further claimed that having weeklymeetings with the leaders “is how we built trust” with refu-gees. Other agencies also relied on commune leaders to getthings done.

This reliance eventually had serious consequences. AnICRC official illuminated how commune leaders could besimultaneously helpful and obstructive. The primary reasonICRC was able to locate unaccompanied children (UAC) inNgara faster than anywhere else in Central Africa was becausethe Ngara refugee camps “were better organized. They cameas intact communities with the same authorities they had inRwanda.” Coordinating with the commune leaders in chargeof the Rwandan camps helped ICRC and partner NGOs findUAC quickly. But once found, the commune leaders would“pressure the UAC not to go back” to Rwanda to reunite withtheir parents. ICRC officials assumed that the leaders did thisbecause successful repatriations threatened their influenceover refugees. As a result, “80-90 percent of the UAC refusedto return” to Rwanda.

If working with commune leaders could be so problem-atic, were other community leaders, such as church officials,available? UNHCR apparently did not explore the options.Besides, coordinating with the leaders yielded results. SomeNGOs criticized UNHCR efforts to build trust with suspectedgenocidaires, leaders who were directly involved in genocidalactivities. “UNHCR was mainly talking to commune leaders,many of whom were not clean,” one official explained. “Dueto this, it took a long time to separate out the intimidators,

47

which could have been possible right from the beginning evenon the basis of suspicions and available evidence.”

Whether this sort of conjecture was accurate or not, UNHCRwas not alone in this practice. Virtually every aid agencyemployed some Rwandan refugees without knowing whetherthey had been involved in the genocide. In response to ques-tions about the past involvement of refugees, aid workersexpressed denial or resignation. One NGO official commented,“I don’t want to know who is who among my workers.”Another admitted that a selected cadre of Rwandan refugees,hired as NGO workers, were used to represent the views of allrefugees at “all the meetings and workshops.” The officialadded that this was simply because “they were good work-ers.” Whether they might have also been genocidaires was aquestion most aid officials avoided.

Reliance on a small cadre of Rwandan refugees for coordi-nating aid activities with refugee populations while also sus-pecting that they might be criminals had repercussions onrelations between refugees and international officials gener-ally, and on UNHCR in particular. Research in the Rwandanrefugee camps in Ngara in 1996 revealed that many refugeesbelieved that UNHCR suspected them of being guilty untilthey could prove their innocence. Refugees felt trapped andfrustrated by this. “UNHCR has never come to talk to us,” amember of a Catholic refugee congregation complained, aview echoed by others. Some also noted that their relationshipwith UNHCR had changed over time. By 1996, after living formore than two years in refugee camps, one church leader said,“When we arrived in Tanzania, it was possible to tell UNHCRthe truth” about which Rwandans were genocidal killers. “Butnow, if you try…, they will call you a [genocidal killer].”25 Suchfeelings over time became major impediments to UNHCR’srepatriation efforts, examined in Chapter 5.

The evolution of relations between the international com-munity and the Tanzanian government was equally problem-atic. At the outset of the Rwandan emergency, Tanzania waswidely praised for its cooperation. But the primary action thegovernment undertook was to step aside and let UNHCR takethe lead in working out security arrangements for the refugeecamps and accrediting NGOs. Tanzania’s surprising acquies-

48

cence inspired one UNHCR official to comment that “in anideal world, we’d like other governments to do [the samething].” Over time, however, the authorities became con-cerned about the Rwandan refugees as a security threat,environmental danger, and drain on resources and came toresent “being ignored by the international community.”26

A UNHCR official recalled how agencies in Ngara grew“extremely arrogant” in their interactions with Tanzanianofficials. “We weren’t thanking [the Tanzanian government]for all they did: for following our instructions, being open,working with us, responding to our ideas,” the official ex-plained. By late 1996, the government’s relations with thehumanitarian community had altogether changed. Previouscoordination arrangements were now irrelevant. As anotherUNHCR official reflected, “the Tanzanians called in theirchits, and said ‘we worked with you [before], but now we arein charge.’“

To many practitioners, good coordination is mostly aboutgood people. If people work hard to get along, they willsucceed (“coordination works because we make it work”).Strong leaders will create well-coordinated systems throughthe force of their personalities (“the leader was a bit dictatorial,but it worked”). In both Sierra Leone and Ngara, forcefulexpatriate women heading UN offices provided the coordi-nating energy. Regardless of what one thinks of Ms. Lwangain Sierra Leone and Ms. Connelly in Ngara, they both deservesome credit for the functioning and partly replicable coordina-tion systems that emerged in their respective settings whilethey were at the helm. And if Ms. Lwanga was more widelycriticized than Ms. Connelly, she also faced a more compli-cated humanitarian situation involving a much larger numberof actors.

The emphasis on personality as the key to effective coor-dination, so often mentioned in interviews, sheds light on theessential insularity of the international humanitarian regime.Coordinators—and most of the other humanitarian officialsand organizations they coordinate—are outsiders. Ms. Lwangabecame unpopular in part because her desire to engage thelocal authorities conflicted with the interpretations of humani-tarianism and neutrality held by most international agency

49

officials. But her policies also took place while other humani-tarian actors, directly or indirectly, marginalized nationalgovernment participation. In Ngara, Ms. Connelly and herstaff were hailed as effective coordinators, even by TCRS.Coincidentally or not, however, TCRS, the NGO that had themost troubled relationship with Connelly and UNHCR, wasalso the international NGO with the largest local staff, particu-larly at managerial levels. Aid officials also praised the rolethat Tanzania, as host government, played in the response. Yetits primary contribution during the early stages of the crisiswas essentially to get out of the way and let foreign agenciestake the lead.

Perhaps humanitarian coordination must be dominatedby foreigners. Certainly key elements of the apparatus areinternational, even though in many emergencies more assis-tance is provided by local institutions and individuals. Evenif the major actors coordinating and to be coordinated areexternal, however, the challenge of engaging indigenous ac-tors must be addressed. In Sierra Leone and Tanzania alike,the national governments eventually became less than pleasedby the behavior of the humanitarians in their midst. Ulti-mately, it is not just refugees who are guests of the “host”government. International humanitarians are, too. Indeed, theultimate success of their labors, these two situations suggest,depend more largely than they may realize on collaborativepartnerships with national governments.

51

CHAPTER 4

COORDINATION AMONG FUNCTIONS

This chapter examines selected aspects of three key func-tions of humanitarian coordination: strategic planning, re-source mobilization, and security information management.In each instance, innovations in coordination are highlighted.

In Sierra Leone, the most effective coordination throughstrategic planning took place at the sector level, where thedynamics of the much-praised Committee on Food Aid areexamined. In Ngara, emphasis on assertive coordination con-centrated authority in a few hands, with UNHCR selecting alimited number of NGOs as partners.

The discussion of resource mobilization highlights theeffects of ECHO’s decision to channel funding for NGOsthrough UNHCR, and the weaknesses of the ConsolidatedAppeals Process in Sierra Leone.

The discussion of security information management showsthat fluctuating security caused the agencies to place a pre-mium on a consolidated response to sudden changes. Securityinformation management in Sierra Leone proved to be divi-sive, but in Ngara security information management mea-sures strengthened interagency relations.

Strategic Planning

Sierra Leone

The Committee on Food Aid stands out as a particularlyeffective example of sector coordination. Aid officials consid-ered it unusually innovative because it managed to bring allparties involved with food relief into a cohesive structure andmaintained a concerted strategy that responded to dynamicchanges on the ground.

Interviews with CFA officials suggested that sector coor-dination took place only because key relief personnel wantedto work together. One commented that “successful coordina-tion depends entirely upon the individuals—usually directorsof UN and NGO organizations—charged with such coordina-

52

tion.” The personalities of such individuals play a particularlykey role in effective field coordination because “few head-quarters demand that coordination take place in the field—headquarters are almost exclusively interested in resourceacquisition and program growth.” Donors “are only begin-ning to require evidence of coordination on the ground which,with diminishing donor presence in the field, is easy enoughto fake.”

Here again, however, the role of personality in successfulcoordination is exaggerated. The CFA was a good deal morethan congenial people getting along with each other. Individu-als were part of a functioning structure reinforced by activedonor participation. The shared goals and approach of OFDAand ECHO, two primary CFA donors, became the essentialglue. They made CFA participation a prerequisite for receiv-ing support. “Coordination implies consensus on rules andregulations,” an ECHO official explained. ECHO made it “aprecondition that the NGOs agree to participate in the coordi-nation mechanism” for their sector. “If an NGO doesn’t coor-dinate with others, we won’t fund them.”

OFDA imposed the same conditions. Both donors moni-tored CFA coordination through their own attendance at CFAmeetings and by undertaking joint field assessments of foodaid operations. Well beyond the dynamics of personalities,CFA coordination succeeded in large part because donorsenforced it.

The origins of the CFA are significant because, as afounding NGO official recalled, “it sparked coordination asa model for other sectors.” NGOs created the CFA in 1995when “We were [all] fed up with people dumping foodeverywhere. You had massive IDP populations, and some-thing had to be done to coordinate food relief.” NGOs alsosought to counter “double-dipping,” that is, the presence ofthe same individuals on the food distribution lists of severalagencies. “The only way to crack [double-dipping] was tounify, coordinate [food distribution] in one system, and getone caseload list.”

An NGO official credited HACU with a key contributionto the CFA’s development. “HACU was brought into coordi-nation in 1996,” the official said, “but found that UN agencies

53

didn’t want to be coordinated.” He said UN agencies resistedbecause coordination could have revealed their inadequacies.If HACU coordinated UN agency activities, the official cyni-cally asked, “It would show up the fact that [UN agencies]weren’t doing much, and how can you coordinate nothing?”

As a result, another added, HACU turned to sector coor-dination, where HACU helped formalize the previously infor-mal coordination for CFA already put in place by NGOs.“HACU gave us meeting space, a central contact point, arecordkeeping list of who’s doing what and where, [and] apublished directory of NGOs and UN agencies. HACU alsohad the resources to photocopy large documents, host confer-ences, and hire a specialist for developing and maintaining afood registration database.” Equally important, the officialadded, HACU saw its role as “facilitators, unlike the HACUoffice in Liberia.”

What emerged was a three-part committee structure forthe food sector, chaired by WFP. At the top was a body alsoknown as the Committee on Food Aid (the CFA was the namefor both the overall structure and its main body). Meetingweekly, its members were comprised of donors, host govern-ment representatives, and managers of the four food pipelines(three NGOs and the WFP). The CFA was the policymakingbody for food aid, determining, in the words of one NGOparticipant, solutions to “problems that arise or policies thatneed to be changed.”

Examples of policy issues that might arise included “gen-eral free food distribution; targeted food distribution (such asvulnerable groups, institutional feedings in hospitals andschools for the blind); food for agriculture programs; and foodfor work programs.”

If changes were agreed on, the CFA consulted with themiddle committee in the structure, the National TechnicalCommittee (NTC). The NTC linked the central coordinatingCFA body with the outlying Regional Technical Committees(RTCs) which implemented policy. The central NTC roleswere to supervise the RTCs and solve operational problemssuch as crowd control during food distributions. The NTC also“passed up issues for the CFA to decide on.” NTC member-ship was mainly comprised of technical representatives from

54

agencies involved in the CFA—those with relevant expertisefor dealing with the difficulties of food programs.

The RTCs constituted the coordination nexus in the re-gions. Their membership was diverse: officials from localimplementing partner offices, local NGOs involved in foodrelief, local government officials, and beneficiary representa-tives such as paramount chiefs and leaders of IDP popula-tions. It is at this level, CFA participants reported, that theywere able to “build local capacity.” Accordingly, in additionto carrying out food distribution obligations, the CFA alsosponsored workshops for RTC members. The following chart,developed by CFA founders, conceived of the structure as atriangle. (See Figure 1.)

Figure 1: The Committee on Food Aid in Sierra Leone

Adapted from a 1997 CFA memorandum and used with permission

Membership Responsibilities

Food importers anddistributors (WFP,international NGOs),donors, HACU,national government

As above, withother food imple-menting agencies

As above, withbeneficiaryand local gov-ernment repre-sentatives

Policymaking bodylocated in Freetown

Problem-solvingbody on technicalissues, located inFreetown

On-the-ground policyimplementerslocated infood distribu-tion areas

CFA

NationalTechnical

Committee(NTC)

RegionalTechnical

Committees(RTCs)

55

The changing civil war context increased the significanceof the RTC. During the coup period between May 1997 andFebruary 1998 when links between the operations centers inConakry and Freetown and Sierra Leone’s outlying areas werereduced, the intermediate NTC level evaporated, with theCFA’s central committee and the RTCs absorbing all theirduties. Following the return of the international community toFreetown, CFA members began to insert the NTCs back intothe structure.1 By the end of 1998, the issue had yet to be fullyresolved.

Following its entrance into Sierra Leone in 1996, HACUbegan developing an overarching coordination structure,envisioned as a set of concentric circles. The inner circle wascomprised of a body called the Consultative Forum and theInter Ministerial Committee, described in Chapter 2. Thiswas surrounded by a circle of sector committees, includingthe CFA, which itself was encircled by regional committeesfor food (the RTCs), agriculture and others.

Outside of these circles were various members of localgovernment offices (the District Demobilization Committeeand district officers) and two sets of civil society members(the Chief’s Court Committees and community-based or-ganizations). A simplified version of HACU’s National Hu-manitarian Coordination Structure is pictured in Figure 2(page 56).

Figure 2 illustrates how sector committees coordinatedwith local groups. But its depiction of the central circle is lessinstructive. The Consultative Forum contained internationalactors interacting with government ministries, but these alsoacted independently of the overall structure. Donors such asOFDA and ECHO preferred funding international NGOs overthe government, much to its dismay. While the ConsultativeForum’s influence over general humanitarian policy was thusminimized, the actions and policies of international and na-tional actors at the regional sector level were much moreintegrated.

Though it may appear that local civil society and govern-ment representation were marginalized in this elaborate coor-dination structure (they are positioned outside of the circles),regional sector committees generally had significant local

56

Reg

iona

lE

duca

tion

Com

mitt

ee

Reg

iona

l Chi

ldP

rote

ctio

nC

omm

ittee

Reg

iona

lN

on-F

ood

Com

mitt

ee

Reg

iona

l Hea

lthan

d N

utri

tion

Com

mitt

ee

Reg

iona

l Wat

eran

d S

anita

tion

Com

mitt

eeR

egio

nal I

nter

-A

genc

y F

orum

Reg

iona

lT

echn

ical

Com

mitt

ee o

nF

ood

Aid

(CF

A)

Reg

iona

lA

gric

ultu

reC

omm

ittee

Reg

iona

lD

emob

iliza

tion

Tec

hnic

al C

omm

ittee

Dis

tric

t Lev

elD

emob

iliza

tion

Com

mit

tee

Chi

ef's

Cou

rtC

omm

itte

es

GoS

L N

CR

RR

Dis

tric

t Off

icer

Com

mu

nity

Bas

edO

rgan

izat

ions

Dem

obili

zati

onA

dvi

sory

Gro

up

Wat

er a

ndSa

nita

tion

Com

mit

tee

GoS

L I

nter

-M

inis

teri

alC

omm

itte

e

Hea

lth/

Nu

trit

ion

Com

mit

tee

Ed

uca

tion

Com

mit

tee

Non

-Foo

dC

omm

itte

e

Chi

ldP

rote

ctio

nC

omm

itte

e

Sens

itiz

atio

nC

omm

itte

e

Agr

icu

ltu

reC

omm

itte

e

Com

mit

tee

on F

ood

Aid

Nat

iona

lT

echn

ical

Com

mit

tee

Con

sult

ativ

eFo

rum

Figu

re 2

: Nat

ion

al H

um

anit

aria

n C

oord

inat

ion

Str

uct

ure

in S

ierr

a L

eon

e

(Sou

rce:

HA

CU

, Fre

etow

n, S

ierr

a L

eone

, 199

8)

Fre

eto

wn

: C

onsu

ltativ

eF

orum

: D

onor

s,N

CR

RR

, N

GO

s, a

nd U

Nre

ceiv

e fin

ding

s fr

om a

llC

oord

inat

ion

For

a. A

dvis

esG

oSL

Inte

r-M

inis

teri

al C

omm

it-te

e on

Hum

anita

rian

Pol

icy.

Fre

eto

wn

: S

ecto

ral

Coo

rdin

atio

n F

ora

guid

ean

d es

tabl

ish

natio

nal s

tan-

dard

s an

d po

licy

for

hum

anita

r-ia

n op

erat

ions

in c

lose

l ink

age

with

reg

iona

l cou

nter

part

s.

Re

gio

na

l L

ev

el:

Sec

tora

l Coo

rdin

atio

nF

ora

over

see

and

impl

emen

thu

man

itari

an p

olic

y on

the

grou

nd,

wor

king

clo

sely

with

dist

rict

leve

l ent

ities

and

bene

ficia

ry g

roup

s.

57

representation. While many sector committees were chairedby government officials, WFP’s role as CFA chair aroused ameasure of criticism.

Aid agency officials who were CFA members explainedthat food aid arrangements should be considered differentlyfrom the health and nutrition committee, for example, whichwas chaired by Sierra Leone’s minister of health. That commit-tee, a donor official noted, “is really run by the government”while the food sector “cannot be in the hands of the govern-ment because in the past politicians were handing out foodaccording to political concerns, not needs.” Two internationalCFA members went further. “Sierra Leone is a corrupt coun-try,” one said, adding that food is money and power. The othernoted that “the primary reason the WFP chairs the CFA is thatthere are four [international] agencies responsible for fooddistribution and they won’t hand it over to the government. …The record of misuse or diversion of food in the past is anindictment of the government, and even the government isaware of this.”

If food represented power in Sierra Leone, then the aidagencies distributing it became very powerful. This realiza-tion was not lost on government officials, some of whomstrongly criticized WFP, which itself was defensive on thispoint. “The government doesn’t understand why we are here,”one WFP official observed. “It doesn’t understand that foodisn’t free.” While explaining that efforts were underway to“develop [government] counterparts for each [CFA] commit-tee level,” the official maintained that “we have to rememberthat the people of Sierra Leone are what we’re here for.”

This last comment illuminates the CFA’s main strengthand primary weakness. A decidedly innovative and effectivecoordination structure, it was designed to address an em-phatically humanitarian objective: feeding people. Determinedto coordinate food distribution efficiently and comprehen-sively, remarkably adaptable to dangerous and changing cir-cumstances, the CFA developed an admirable record of hu-manitarian achievement. Yet it was also built on the assump-tion that central government involvement would harm foodsector coordination. As a result, international actors becameaccountable for feeding Sierra Leoneans, and, whether they

58

sought to or not, significantly enlarged their influence in SierraLeone. Given the political potency of delivering food to thecitizenry, it is not surprising that the government would pressfor more control over this sector. The CFA thus illustrates howeffective humanitarian solutions can weaken links betweenrelief and development. The CFA fed Sierra Leoneans whileconstraining government participation, thereby limiting thelonger term development value of its contribution. Thoughcapacity building was a component of the CFA’s work in theregions, it was not a priority at the national government level.

Despite this limitation, many aid officials believed thatsuch a structure was applicable to other humanitarian emer-gencies. It was replicated, to some extent, in Liberia. An officialfamiliar with CFAs in both countries commented that “theSierra Leone CFA works better than that in Liberia mainlybecause the Sierra Leone WFP director—WFP being the larg-est CFA partner and the CFA chair in both countries—encour-ages coordination, while the Liberia WFP director does not.”To ensure success, the official believed that “unless a directorsees the need for coordination, no one … will push for it.” Adonor official in Sierra Leone agreed that the CFA in Liberiawas “not nearly as effective as in Sierra Leone.” As many CFAmembers commented, sector coordination can only succeedwith a collective determination among international actors—but most especially among donors—that it do so.

Ngara

For three decades, Tanzania garnered praise for its treat-ment of refugees. Its refugee settlements “probably receivedthe most attention and study as possible models for replication[in Africa].”2 Yet its response to the massive Rwandan influxin 1994 was different. “The Tanzanians had no staff to handlesuch a large emergency,” said one NGO official who hadworked for years with refugees in Tanzania. As a result, theauthorities essentially deputized UNHCR to manage the refu-gee crisis.

To work with refugees in Ngara, an international NGOfirst had to be recognized as an implementing partner ofUNHCR. From the outset, “The Tanzanian government wanted

59

everything to go through UNHCR. They made the UNHCRrepresentative in Ngara the single point of entry” for allNGOs, the NGO official said. This was a new policy for theTanzanian government, which had previously assigned agovernment commandant to coordinate assistance for refu-gees in camps or settlements. In Ngara, however, “it becameclear that [international agencies] weren’t working with theTanzanian government. The settlement commandants had noauthority over NGO activities” as they had following previousrefugee influxes.

The government’s approach to sovereignty thus playedan important role in the emergence of an effective coordina-tion scheme in Ngara. Its voluntary relinquishment of controlover the selection of agencies that would work with Rwandanrefugees inside their country became the catalyst for the devel-opment of highly focused strategic planning. During the earlystages of the emergency, the head of UNHCR’s Ngara Sub-Office, Maureen Connelly, became gatekeeper for all NGOsthat wanted to work in the camps, many of whom wereamazed at her ability to pull together a coordination schemequickly. As one recalled, “She in effect called all the NGOrepresentatives together and assigned sectors to each NGO.”

Much of the strategic planning took place in UNHCR’stent at night. Every evening, NGO representatives met withUNHCR officials. Amid seeming chaos outside, lasting deci-sions were made quickly. An NGO recalled how “Maureenwould say, ‘You do this, you do that.’ If you wanted a contract,you’d better be at those meetings.” Another pointed out thatit took Connelly “one to two weeks to work out [the coordina-tion structure] because it had never been tried before” for arefugee operation on Ngara’s scale. She put a premium oncoordination meetings, which she described as “unwelcomewith so much operational activity going on but necessary toensure cooperation and coordination at the management level,as well as within and between sectors.”3

UNHCR set out three criteria for implementing partnersin Ngara: presence, size, and experience. An NGO officialrecalled that, “Maureen chose only those NGOs who had thecapacity to carry out sector specialties: Oxfam for water, CAREfor food, and so on. She was looking for performance, NGOs

60

who could get the job done.” Although her tight control overthe selection process became controversial, her directive ap-proach was widely recognized as essential to coordinating theNgara response successfully. An NGO official estimated that20 or 30 aid agencies in Ngara sought to be involved, out ofwhich UNHCR chose 12 to implement programs for 500,000refugees.4 Agencies that arrived with their own funds encoun-tered the “Maureen-bar-the-door” approach. One of Connelly’searly reports to Geneva commented that “Whilst [the] em-bargo on participation on demand was not welcomed by manyagencies it did enable real coordination, with standards beingset and maintained.”5 Indeed, the approach was not onlyhighly innovative but, particularly in the early emergencyphase, successful as well.

A kind of mythology soon took shape. One observer notedthat “The leadership and quality of the UNHCR team seemedto be an inspiration to the NGOs.”6 Another remarked that“The limited number of agencies and personnel involved inthe operation engendered an unusually collaborative approachbetween UNHCR and NGOs.”7 But some aspects of UNHCR’svetting process were less successful than others. The emphasison getting NGOs that were in Ngara and had the necessaryinstitutional capacity and experience to carry out assignedactivities did not always work as planned. The NGO chosen tohandle community services for Benaco, one of the world’slargest refugee camps, was the Disaster Relief Agency (DRA),whose initial team lacked broad emergency and communityservice experience. UNHCR also assigned TCRS the task ofdeveloping a centralized trucking fleet to accommodate themajor transport responsibilities, despite its lack of experiencein the sector.

Some NGO officials questioned whether the three selec-tion criteria—presence, size, and experience—were the onlyfactors in UNHCR’s choice of NGO implementing partners.One NGO official said, “UNHCR’s claim that only expertNGOs were engaged is not very correct. Many NGOs werenew to the situation, many were handling their assigned taskfor the first time ever (at least on a very large scale like inNgara), and a large number of staff engaged by the NGOs werevery raw, without any previous [emergency] experience.”

61

One of the world’s largest and most experienced reliefNGOs, World Vision, had arrived in Ngara early on but wasnot selected as an implementing partner. An NGO officialpresent in Ngara at the time speculated that World Vision was“pushed out” by UNHCR because it had begun to workindependently of UNHCR and other NGOs. World Visionstaff “were handing out blankets [to incoming refugees] at theborder. World Vision had Tanzanian experience, but this[action] was thought to be inappropriate. World Vision didn’tget any contracts” with UNHCR in Ngara.8 A UNHCR officialemphasized that “The action taken by World Vision [to with-draw] was in fact the decision of one individual and not byWorld Vision as a policy—and was never viewed as such.”

The rejection of “participation on demand” by NGOsrequired a resolute personality and Connelly, by all accounts,had one. She was a strong, capable, and effective leader, ifoccasionally controversial. Her emphasis on simplifying coor-dination arrangements by limiting the number of participantsmay have attracted criticism from some NGOs, but that didnot seem to affect her policies. Beyond Connelly’s forcefulpersonality, the strategic coordination plan that she and herstaff (especially Deputy Head Jacques Franquin, who hadarrived with her during the Burundian emergency) estab-lished lasted long after she left. Important though her contri-bution was, UNHCR’s success ultimately hinged not on per-sonality but on the high degree of authority it exercised,reflecting both an unusual delegation from the Tanzaniangovernment and a high degree of control over the allocation offunds, as will be explored in the next section.

Resource Mobilization

Sierra Leone

One of HACU’s primary coordination responsibilitiesinvolved managing the Consolidated Appeals Process for UNagencies operating in Sierra Leone. The CAP provides a meansfor coordinating UN agency activities by bringing each agency’sfunding request into a single document, the UN ConsolidatedInter-Agency Appeal for Humanitarian Assistance. HACU

62

officials, who believed that the CAP in Sierra Leone succeededin its aim because it was developed in collaboration withdonors, viewed HACU’s credibility as a key factor in itssuccess. “We know the donors, and we work with them,” oneHACU official explained. Referring to a 1998 consolidatedappeal, the official said that “We told [the donors] what wecould do [and] we’ve done it, so that’s why we got the money.”In consolidated appeals, HACU often received a higher per-centage of requested funding from donors than many otherUN agencies, sometimes far more. The official further con-tended that HACU’s credibility allowed the CAP to be seen as“a major information base for the donors,” who are able totrust HACU because “we do information dissemination andcoordination strategy.”

While enthusiastic about the CAP process, HACU officialsacknowledged the uneven quality of the resulting appeals.“The CAP often describes a plan that can’t be implementedand is based on conjecture,” one official said. Two separate UNfunding appeals for 1998 illuminate this problem. The firstwas a “flash appeal” for March 1 to May 31, 1998 that imme-diately followed the UN’s return to Sierra Leone in early 1998.The second was a 10-month appeal that included the earlierthree months but was extended through December 31, 1998.Assuming a defensive posture, the latter document high-lighted how some of the three-month objectives had beenachieved even if funds had been difficult to secure. “While theofficial response to the [Flash] Appeal may seem muted, manymore of the objectives set for the initial three-month periodhave been satisfied than might first be assumed.” This oc-curred in part because much of the donor funding for humani-tarian assistance had “been provided outside the frameworkof the Consolidated Flash Appeal.”9 A number of donorofficials had viewed the appeal as unrealistic and lackingcredibility. Some chose to simply ignore it entirely and targettheir funding to specific UN agencies in which they hadconfidence.

Most UN agencies in Sierra Leone—Food and AgricultureOrganization (FAO), UNICEF, UNDP, UNHCR, WFP, WorldHealth Organization (WHO), and OCHA/HACU itself—depend at least partially on the CAP process for funding.

63

Initial sections of the two CAP documents examined the scopeof humanitarian need and how UN agencies proposed toaddress the problem. The central sections of each documentwere the project summaries that each UN agency proposed toundertake. The amount of funding sought was US $11,225,538for the initial three months and US $20,253,640 for the ten-month period.

The most glaring weakness was the impression that fund-ing requests by different UN agencies, though presented inone document, were not well coordinated. In the flash appeal,UNDP gave priority to “national coordination,” arguing that“it is critical that the government is able to establish capacityon the ground.”10 Meanwhile, HACU sought a similar sum tosupport coordination activities such as information dissemi-nation and the formulation of interagency strategies. Neitherproject summary indicated how these separate activities wererelated, which appeared to put the UN’s two primary coordi-nation vehicles in competition with each other. Forced tochoose which UN agency should carry out coordination ac-tivities, donors chose HACU. The regular appeal documentstated that “UN-HACU received 91 percent of funds requestedin the Flash Appeal” while “no funds were received by UNDPfor National Coordination.”11 In the end, the CAP had notproposed a method for coordinating two separate UN coordi-nation vehicles but simply highlighted their differences.

The CAP process in Sierra Leone (and, for that matter,elsewhere) has had many detractors. Among the reasons whyits credibility is so low, a HACU official admitted, was theflash appeal’s unrealistic projections of what UN agenciescould accomplish in three months. NGO and governmentofficials also pointed out that the CAP was not well coordi-nated with funding requests either from NGOs or from theSierra Leonean government’s funding requests, even thoughall of them were tapping into the same donors. As one NGOofficial explained, the CAP was not truly consolidated becauseit addressed resource needs only for UN agencies in SierraLeone, while excluding NGOs. In the eyes of one governmentofficial, the CAP heightened competition for donor fundingand the existing inequality between UN agencies and theSierra Leonean government. The UN received more donor

64

support than the government because it had “the machinery todo [fundraising] better,” the official added.

Another recurrent criticism reflected the perception thatUN agencies emphasized their individual objectives over theactual needs of beneficiaries. According to one donor official,this transformed the CAP into “a bloated laundry list. The UNgives unrealistic proposals to donors, so donors don’t fundthem.” Another donor official complained that because “theneeds were not presented by sectors but rather by UN agenciesCAP documents became “non-operational” and thus “a prob-lem for donors” who were seeking to support specific projectsin particular sectors instead of an agency’s general financialneeds. Some donor officials grew exasperated with the entireprocess. One argued that the CAP should be ignored outrightbecause the documents are “basically preposterous. They’retotally ridiculous.”

This negative view was not shared by all donors. UNagencies in Sierra Leone do receive funding through the CAPprocess. OCHA is also working to reform the CAP, and somedonor officials are urging their field representatives to becomeactively involved in working to improve it. “If we don’t getinvolved,” one senior donor official asked, “how will it getbetter?” To this, a field representative responded, “Our head-quarters officials want us to support the UN CAPs, but I don’tknow why we should. If we go on joint assessments [with UNagencies], it looks like we’re approving the CAP. But I’m notpaid to do their work.” Other donor field staff disagreed. “Ilike it,” one official said, noting that in each CAP document,“all parts of UN needs are in one place.” For this official, CAPdocuments seemed like restaurant menus, allowing donorsthe opportunity to select which UN agencies they wanted tosupport.

But the CAP was not conceived as a menu pitting UNagencies against each other in competition for donor support.Its purpose was in fact the opposite: to present donors with away to fund the UN in a more strategic and coordinatedfashion. In Sierra Leone, HACU had the responsibility ofputting the CAP together, but lacked the authority to ensurethat individual UN agency funding requests were coordi-nated with each other and with other humanitarian actors,

65

NGOs, and the government. As a result, the main utility of theCAP in Sierra Leone—creating a single document that in-formed donors of the UN system’s humanitarian objectives—exposed cracks in UN agency coordination. One observernoted that, as a result, the role that OCHA, including itscountry offices such as HACU, plays in the CAP processultimately provides “no value added” to the goal of coordinat-ing UN funding requests. “OCHA simply staples together”individual UN agency funding requests and hands out theresult as a CAP document, he said.

Ngara

Considering that more than half a million Rwandan refu-gees in Ngara needed humanitarian assistance, it is somewhatsurprising that adequate funds were generally so easy to raise.A UNHCR official commented that from the outset of theemergency in April 1994, “we had plenty of money. It wascoming out of our ears, and that solved a lot of problems.” Theprimary reasons for this, the official related, were the massivesize of the emergency and the press attention it received.

As with other elements in Ngara, the Tanzanian govern-ment was a key facilitator of financial coordination. At theoutset it allowed planes to arrive from the Tanzanian capital,Dar es Salaam. “The Tanzanian Regional Commissioner [forthe Kagera region] was terrific” in expanding access to thisformerly remote Tanzanian district, said one UNHCR official.The planes were ostensibly for the press, but key officials fromthe European Union and the United States came, too. Thedonors came up to Ngara and were very pleased with theextent of the response to the emergency, a UNHCR officialrecalled. Ready access to key players in the emergency stokedthe interest of Western donor governments that were alreadybeing pressed to respond to the public outcry.

A new centralized coordination system for funding soonemerged in Ngara, though its origins are somewhat unclear.By one account, UNHCR had been “waiting for a chance toimplement a new emergency response system which had beendesigned following the conclusion of the Gulf War in 1992.”12

But an ECHO official asserted instead that the new funding

66

arrangement was simply an isolated reaction to the Rwandanrefugee emergency. “After the massive influx of Rwandanrefugees into Tanzania,” the official explained, “we decided toput all financial support through UNHCR in support of coor-dination.” Underscoring the exceptional nature of its policytowards UNHCR in Ngara, ECHO decided to supportBurundian refugees in Ngara differently. Activities forBurundian refugees in the Lukole camp who had remained inTanzania well into 1994 were funded more directly to NGOs.

ECHO’s decision to fund UNHCR directly was unusual.One official called it “a one-time only situation, specific to theGreat Lakes crisis,” highlighting a hesitancy among manydonors to fund UN agencies directly. Under normal circum-stances, the ECHO official continued, “our preference is tofund European NGOs. You don’t have to go through theUnited Nations to support refugees.” Besides,” he contended,“it’s not good for UNHCR to be funded so much by onedonor.” The special funding arrangement in Ngara reflectedthe fact that “the coordination by UNHCR was so good.”Coordination met ECHO’s criteria: the mechanism was effi-cient, it supported coordination with all the agencies involved,and it allowed ECHO to have a voice on key political orcoordination issues in the camps. By contrast, humanitarianoperations for Rwandan refugees in Zaire involved competi-tion for funds between NGOs and UNHCR, making it difficultfor ECHO and other donors to develop a coordinated strategy.

While directly funding UNHCR-Ngara was unusual forECHO, it was standard practice for the primary U.S. govern-ment donor agency for refugee crises, the State Department’sBureau for Population, Refugees and Migration . Unlike ECHO,BPRM normally provides direct funding to UNHCR “by andlarge without earmarks,” one BPRM official noted. “We don’tfund projects; we fund UNHCR’s general program.” ForBPRM, the Ngara case was notable only for the size of itscontribution.

Although the common approach adopted by the twomajor humanitarian donors had a positive impact on coordi-nation, the ceding of donor funding authority to UNHCRdirectly affected UNHCR-NGO relations. Whether the effectswere positive or negative and whether gains in coordinationoutweighed other perceived drawbacks depended on several

67

factors. Some NGO officials found that, as UNHCR’s powerexpanded, their own leverage with UNHCR decreased. Forsome, that was fine because, as one explained, “UNHCR set itup so that the NGOs didn’t have to make decisions. In return,they got money. So with many NGOs, Maureen [Connelly]and UNHCR were very popular.” Many NGO officials de-scribed Connelly and her UNHCR colleagues as excellentlisteners who responded to their concerns.

An early UNHCR document characterized UNHCR-NGOrelations as “excellent” during the early stages of the emer-gency, an assessment confirmed by NGO participants. Later,however, some NGOs came to believe that UNHCR was favor-ing certain NGOs over others. This belief was tied to theirsuspicions of a direct correlation between the percentage offunding that UNHCR gave to an NGO for Ngara operationsand that NGO’s loyalty to UNHCR. An official from an NGOthat received some of its funding from non-UNHCR sourcessaid that his agency “could tell UNHCR ‘no’, but NGOs [receiv-ing all their funding from UNHCR] couldn’t do that, so theywere easier [for UNHCR] to control.” Another NGO officialwent further, saying “You can bet your last dollar that thoseNGOs that were 100 percent funded by UNHCR had to stay inline with Maureen and jumped when they were told to jump.”

Perhaps it is not surprising that UNHCR’s funding andinfluence were linked. What is more notable, however, is thata UN agency wielded such authority over humanitarian actorsthat were not part of the UN family. UNHCR’s positionundoubtedly helped establish and maintain an effective coor-dination structure. Yet its control over NGOs may have had asignificant cost, contributing to the ineffective response of thewider humanitarian community to the government’srefoulement of Rwandan refugees in late 1996. A more looselystructured coordination system might have encouraged cre-ative problem-solving and afforded NGO views more weight.At any rate, the authority vested in UNHCR by the nationalgovernment on the one hand and donors on the other leftNGOs with limited leverage. For donors and most NGOs,reduced influence with UNHCR and the government wasmore than offset by the resources and sector responsibilitiesthey were awarded.

68

Security Information Management

Sierra Leone

The RUF’s tactics of roaming the forests and thriving onsurprise made it difficult to track and its attacks difficult topredict. The character and tactics of the RUF heightened theneed to maintain reliable information on security for all hu-manitarian actors in Sierra Leone as the basis for commonresponses. Despite the fundamental need to update and shareinformation continuously on security, however, what devel-oped was not a consolidated approach to security manage-ment but rather two parallel information management sys-tems that often presented conflicting evaluations of safety anddanger.

The evolution of the two systems took place in two stages.During the junta period (May 1997 to February 1998) when UNagencies, the Sierra Leonean government, members of theinternational diplomatic community, and some NGOs shiftedtheir base of operations to Conakry, the security informationmanagement system that emerged there interpreted the situ-ation in Sierra Leone in ways that clashed with views held bythe ICRC and NGOs that retained their central operationalbases inside Sierra Leone. The second stage arose in March1998 immediately after ECOMOG soldiers captured Freetownand drove the RUF and its coup partners back into the inlandforests. The separate systems then split more closely alongcivilian and military lines. During both the junta and post-junta periods, the two groups managed and shared securityinformation separately among their members. The core mem-bership of one group included the SRSG, UNOMSIL, andECOMOG; the other group was comprised of the ICRC andNGOs that remained inside Sierra Leone during the Juntaperiod.

One NGO director in Freetown traced the origins of sepa-rate information management groups to the period before theMay 25, 1997 military coup. For coordinating security mattersin the field, the UN Field Security Handbook directs that “onesenior [UN] official to be titled Designated Official... willundertake overall and special responsibility for the security

69

and protection of the [UN] organizations’ staff members andeligible family members and property. … At most duty sta-tions, the Designated Official will be the Resident Coordina-tor/Resident Representative of UNDP.”13 The designatedofficer usually has a UN security officer on staff, as was thecase for Sierra Leone. Although the official is technicallyresponsible only for UN staff, it was also the case that, as oneNGO official pointed out, “the model in many emergencies isthat the UN provides a security system to collect, synthesize,and redistribute security information” to NGOs as well. Insome emergencies, this assistance might include “a sharedradio channel [for UN and non-UN humanitarian personnel,as was the case in Ngara] and a coordinated evacuation planand recommended precautions” for those delivering humani-tarian assistance. Yet, he said, “None of this happened in SierraLeone. We approached the UN security officer about briefingNGOs on security,” only to have this blocked by the UNDPhumanitarian coordinator/designated official. Thus did theenmity between UNDP and international NGOs extend intothe management of security information.

During the junta period, a number of aid officials presentin Conakry reported that good security information aboutSierra Leone was hard to come by and even more difficult toshare. “The problem was that [security briefings] took place ina public forum,” one NGO official based in Conakry recalled,“so people didn’t say much.” A UN official also present inConakry explained that the UN security officer and others“were briefing people at 7 p.m. every night.” But since themeetings were large and may have included reporters or eveninformers for the junta, “you couldn’t pass on [sensitive]information, and sometimes there was nothing to report.”Often security briefings were based on nothing more thanradio broadcasts. As a result, an NGO official described howhis colleagues, together with “UN allies,” regularly gatheredinformally to exchange information considered too sensitivefor public meetings. This included security information com-ing directly from Sierra Leonean staff members still insideFreetown and regional capitals, and from members of regionalsector committees, in particular those connected to the CFA,which remained in operation during the junta period and

70

relayed security updates by telephone and radio to officials inConakry.

An important first-hand source of security informationwas generally dismissed in Conakry: NGOs and the ICRC thatcontinued working inside Sierra Leone during the junta pe-riod, security constraints notwithstanding. “We were never asable to move about as freely [in Sierra Leone] as during thecoup period,” stated one of those present. Particularly in theearly months of the junta period, NGO and ICRC officialswould regularly drive from Sierra Leone across the border toGuinea to buy supplies and consult with aid officials based inConakry. They found that “while we were all saying that thesituation [in Sierra Leone] was not as bad as it was perceivedin Conakry,” their perspective was not believed by most UNofficials, ECOMOG and ECOWAS personnel, and members ofthe diplomatic community. “Information management inConakry was terrible,” concluded one aid official who regu-larly visited the Guinean capital from Sierra Leone. “Therewere incredible discrepancies. People [based there] boughtthe information that served their interests.” As a result, theUN resisted assessments by NGO and ICRC staff working inSierra Leone of the security situation inside Sierra Leone.

While the parallel information systems produced differ-ent assessments of risk and resulted in different levels ofhumanitarian engagement and programming, both systemswere, in different contexts, accurate. With the RUF out of theforests and in control of the capital, the threat of attacks on aidproviders in rural areas was dramatically reduced. Thoseareas, ICRC and some NGOs maintained, were reasonablysafe. Freetown, however, was entirely different, and ICRC andNGOs based inside Sierra Leone were regularly harassedthere. Journalists in Freetown reported how the RUF and theAFRC were committing atrocities there. The media reportedthe atrocities and the news circulated throughout the aidcommunity in Conakry. Secure rural areas might have beendefined and demarcated for aid operations, but this did notoccur. “[UN Secretary-General ] Kofi Annan told the UN toisolate the junta,” a UN official who was in Conakry at thattime recalled, “and refuse to recognize it. The UN had taken aposition of strong support for [ousted President] Kabbah.” As

71

a result, each group emphasized the security information thatsupported its views: Conakry-based agencies focused on dan-gers in Freetown while Sierra Leone-based agencies empha-sized safety and access in the rural areas.

The lack of collaboration on security information contin-ued following the February 1998 return of President Kabbah toFreetown and the RUF to the forests, and in fact worsened overtime. But membership in the two security information groupschanged slightly. One group was comprised of most interna-tional humanitarian actors: NGOs, the ICRC, and all UNagencies except UNOMSIL. It resumed the coordination ar-rangements for sharing security information that had existedin Freetown prior to the coup. The other group was composedof the primary military and political actors in Freetown:ECOMOG, which had essentially assumed the role of theKabbah government’s defense force, UNOMSIL’s militaryand human rights observers, and the Sierra Leonean govern-ment.

In the fall of 1998, the government, UNOMSIL andECOMOG officials believed that the RUF was no longer at-tacking but had instead assumed a defensive posture. “I don’tthink the war is escalating,” a high-ranking Sierra Leoneangovernment official said at that time. Others in the humanitar-ian community questioned this view. “UNOMSIL doesn’twant to give out bad news,” a donor official in Freetownexplained, preferring to paint a rosy picture of the war situa-tion. Another aid official contended that “UNOMSIL [is] be-having like there’s an almost complete return to normalcy.” Athird aid official explained that the official way to get informa-tion on the security situation “was to go and ask ECOMOGofficers, who say ‘the situation is rosy.’“ The only reason theywould say that, however, was “because security is their re-sponsibility.” Subsequent events proved the rosy perspectiveincorrect. Late in 1998, the RUF was actually gaining strengthand would enter Freetown in January 1999. The securityconcerns voiced by the ICRC/NGO/UN agency group in 1998had been accurate.

The dramatic difference in security interpretations re-flected in part the presence of two security systems existingwithin the UN structure. One was headed by the UN security

72

officer, who reported to the humanitarian coordinator at UNDP;the other was headed by UNOMSIL military observers. In thefall of 1998, the former system appeared to have credibilitywith a large proportion of the humanitarian community inFreetown. Information was collected from a wide variety ofsources, analyzed, and shared not only with UN personnel butwith NGOs and the ICRC as well. The UN security officer gavebriefings at weekly NGO meetings on security concerns andNGO officials also shared relevant security information fromtheir own observations. The security officer also met weeklywith the UN’s security management team (SMT) to give amore specialized briefing to UN agency heads. SMT memberswere then supposed to relay important information to their

Figure 3: Parallel Security Systems for Sierra Leone Duringthe Junta Period (May 1997-February 1998)

Agencies Based inConakry, Guinea

Agencies Based inside Sierra Leone

Membership

SecurityPerspective

Mode ofCommunication

Other

UN agencies,ECOMOG,ECOWAS, someinternational NGOs

ICRC, some interna-tional NGOs

Urban focus: empha-sized serious dan-gers in Sierra Leone,particularly inFreetown, wherejunta leaders resided

Rural focus: assessedcountry risk asserious, but de-scribed access tobeneficiaries in ruralareas as very good

Formal dailybriefings, allowingbroad access tobriefings

Shared informationinformally amongthemselves andsplinter groupmembers

Splinter group ofNGO and UNofficials developeddiscreet meetings toexchange sensitiveinformation

73

staffs. The UN security network included a team of nationaland international wardens responsible for passing along secu-rity information to UN staff quickly.

Although UNOMSIL was represented at the SMT meet-ings, a UN official explained that “UNOMSIL has its owninformation gathering system that is separate” from the UNDP/UN security officer’s system described above. Some UN offi-cials reported that UNOMSIL was not allowed to exchangemost of the security information it collected with non-UNOMSIL officials. As a result, there was “no formal systemfor information management or information sharing” betweenthe two UN security systems. Informal information exchangebetween UNOMSIL officers and NGO and ICRC officialsrarely took place.

A high-level UNOMSIL officer portrayed differences be-tween UNOMSIL and the UN security officer as a function ofthe ways that military and civilian systems operated.UNOMSIL, the official explained, is a military unit (though italso hosted human rights monitors from the UN High Com-missioner for Human Rights), and for military units there arethree types of security information. The first type, dubbed“absolutely correct,” is information that a UNOMSIL officercan verify first-hand. The second-best type of information is“information received from somebody credible” from thecivilian side of a humanitarian operation such as an officialfrom another UN agency or NGO. This kind of “crediblesource” provides information that “may or may not be true.”The final and least reliable kind of security information is“rumors.”

The UNOMSIL officer’s description effectively down-graded the credibility of the civilian security structure, led bythe UN security officer (himself a former soldier) but includ-ing every UN agency outside of UNOMSIL, all internationalNGOs, and the ICRC. Although many of these civilians had afar greater experience in Sierra Leone than their UNOMSILcounterparts, UNOMSIL nonetheless considered their infor-mation largely unreliable. UNOMSIL thus positioned itselfdubiously as a separate entity with superior information. Theseparate security coordination arrangements for this secondset of parallel systems are described, in simplified form, inFigure 4 (page 75).

74

The changing alignments in the two sets of parallel struc-tures produced an ironic shift in how Sierra Leone’s securitysituation was perceived. During the junta period, many UNand other officials in Conakry maintained that Sierra Leonewas exceedingly insecure while NGOs and the ICRC activeinside Sierra Leone were reporting unprecedented successesin delivering assistance to IDPs. After the entire humanitariancommunity returned to Freetown, however, perceptions re-versed. Officials from UNOMSIL, ECOMOG and the govern-ment, among others, described a steadily improving situationwhile NGOs and ICRC officials noted an increasingly insecuresituation across the country.

The UN security officer’s situation vis-à-vis the humani-tarian community was in many respects similar to HACU’s.Based in Conakry during the junta period, his credibility withmany NGO and ICRC officials suffered, just as HACU’s ap-parently reluctant membership on the Humanitarian Exemp-tions Committee in Conakry damaged its reputation amongaid agencies outside the UN. But back in Freetown afterECOMOG had chased RUF and AFRC forces away early in1998, the security officer, Grahame Membry, and HACU offi-cials, led by their chief, Robert Painter, worked hard to repairtheir damaged relationships with NGOs and the ICRC. AmongMembry’s tasks was to bridge the traditional separation be-tween UN and NGO agencies on security concerns. A seniorUN security official summed up the difference by explainingthat the UN “has a structure and discipline” whereas NGOs“don’t want to work as one corporate body.” Despite this, anddue to a collective concern about security issues, UN and non-UN agencies regularly exchanged relevant information.

The exception was UNOMSIL. While the aid communitygenerally shied away from battlefield-oriented activities,UNOMSIL coordinated its military observer work withECOMOG troops who were battling the RUF on behalf of theSierra Leonean government. Committed to the government’sobjectives, UNOMSIL’s perspective of the security situationwas influenced accordingly. Its mid-1998 entrance in SierraLeone and its separation from aid organizations exacerbatedthe existing divide between agencies that strongly supportedthe government side and others struggling to be seen asimpartial.

75

It was ironic that UNOMSIL, a UN institution created tofacilitate the return of peace in Sierra Leone, was viewed as athreat to security by so many humanitarian officials inFreetown. An ICRC official expressed a widely held concernabout the “clear discrepancy” between the roles of these dualUN systems. He argued that “UNOMSIL has undermined theUN security officer’s role. The UNOMSIL military observersact as if it’s secure to move around. On the other side, the UNsecurity officer is saying that it’s dangerous to move around.”The ICRC official was also concerned that UNOMSIL’s mobil-ity in conflict zones would eventually jeopardize the move-ments of all humanitarian agencies.

ICRC and others sharing this view, however, did not offeran alternative that would have enabled UNOMSIL to carry outits important work without ECOMOG’s assistance. That said,the lack of coordination between UNOMSIL and the UN

Figure 4: Parallel Security Systems in Sierra Leone Followingthe Junta Period (February 1998-January 1999)

"Military" Actors "Civilian" Actors

Membership

SecurityPerspective

Mode ofCommunication

UNOMSIL (militaryobservers and humanrights monitors fromUNHCHR)

All UN bodies exceptUNOMSIL, interna-tional NGOs, and theICRC

Viewed general secu-rity situation as im-proving, even "rosy"

Viewed general secu-rity situation as highrisk and dangerous

Internal securitybriefings, collabora-tion with ECOMOGand the Sierra Leonegovernment; limitedcoordination withother UN bodies,international NGOsand the ICRC;attended UNSecurity Manage-ment Team briefings

Led by UN securityofficer who regularlybriefed UN SecurityManagement Teamand exchangedsecurity informationwith internationalNGOs at the weeklyNGO meetings

76

security officer (and, by extension, all other UN agencies)effectively threatened the entire humanitarian operation.UNOMSIL’s stance on this issue seemed unnecessary andeven reckless, given the potentially dangerous consequencesthat its work posed for all humanitarian personnel in SierraLeone. As one UN official warned, “The humanitarian com-munity can’t survive here if we don’t work well together” onshared security concerns. The connection between coordina-tion and effective security information could not have beenmore clear.

Ngara

UNHCR and NGO officials present during early days ofthe emergency in 1994 did not devise a scheme for sharinginformation on security in Ngara. The camps seemed safe andreasonably secure for aid officials and, to a large extent, forrefugees as well. With the humanitarian operation straining toaddress the needs of hundreds of thousands of people, secu-rity was not an overriding priority. Massive logistical prob-lems involved in providing water, food, and shelter to refu-gees received top billing. However, as one UNHCR official inNgara recalled, with the wisdom of hindsight, “We werenaive.”

UNHCR rather than the Tanzanian authorities was re-sponsible for refugee camp security. A UNHCR official re-called how “the Tanzanians realized that they lacked thecompetence to deal” with security for such an enormouspopulation of refugees and trusted UNHCR staff to handlesecurity concerns. Information-sharing between UNHCR andthe government on security issues, however, took place not inNgara but in Tanzania’s distant capital, Dar es Salaam. There,the UNHCR official explained, his colleagues were able toinform the Tanzanian authorities of the situation in the camps“and give them the illusion that they’d participate in theoperation.”

The fact that UN work in Ngara was “never obstructed bythe government was very unusual,” a researcher on the GreatLakes emergency observed. But an NGO official in Ngararecalled that the authorities “had no staff to handle the emer-gency.” As a result, “the Tanzanian government said, ‘We’llwatch the borders, but, UNHCR, you handle the refugees.’“

77

Reflecting on the government’s decision to relinquish its secu-rity role for refugees in its country, the official concluded that“it was almost as if the Tanzanians gave up their sovereigntyover that area of Tanzania.”14 Its approach was consistent withthe authority it ceded UNHCR over the management of thecamps.

Because UNHCR at the time had yet to “realize the nega-tive influence of the refugee leaders,” as one UNHCR officialrecalled, it initially hired Rwandan refugees as camp securityguards. (It was reluctant to hire Tanzanian policemen to carryout this assignment for fear that they might abuse their powerand assault refugee women.) As it turned out, the guardshelped the commune leaders, many of whom were likely tohave been involved in the genocide, to assert firm control overthe camps. “It took us three months to realize what kind ofpeople we were dealing with,” a UNHCR official recalled.“We were so focused on humanitarian concerns that we forgotabout the genocide during the first three months. This was abig mistake.”

The mistake eventually led to a significant security inci-dent in Benaco camp, which began with the arrest of 14 Rwandanrefugees by Tanzanian policemen. They had acted on informa-tion from refugees that these 14 had organized genocidalkillings in Rwanda. The policemen then asked UNHCR whatthey should do with the suspects. Although UNHCR officialssaid that they had decided to relocate the individuals to arefugee settlement in another part of the country, the Tanzani-ans considered the relocation process too slow. Without noti-fying UNHCR in advance, they released the suspects from theNgara jail and allowed them to return to the camps.

Among the group was the notorious Kibungo Bourgmestre,Rémy Gatete, who had become “famous for [his] enthusiasmfor the killing business” during the genocide15 and whomUNHCR soon realized was indeed very dangerous. On July15, 1994, a UNHCR protection officer asked Gatete to report toUNHCR’s field tent in Benaco camp. Gatete arrived later thatday with 5,000 refugees brandishing sticks and machetes. “Upto this point,” a UNHCR officer present that day said, “therelations between the Rwandans and UNHCR were verygood, because [the refugees realized that] we were trying tosave their lives.”

78

But Gatete would change this. The confrontation wasbroadcast over a VHF radio network established the weekbefore, allowing aid workers the opportunity to listen to thedeliberations between Gatete and UNHCR officials. In theend, the Tanzanian police dispersed the crowd by firing shotsin the air. Ten days later, Tanzanian authorities transportedGatete away from the refugee camps a second time. He did notreturn. When Gatete arrived in Dar es Salaam, one aid officialrelated that “he was met by the French ambassador, who flewhim to France to be with [former Rwandan President]Habyarimana’s wife,” who was also thought to have beeninvolved in genocidal activities. The official also related arumor that Gatete later left to join Hutu extremists in a Rwandanrefugee camp in Congo-Brazzaville.

The Gatete incident transformed the aid community’sperceptions. People in the camps were no longer simplyrefugees in need of assistance; they were a potentially danger-ous population led by criminals capable of extreme violence.This shift was accentuated by Western press coverage of theincident, which contributed to a new emphasis on protectingexpatriates from refugees regardless of expense. Donors sup-ported this change and “simply announced that funding forsecurity was no problem.”16 In a costly move, all NGO opera-tions were shifted away from the camps and resettled on a hill15 kilometers away. Surrounded by thinly populated hills, thenew barricaded home for expatriate staff was called “K-9”.

If the release of the 14 suspected criminals had under-mined UNHCR’s credibility with NGOs, its handling of Gatetewent a long way to restoring it. Following his removal from theregion, UNHCR and NGO officials working in Ngara filed ajoint letter with the district commissioner requesting that theTanzanian authorities provide security in the camps. Thiseventually led to the appointment of a former Tanzanianpolice chief as the head of UNHCR’s security office. UNHCRalso hired Tanzanian policemen from districts across the coun-try and developed a system in which Tanzanian police becamewardens of Rwandan refugee guards. An NGO official com-mented that since UNHCR “had no enforcement capability ofits own” and had offered much higher salaries than the police-men normally received, “they gutted the police staffs in other

79

parts of the country.” UNHCR also requested that the Tanza-nian army not be allowed inside the refugee camps but insteadshould patrol surrounding areas, although at times the troopsentered the camps.17

UNHCR also developed staff security systems for NGOsand formed a security committee for UN and NGO personnel.Its formation was another unusual aspect of the coordinationsystem in Ngara. The committee shared information on secu-rity concerns for aid staff members and developed an evacu-ation plan. UNHCR officials emphasized that “this was a jointagency system” involving UN agencies and NGOs working ina consolidated security information system. Hence the divi-siveness on matters of security coverage between UN andother aid personnel that characterized the situation in SierraLeone and elsewhere did not arise.

Indeed, in accordance with UN regulations the UNDPresident representative in Dar es Salaam had appointedUNHCR as the local security head for Ngara District butcautioned that the UN should provide security, includinginformation and evacuation, only to UN personnel. UNHCRin Ngara rejected this policy. UNHCR officials apparentlyinformed UNDP that they would be sharing all security ar-rangements with their partner NGOs. Perhaps the most sig-nificant indicator of this joint UN-NGO security arrangementwas the use of a shared radio frequency for all humanitarianpersonnel. This was a critically important security measurebecause it provided all personnel with equal access to infor-mation. Security regulations also required all agency staff tocarry a battery-charged hand radio with them in the camps.

Other security measures were designed only for expatri-ates. Out of concern that some of the Rwandan refugee work-ers were relaying sensitive information to genocidaires in therefugee camps, the security committee held weekly meetingsthat required every UN and NGO agency head to attend. Eachagency head then shared security information with their expa-triate staff in private conversation. Nothing was to be writtendown, a precaution invoked to prevent Rwandan personnelfrom intercepting the information.

Many of the innovations described in this chapter appearto be replicable. Some, such as the CFA, have already been

80

adapted to other sectors in Sierra Leone and to the humanitar-ian food sector in Liberia. For security arrangements, both theSierra Leone and Ngara cases illuminate how NGOs and UNactors can coordinate on shared security concerns even whenUN regulations restrict such interactions.

Other innovations may be somewhat more difficult toreplicate, given the particularities of these two settings. InSierra Leone, the CAP illuminated how a coordination innova-tion can threaten coordination. The two CAPs considered herelacked cohesion because they neither explained program rela-tionships across UN agencies nor clarified how UN agencyfunding coordinated with non-UN agencies. In Ngara, UNHCRdepended on the support of others to initiate many of itscoordination ideas. Whether other host governments willgrant UNHCR the power to restrict the number of NGOparticipants, as the Tanzanian government did, also remainsuncertain. Finally, applying the centralized funding mecha-nism initiated in Ngara to other refugee settings may beconstrained by ECHO’s stated reluctance to do so.

81

CHAPTER 5

COORDINATION WITHIN PROGRAMS

This chapter examines two emergency assistance pro-grams: one from Ngara, the other from Sierra Leone. Bothprograms were developed in direct response to specific chal-lenges and both adjusted coordination arrangements to meetrapidly changing circumstances. The challenges themselves,however, were quite different, as are the lessons that can bedrawn from the experiences. The first case involves the repa-triation of refugees to Rwanda from Ngara District, Tanzania,in the last weeks of 1996. The second concerns emergencyeducation programs for Sierra Leonean and Liberian refugeesin Guinea. Since the repatriation effort proved the more com-plex and controversial, it is accorded more space and de-scribed first.

The Ngara and Sierra Leone experiences are joined by aninteresting paradox. Repatriation is a central concern in virtu-ally every major refugee emergency. Once it is considered safeto go home, aid agencies and governments assist refugees toreturn. Yet many agencies hesitate to support educationalservices during humanitarian emergencies, fearing that set-ting up schools in refugee or IDP camps will make the campsseem more permanent and eventual repatriation more diffi-cult. The two cases paired here offer the antithesis of standardviews: a repatriation program from Tanzania to Rwanda thatwas unusually controversial and an emergency educationprogram for Liberians and Sierra Leoneans in Guinea that waspopular and well supported.

The pairing of these two cases challenges conventionalwisdom in yet another way. Although repatriation duringemergencies is supposed to be voluntary, the refugees inNgara were rarely consulted and their leaders eventuallysought to block the repatriation effort. Orchestrating the repa-triation of Rwandan refugees took place without refugee—or,in the end, UN—input. In contrast, the emergency educationprogram in Guinea arose in response to refugees who hadstarted schools on their own. Together with their Liberiancounterparts, Sierra Leonean refugees served as central fig-

82

ures in the development of a well-coordinated educationprogram.

Coordinating Repatriation: The Ngara Case

Changing Circumstances

Although humanitarian work these days is rarely easy orpredictable, the Rwandan return from Tanzania in December1996 stands out as one of the most controversial and challeng-ing coordination efforts in decades. Repatriation involved amassive population ostensibly under the control of leaderswho did not want to return home. And, because of theirparticipation in war crimes and crimes against humanity, theymay not have even qualified as refugees.1 In addition, threefactors preceding the Rwandan emergency’s suddendénouement in Tanzania made this coordination challengeparticularly difficult.

First, ever since the Rwandan refugees entered Tanzaniain April 1994, UNHCR-promoted repatriation strategies failedto spur significant movement. One initiative involved spon-sored visits by refugee delegations to Rwanda in the vain hopethat these delegations would persuade large numbers ofRwandan refugees to return home. A second initiative fea-tured visits to the camps by high-level Rwandan and Tanza-nian authorities, which also yielded disappointing results. Athird used videos with statements by various leaders encour-aging refugees to return and depictions of peaceful conditionsinside Rwanda. Even a cameo appearance by former U.S.President Jimmy Carter’s in a video failed.2

Refugees responded to such initiatives by questioningtheir credibility, particularly in view of the tens of thousandsof Rwandans who had been imprisoned without charges ortrials for suspected participation in the 1994 genocide.3 Theywere also under pressure from their leaders not to repatriate.In the end, the UNHCR-led repatriation campaign never gath-ered steam. During the first 11 months of 1996, only 4,200 of theestimated 534,000 Rwandan refugees in Ngara District—lessthan 1 percent—returned to Rwanda.4 An NGO official com-mented that the repatriation could not even keep up with thepopulation growth rates in the camps.

83

Second, public announcements by European Union, UnitedStates, and Tanzanian officials during 1996 eroded UNHCR’scommand over the camps and made a final, aggressive repa-triation effort seem almost inevitable. In January, TanzanianDefense Minister Edgar Majogo announced that the refugeepresence in Tanzania “must be temporary [and] cannot be acontinued stay.”5 In April, USAID Administrator Brian Atwoodand European Union (EU) Commissioner for HumanitarianAid Emma Bonino publicly suggested in Ngara that the UNSecurity Council consider setting “a date for closing the camps.”Over time, the two donors had grown concerned over UNHCRcamp management. An ECHO official feared that “we weresedentarizing the refugees into Tanzania,” making them moredifficult to repatriate. ECHO’s chief, Emma Bonino, sug-gested that “Keeping the current level of humanitarian aid isunsustainable in the absence of concrete prospects for a solu-tion to the refugee crisis.”6 U.S. officials were equally agitated,claiming, said a UNHCR official, that “UNHCR didn’t wantthe refugees to go home because they’d lose their jobs.” As aresult, “Congress said, ‘we’re not going to fund this anylonger.’“ All signs suggested that the refugees would headhome soon. The only remaining questions, it seemed, werehow and when.7

Third, the startling voluntary return of 600,000 Rwandanrefugees from Zaire in November 1996 immediately reverber-ated across Ngara. In the wake of civil war between Zairianinsurgents (led by Laurent Kabila and supported by theRwandan army) and Zaire’s national army (supported byformer Rwandan government soldiers and Interahamwe mili-tiamen), many Rwandan refugees there returned home. Thisundermined somewhat the widely held assumption thatRwandan refugees comprised a “willing and loyal commu-nity” that chose to support their camp leaders. It also providedthe Tanzanian and Rwandan authorities with the precedentthey sought.8

The two governments reacted swiftly. Days after Rwandanrefugees began to repatriate from Zaire, Tanzania’s PresidentBenjamin Mkapa met with Patrick Mazimhaka, Rwanda’sRehabilitation Minister, in Dar es Salaam. President Mkapapledged to “facilitate the voluntary repatriation of all Rwandan

84

refugees camped in Tanzania” and asked for internationalhelp in the effort.9 Minister Mazimhaka declared that Rwanda“was ready to receive all refugees from Tanzania.”10 UNHCRthen dispatched a special envoy to Tanzania, Sergio Vieira deMello, to help coordinate a new repatriation initiative. OnDecember 5 the Tanzanian government and UNHCR an-nounced a plan. “We are putting out the message that it is timeto go back” by the end of the month, UNHCR’s deputy directorfor Tanzania, Lennart Kotsalainen, said following the Dar esSalaam consultations. “Hopefully by the end of this eveningthey will get the message.” Kotsalainen outlined the plan ofaction: “The police would protect the refugees, supported byTanzanian troops if necessary... [while] the United Nationswould provide transport across the border.” The transferwould be coordinated with the Rwandan authorities, who“had given assurances that they would receive all refugeesfrom Tanzania.”11

The tripartite repatriation plan was unusual. Althoughthe UN 1951 Convention on Refugees and the 1967 Protocolare designed to ensure that refugees decide when to return totheir homes, this repatriation was not going to be voluntary.The joint statement declared that the government of Tanza-nia—not the refugees—had “decided that all Rwandese refu-gees can now return to their country in safety.”12 As UNHCRspokesman Peter Kessler commented, “UNHCR is alerting therefugees in Tanzania that this is the time to begin their re-turn.”13 Human rights groups criticized the plan. AmnestyInternational attacked UNHCR for “effectively rubber-stamp-ing” a decision which “contravenes its own basic principles ofprotection of refugees.”14 UNHCR did not immediately re-spond to this critique. However, a series of unexpected eventswere about to unfold that would significantly change UNHCR’srole in the repatriation effort.

The Tanzanians Take Charge

It took only a week for the repatriation plan to be under-mined. On December 12 refugees abandoned their camps andheaded not toward Rwanda but into the Tanzanian interior.Directed by their leaders, the flight of refugees away from the

85

border, said a UNHCR official, was not “one of the scenariosthat was in our plan.”15 Yet the plan’s organizers quicklyresponded to this dramatic shift. The next day, Radio Rwandareported that Dr. Ephriam Kabayija, chairman of the nationalcommission in charge of refugee repatriation, had led a secondRwandan government delegation into Tanzania. He met withTanzanian and UNHCR officials in Karagwe, just north ofNgara. According to one of the Rwandan delegates, the meet-ing “confirmed the total repatriation” of Rwandan refugeesand reviewed how “Tanzania and the UNHCR [would] con-duct the repatriation, while Rwanda and the UNHCR [would]deal with the reception of the Rwandan refugees.”16 However,although aid agencies on the Rwandan side of the borderprepared to receive the immense refugee influx17 and Rwandanauthorities seized 70 suspected Hutu genocidaires near Ngara,18

the Tanzanian army took charge. From this moment on,UNHCR’s role was minimal.

Troops set roadblocks and sealed off access to the formercamps. Helicopters redirected refugees toward the Rwandanarmy at the border. They had no choice but to return to theircountry. The rate of return soon reached 10,000 an hour,matching the rate of their entry into Tanzania 31 monthsearlier.19 The Tanzanians carried out this military maneuver inthe presence of few outside observers, because Tanzanian andRwandan authorities had limited the access of aid agencyofficials and journalists on both sides of the border.20 Inresponse, UN officials focused not on the limited access issue(one merely commented that it was the “official policy” of theTanzanian government)21 but on how Rwandan refugee lead-ers had “fooled us incredibly” by herding terrified Rwandanrefugees away from Rwanda.22

Members of the international community described theunfolding events in two different ways. Amid reports thatTanzanian authorities were employing tear-gas and batons torepatriate Rwandans, a reporter noted how UNHCR officials“appeared content to watch events take their course.”23 “[Tan-zanian] helicopters are circling around,” an UNHCR officialreported, “but just look at how the movement is going.”24

“Many people are walking in the right direction,” anotherUNHCR official observed.25 A WFP document described

86

Tanzanian authorities as merely “assisting with the return ofRwandan refugees.”26 As the refugees crossed into Rwanda,President Pasteur Bizimungu greeted them: “Welcome,Rwanda is your home, help bring peace to rebuild Rwanda.”27

Many journalists and human rights organizations, how-ever, were highly critical toward the activity. “At the border,”British journalist Louise Tunbridge reported, “Tanzanian andRwandan officials kept similarly tight control,” adding thattheir message “was that nothing and nobody would be al-lowed to stop the dismantling of the camps.”28 Anotherreporter wrote that two Roman Catholic priests from Europewere expelled from Tanzania “for allegedly opposing theforced repatriation of some 600,000 Rwandan refugees.”29 Athird observed that “the UN High Commission for Refugeesand aid agencies were muted in their criticism of an operationfew could define as voluntary.”30 U.S. journalist JenniferLudden observed that “The forced return violates the interna-tional principle of voluntary repatriation. Yet even the UnitedNations refugee agency has approved Tanzania’s effort.”31

Human Rights Watch described the event as a forced repatria-tion “that UNHCR and the international community havewatched with little protest.”32

Human Rights Watch later reflected that the refoulementwas “watched with virtually no protest by the internationalcommunity.”33 It also faulted the international community ashaving been “unwilling,” and the Tanzanian government as“unable,” to “screen out combatants or those suspected ofgenocide” from the larger refugee population since their ar-rival in Tanzania in 1994.34 Yet during those years, HumanRights Watch and other human rights groups had treatedNgara as a refugee backwater, focusing instead on Rwandaand Rwandan refugees in Zaire. Its criticism of UNHCR andthe Tanzanian government, though relevant, was very late incoming.

UNHCR downplayed its handling of refugee rights byemphasizing that the power of the Rwandan refugee leaderswas so great, as one official commented, that “voluntaryrepatriation was never a possibility.” Agency officials de-scribed the Tanzanian and Rwandan governments in largelyuncritical terms. One official explained that UNHCR had

87

joined the initial repatriation plan “so that it could betterprotect and assist the refugees” that the Tanzanian andRwandan governments had already decided to return home.Another emphasized how “the return followed bilateral con-tacts between Rwanda and Tanzania.”35 A third blamed campleaders for leaving “the Tanzanian government with littleoption but to contain this mass internal movement and redi-rect it towards the Rwandan border.” Emphasizing UNHCR’sreactive role, these officials also pointed out the decline inUNHCR’s influence on Rwandan refugee issues in Tanzaniasince the emergency in the spring of 1994. They also did notmention the donor governments who had sought to close therefugee camps.

Shifting Coordination Roles

If, as one observer mentioned, “coordination rests wherethe power is, “ then the forced repatriation effort from Tanza-nia in late 1996 demonstrates how sudden, dynamic changesin power relations severely challenge coordinated responses.Over the course of a few weeks, UNHCR’s role as the primarycoordinator for Rwandan refugee activities, carried out effec-tively over a period of 31 months with authority delegated bythe Tanzanian government, was rescinded. The Tanzaniangovernment, whose role at the outset of the emergency waslimited to facilitating the UNHCR-led humanitarian response,took firm command of repatriation in December 1996, leavingUNHCR struggling to find its role.

Piecing together a comprehensive picture of coordinationresponsibilities for the repatriation effort is difficult. There aredifferent recollections among officials in Ngara of what actu-ally took place, and secrecy surrounds some of the activities.Within these constraints, the roles of five groups in the evolv-ing coordination scheme—refugee leaders, the Tanzanian andRwandan governments, UNHCR, and NGOs—will be ana-lyzed in three separate stages.

UNHCR and Rwandan Refugee Leaders. In Ngara, Rwandanrefugee leaders both helped and disrupted humanitarian ef-forts. The centralized coordination scheme relied on theirability to help distribute goods and resolve internal commu-

88

nity problems. UNHCR officials defended their working withthe leaders by arguing that they had no choice. However, theleaders’ involvement with the humanitarian community (evenif many aid officials found it distasteful) enhanced their legiti-macy with other refugees, making voluntary repatriation evenmore difficult. UNHCR’s emphasis on confidence-buildingmeasures to convince the refugees to return home36 neverpersuaded the leaders, who worked actively for the oppositeobjective.

The leaders’ power and influence on repatriation matterscontrasted with UNHCR’s eroding credibility with refugees.One analyst characterized UNHCR’s attitude towards repa-triation as ambivalent during the early months of the emer-gency. The agency relayed a succession of contradictory mes-sages to the refugees: “safe, go now (mid-August, 1994);unsafe, do not go (mid-September 1994); safe again, go now(mid-December 1994).”37 One NGO did not blame the refu-gees’ growing skepticism about UNHCR since “the refugeeshave been lied to by UNHCR since day one” of their sojourn inTanzania. Searching for answers, UNHCR occasionally soughtassistance from NGOs to spur voluntary repatriation, in oneinstance asking NGOs to persuade particularly vulnerablerefugees (e.g., widows and elders) to repatriate.38 When NGOsasked a UNHCR official why UNHCR couldn’t approachvulnerable refugees directly, one NGO recalled, the UNHCRofficial responded, “because the refugees don’t trust us.”

But the reality was more complicated. As another UNHCRofficial noted, the aid community in Ngara had collectivelynudged refugees closer to their leaders by emphasizing thegenocidal crimes that some had carried out while “pass[ing]no views on” the deaths of many refugee family membersduring the civil war that preceded the genocide. This, theofficial concluded, had effectively “reinforc[ed] the socialcohesion and the role of the [refugee] leaders.”39 Overlookingcivil war casualties may have seemed appropriate, given theimmensity of the genocide, but in the eyes of refugees it alsomade the international community defenders of the Rwandanregime and opponents of the refugees.

In December 1996, with its powers of persuasion in declineand repatriation inevitable, UNHCR sought, in the words of

89

one official, finally to “break the grip of the refugee leaders.”While building a transit facility for leaders who would bearrested by Tanzanian authorities, agency officials spoke di-rectly with refugees, including their leaders, to “explain thesituation” about the refugees’ impending return. Officials alsohoped to facilitate the initial repatriation of approximately15,000 Muslim refugees who had already indicated their de-sire to return to Rwanda.40 But when the Rwandan refugeeleaders unexpectedly drove the refugees away from the bor-der, Tanzania took control of the coordination effort andsidelined UNHCR. A UNHCR official later reflected that “Ourmistake was to make [UNHCR] too transparent and let therefugee leaders know what we had in mind.” It also appearsthat UNHCR did not anticipate that the Tanzanian govern-ment would reclaim its authority so completely.

The Tanzanian and Rwandan Governments. Following theNovember 1996 agreement in Dar es Salaam giving Rwandansuntil year’s end to repatriate, a meeting between the parties(Tanzania, Rwanda and UNHCR) in Ngara on December 12transformed the arrangement from a collaborative effort intoa military operation. A UNHCR official who was presentrecalled that the meeting focused on the details of the refugees’return. Before lunch, the head of the Rwandan delegation, Dr.Kabayija, was persuaded to speak to refugees in the camps.The UNHCR official reported that Tanzanian and Rwandanofficials drove “in a motorcade through the camps but neverstopped anywhere. Maybe they were too afraid to talk.” Afterreturning to UNHCR offices for lunch, the motorcade drovepast the Rwandan camps again on its way to Rwanda. A halfhour later, according to the UNHCR official, refugee leaderscarried out what was clearly a well-coordinated plan of theirown to create “absolute chaos.”

Leaders in the three main camps near Ngara “told refu-gees to go... in all directions.” This was their last stand, theofficial explained, an act of desperation reflecting the realiza-tion that they were losing their grip on the refugees. By forcingthem to attempt a desperate escape, they hoped to create anoutcry in the West that would lead people to come to theirrescue.

At this point, in the mid-afternoon of December 12, the

90

framework for coordination between the two governmentsand UNHCR ended. The response orchestrated by the refu-gee leaders ruptured the tripartite plans. Returning fromRwanda and with assistance from Rwandan officials, Tanza-nian authorities immediately undertook a military responseto the refugees’ actions. In the process, Tanzanian officersturned on UNHCR, resurrecting a theme earlier voiced byU.S. government officials. They blamed UNHCR (and addedNGOs to the criticism) for instigating the refugees’ attemptedescape because, as one UNHCR official recalled, “they thoughtthe NGOs and UNHCR didn’t want to lose their jobs” provid-ing refugee services. UNHCR and the NGOs thereafter re-ceived only limited access to the refugees, whom the Tanza-nians proceeded to round up and march to the Rwandanborder.

UNHCR and the NGOs. In Ngara, UNHCR saw its roleshrink. After the Tanzanian army moved to redirect thefleeing refugees towards Rwanda, the agency adjusted itsmandated responsibilities to fit a difficult situation that hadmoved out of its control. Blaming the refugee leaders (not theTanzanians) for ending what one UNHCR official called “thebalanced repatriation plan that had been agreed after muchnegotiation,” the agency assumed a secondary position in therepatriation effort. Its new role became “primarily one ofattempting to ensure that the basic [refugee] rights wererespected during this complex movement.” But in the radi-cally changed context, basic rights meant primarily providingfood and water to refugees as they returned home.

It is worth noting how aid officials assessed Tanzania’sactions. An NGO official in Ngara at the time recalled that“NGOs had to go along with UNHCR announcing [the Tan-zanian move against Rwandan refugees] as a voluntary repa-triation.” Whether UNHCR officials actually believed it ornot, this particular rationale, the official explained, was basedon the following reasoning: “refugees had disobeyed byrunning into the forests... thus calling for disciplinary actioninvolving [Tanzanian and Rwandan] armed forces.” Butsome NGO and UNHCR officials also wondered whetherappropriate disciplinary action against refugees should in-clude refoulement. The judgment provided by a U.S. govern-

91

ment official was more direct: “If you consider it was a forcedrepatriation, it was well run.”

As the Tanzanian military took control of the repatriationoperation, each agency had to decide how to respond. OneUNHCR official described how, after much soul-searching,local UNHCR staff decided “to do what we could to soften thereturn” of Rwandan refugees and to coordinate the effort.Most if not all NGOs agreed to work with UNHCR to assist therefugees during their return. One NGO director commentedthat once this was agreed, “the relationship between UNHCRand NGOs throughout the [subsequent relief] operation wasvery good.” Although the Tanzanian authorities preventedcertain NGOs and a number of specific individuals fromparticipating, most aid agencies were able to carry out thework.

Coordinating assistance to refugees during their march toRwanda was spearheaded by a UNHCR official who had beeninvolved with the Rwandan refugee repatriation from Zaire afew weeks earlier. He applied lessons learned from the Zairerepatriation, including the significance of keeping peoplemoving (to avoid cholera caused by people clustering in onearea); stationing food and water along the way; using string tokeep children attached to their parents or child identificationarmbands to keep families together; and holding nightly meet-ings with all agency officials to adjust their work the followingday.

In the end, what mattered most during the refoulement wasnot international rights or principles but power relations, andin this the Tanzanian and Rwandan governments won out.The forced repatriation appears to have inspired the Tanza-nian government to continue ignoring international criticismand solve refugee problems by force if necessary. Soon afterthe Rwandan repatriation, Tanzanian security forces roundedup some 80,000 to 90,000 Burundian migrants and spontane-ously settled refugees in Tanzania and sent them to camps inTanzania for Burundian refugees. Once again, UNHCR wascriticized for its response to the roundup.41 And once again,scant international attention was paid to the Tanzaniangovernment’s rough handling of refugees.

92

Coordinating Education: The Sierra Leonean Case

Supporting a Refugee Initiative

Refugee camps are not only dismal and difficult places tolive; they are also boring. Once systems are set up to supplybasic necessities such as food, potable water, sanitation andshelter, the monotony of everyday life sets in. Waiting for foodone day is followed by waiting for water the next. This routin-ized way of living is hardly stimulating, but it is particularlydifficult for children. It is thus not uncommon to find refugeeparents organizing informal education and sometimes evenschools in the camps.

When Sierra Leoneans began to seek refuge in Guineafollowing the outbreak of war in 1991, they found Liberianrefugees developing schools in nearby camps. Aid agenciessoon became aware of these refugee schools. In April of thatyear during a joint delegation visit to West Africa by theInternational Rescue Committee (IRC) and its sister NGO, theWomen’s Commission for Refugee Women and Children(WCRWC), aid officials discovered that UNHCR was lookingfor an implementing partner to create an education program.IRC also found that, unlike other host governments, Guineawas prepared to approve the development of formal schoolingfor Sierra Leonean and Liberian refugees.

IRC started its program later that year with the 12,000refugee students who had been attending schools formed byrefugee communities. The coordination structure that it devel-oped was cohesive and effective. With the education respon-sibilities delegated by UNHCR, IRC focused on working withrefugee professionals to enhance and formalize the educationsystem that the refugees themselves had already created. IRCalso reinforced the existing structure by pledging materialsupport to every existing school that met certain criteria.Before IRC would support an existing school, the refugeecommunity needed to demonstrate its commitment by estab-lishing a school on its own, retaining a student enrollment ofat least 200, and organizing to build classroom walls. IRC thenpromised to pay teachers a stipend, deliver school supplies,and build roofs. It later expanded its package to include

93

digging a well and latrine for each school, providing teachertraining, and developing a health education curriculum. Sevenyears after the program began, IRC’s emergency educationprogram included 160 refugee schools and 72,000 students.

By gradually assuming control over the schools, IRCabsorbed many of the education professionals in the refugeecommunity—like many refugee populations, it contained atalented group of teachers, school principals, and governmentadministrators—into the emerging IRC-led education system.IRC’s coordination work then concentrated on the educationministries in Sierra Leone and Liberia. As increasing numbersof refugee schools received support, the next challenge was torespond to refugee concerns that the education received incamp schools be recognized in the refugees’ countries oforigin. A group of IRC curriculum experts developed a unifiedschool curriculum for both Sierra Leonean and Liberian stu-dents. The team had the benefit of working with countrycurricula that were taught in the same language (English) andwere aimed at preparing students for the West African ExamCouncil (WAEC) exams, a standardized test for all students inthe six English-speaking countries of West Africa.

IRC then coordinated with education ministry officials inSierra Leone and Liberia to ensure acceptance of the curricu-lum in both countries. With IRC facilitating interactionsbetween refugee educators and ministry officials in Liberiaand Sierra Leone, the two governments in the end not onlyapproved the curriculum but recognized the teaching certifi-cates and student transcripts issued to refugee returnees. Ameasure of the program’s success came in 1998 when candi-dates achieved the highest pass rate in WAEC history (94percent). UNHCR has written up the program as a model.42

There are those who wonder whether the IRC project,which has been supported by UNHCR and the U.S. govern-ment, has been too successful. It provides better educationthan is available in the refugees’ communities of origin. Manyin the aid community, donors in particular, believe that theemergency education in refugee schools constitutes a magnetthat discourages repatriation.43 While this viewpoint over-looks the 1959 UN Declaration of the Rights of the Child,

94

which considers access to formal education the right of allchildren, there remains a certain legitimacy to the criticism.Refugees have indeed been reluctant to return to their homesin Sierra Leone and Liberia as long as schools there were notfunctioning. 44 With peace largely restored in Liberia, if not inSierra Leone, efforts have begun to focus on repatriatingLiberian refugees. But while UNHCR continues to supportrefugee schools in Guinea, it and other groups in Liberia havebeen slow to help the Liberian government attract refugeereturnees by restarting village schools inside Liberia.

It is here, in the linkage between international aid agenciesand national governments, that coordination broke down. Aformer UNHCR consultant, sent to Liberia in 1998 to investi-gate how to shift the IRC’s educational success in the refugeecamps to areas within Liberia to which refugees would repa-triate, identified two constraints. The first involved a certifica-tion issue. The Liberian education ministry was not particu-larly concerned about the certification of refugee students; ithappened automatically when refugee students passed theinternationally recognized exams. The ministry focused in-stead on the certification of Liberians who had been trained asteachers in the refugee camps.

The central problem the consultant observed was that theIRC lacked documentation for refugee teacher training. Theministry required documentation of the number of hours oftraining received and its specific content. But attendance listsand time sheets did not exist, nor were many teacher trainingmanuals on file. Ministry officials also expressed concernabout the pay differential for Liberian teachers. They foundthat the IRC was paying refugee teachers in their programseven times the salary that teachers received in Liberia (US $70per month in the camps, US $10 in Liberia). The LiberianTeachers Union would surely complain about such inequityand perhaps demand a sevenfold increase in salary, which,given the ministry’s limited resources, was unlikely to occur.Taking the lack of teacher certification and the pay differentialinto account, the consultant concluded that “there’s no incen-tive for [refugee] teachers to return to Liberia.”

The second problem, the consultant continued, involvedthe anticipated role of the IRC following repatriation. The IRC

95

had proposed to accompany the repatriating refugees intoLiberia and Sierra Leone and restart schools with the return-ees.45 The difficulty the UNHCR consultant identified wasthat the IRC would be operating in Liberia with far moreresources than the country’s Ministry of Education. “How canthe Ministry of Education be the lead agency if it has fewresources?” he asked. Armed with donor funding, IRC couldundermine the Ministry in the areas where IRC worked. Itwould essentially become the Ministry of Education therebecause, the consultant reasoned, if the ministry has no re-sources and presence, it has no authority. A second questionwas equally perplexing. If the IRC education program did notaccompany refugees back to Liberia, would refugees return totheir former Liberian communities? Probably not, it ap-peared. Whether IRC’s successful education program in refu-gee camps in Guinea can be transferred into Liberia remains tobe seen. Negotiations with the Liberian government continue.

Refugee Education in Ngara: A Contrast

A comparison between the coordination of emergencyeducation in the Sierra Leonean and Liberian camps in Guineaand that of the education program in Ngara is illuminating.The Ngara program, too, has received a relatively high degreeof attention, nearly all of it positive.46 Yet while both programsalso reached an unusually high number of refugee children,they were fundamentally different. The Ngara approach wasfairly prescriptive and unusually proactive, organized early inthe emergency phase of the refugee crisis (June-July 1994). Theprogram in Guinea, by contrast, was reactive, entering therefugee camps in Guinea long after refugees had arrived andstarted their own schools. The formats also were different. Theemergency educators who assembled at Ngara sought toadapt for Rwandan refugees UNESCO-PEER’s47 Teacher Emer-gency Package (TEP), a three-to-six-month literacy andnumeracy-based approach developed during the Somali emer-gency.48 A gradual return to teaching the Rwandan curricu-lum would follow. The Ngara approach was also ambitious:the organizers envisioned that TEP schools could reach 40,000to 50,000 refugee students, a goal that was eventually achieved.

96

The Ngara program also involved a much wider array ofagencies in a coordination structure similar to the CFA inSierra Leone. Like the CFA, the top and middle segments of thestructure were dominated by international aid agencies. TheNgara structure was led by the education program manage-ment unit (EPMU), which contained UNESCO-PEER, UNICEF,and the German government donor GTZ. The EPMU wastasked with the day-to-day, week-by-week supervision of theeducation program which included providing technical ad-vice and guidance to the NGOs and Rwandan educators whoimplemented the program. Also, the EPMU served the educa-tion coordination committee, which included representativesof all international agencies participating in the program. Alsoinvolved in the coordination scheme was UNHCR’s educationofficer, a position filled, in an unusual arrangement, by theECHO field officer. The education program was also part ofthe community services committee, a coordinating body runby UNHCR.

Like its counterpart in Guinea, the UNHCR sub-office inNgara gave the education sector a high degree of indepen-dence. A UN official involved in developing the programcredited sub-office head Maureen Connelly for “personallymaking a difference by virtue of allowing greater openness,user-friendliness and flexibility, with an orientation towardsfinding practical solutions.” At the same time, the three NGOsworking in the Rwandan refugee camps in Ngara, each re-sponsible for the education program in a different camp,“probably experienced the EPMU as too ‘top-down’ and au-thoritarian.”

The original developers of the program, a group of UNand donor officials but not NGOs, were concerned that “the‘bottom-up’ orientation and community-sensitive proclivitiesof the NGOs needed to be tempered by a well-defined frame-work to which they signed up and were responsible for.” Thefinal link in this coordination structure, interacting with boththe EPMU and the NGOs, were project development officers,“the key Rwandan focal point for the education program ineach camp.”

The Ngara approach was comparatively more complexand hierarchical than the ICRC model in Guinea, with refugee

97

communities and educators on the receiving end of interna-tional agency directives. This may appear unfair to theRwandan educators, but it was probably quite appropriate;for evidence from Rwandan refugee camps in Zaire indicatesthat the IRC model could not have succeeded with Rwandans.There, Rwandan refugee leaders started the refugee schools,as aid agency officials were largely uninterested in education.It was widely assumed that these leaders used the schoolsprimarily to promote the sort of ethnic history that had pro-vided the rationale for genocidal killing in Rwanda.49

The limited degree of Rwandan refugee contributions tothe coordination structure has been identified as a weakness ofthe Ngara model. A UN education official present there laterobserved that Rwandan refugee teachers asked, “Why are yougiving us [TEP]? It’s not Rwandan.” They would havepreferred to resume teaching in camps immediately accordingto the Rwandan curriculum. It would be unwise, however, tosay that reactive, refugee-responsive program models arealways better. In terms of program coordination, the IRCapproach was simpler and more responsive to refugee initia-tives. It involved one NGO interacting with many refugeecommunities while the government and UNHCR remained inthe background.

The Ngara education approach, as well as the Tanzaniangovernment-led refoulement in December 1996, shed light onthe utility of proactive programming when refugee popula-tions are led by people as dangerous and difficult as those inthe Rwandan camps. The refoulement experience in Ngara alsohighlights the importance of the careful maintenance of goodrelations between international humanitarian actors and na-tional governments. In this case, UNHCR’s relations with theTanzanian government were generally good throughout, andit should be commended highly for this. Still, subsequentevents suggest that the government’s role may have been tooperipheral at the start. Like other governments, the Tanzaniangovernment is naturally sensitive about sovereignty issues,and providing it with a substantive coordination role from theoutset may have reduced the force of the government’s laterresponses.

Inconsistent relations between international NGOs and

98

the Tanzanian government certainly did not improve matters,nor did donor warnings in 1996 that they wanted to withdrawsupport for Rwandan refugees in Tanzania. But the irony isthat refugees, so often passive recipients of humanitariansupport from powerful international actors, ultimately holdsuch sway over humanitarian actors when repatriation issuesare concerned. In both the Tanzanian repatriation programand the repatriation problem for the IRC’s refugee educationprogram in Guinea, prominent refugee leaders (politicians inNgara, teachers in Guinea) wielded considerable power overthe timing and nature of refugee returns.

And for both sets of leaders, the nature of this powernegatively impacted repatriation efforts. In Ngara, Rwandanrefugee leaders ultimately overplayed their hand, as their planto drive refugees into the Tanzania countryside opened thedoors for the final refoulement exercises. In refugee camps inGuinea, Liberian refugee teachers’ reluctance to return toLiberia and the IRC’s difficulties with Liberia’s Ministry ofEducation eventually transformed the IRC’s acclaimed educa-tion program into a magnet keeping refugees in camps andoperating as a disincentive against repatriation.

99

CHAPTER 6

CONCLUSIONS AND RECOMMENDATIONS

Coordination should not, in principle, be difficult to or-chestrate. Humanitarian organizations, after all, are on thescene of humanitarian crises to provide emergency assistanceand to protect the human rights of victims of violence, perse-cution, and other misfortunes. The quality and effectiveness oftheir efforts stands to be enhanced if they work together in asystematic way.

The two cases described in this study, however, demon-strate the difficulties humanitarian actors face in coordinatingtheir work. Rapidly evolving changes in Sierra Leone and inRwandan refugee camps in Tanzania called for creative re-sponses. Humanitarian actors were sometimes equal to thechallenge, but on other occasions failed to keep pace with—much less to anticipate—changes in the political, military, andhumanitarian situation on the ground. Experiences were bothpositive and negative at each of the three levels of coordinationexamined: among organizations, among functions, and withinprograms.

Regarding the coordination of humanitarian organiza-tions, relationships among agencies and the people in posi-tions of responsibility facilitated the creation of agile arrange-ments. The esprit des corps that emerged among agencies anddonor officials at the outset of the Rwandan refugee emer-gency in Ngara District provides one example of this ten-dency. Conversely, however, relationships also impeded co-ordination. The logjam created by three separate UN coordi-nating entities jockeying for position in Sierra Leone providesa second example. Regarding the coordination of basic hu-manitarian functions, the experience was also mixed. Al-though strategic planning for the food sector in Sierra Leonewas largely successful, coordinating the consolidated appealprocess had serious limitations. Regarding the coordination ofhumanitarian programs, the repatriation of Rwandan refu-gees from Tanzania illuminated how dynamic changes in thepolitical context could severely strain constructive partner-ships and alter relationships among the actors involved. On

100

the other hand, the IRC’s educational programming in Guineastabilized the lives of many Sierra Leonean and Liberianrefugees while constraining repatriation efforts.

Three commonly held beliefs about coordination emergedamong many of those interviewed. First, most assumed thatthe task of coordination naturally centered on UN agenciesand international NGOs, not donor and national govern-ments. In point of fact, it was both sets of governments thatestablished and managed the framework for coordination.Second, many officials countered that money impeded coordi-nation, since institutions seek to command the lion’s share ofthe action in a particular sector, camp, or region, even at theexpense of coordination with other humanitarian actors. YetUNHCR’s control over resources had positive effects in Ngarawhile, conversely, the absence of similar resources and au-thority vested in Humanitarian Assistance Coordination Unitin Sierra Leone undermined its coordination prowess. Finally,many officials assumed that coordination depended foremoston the personalities of those at the helm, while coordinationstructures were of secondary importance. The dynamics onthe ground, however, underscored the importance of agreedcoordination structures that reinforce assertive and creativeleadership. The following are conclusions that address theseand other issues emerging from the study.

Coordination: Donor and National Government Roles

Donors and national or host governments often seem onthe periphery of coordinated humanitarian action. The twocases examined here demonstrate, however, that governmentsare absolutely crucial to providing the framework withinwhich the challenge of humanitarian coordination is set.

The Power of Donors

No single set of humanitarian actors has more power toinfluence coordination positively than donors. Through theirdecisions, they sculpt humanitarian action. Donors can stabi-lize an entire coordination system, such as when BPRM andOFDA put their funding through UNHCR for the Rwandan

101

refugee response in Ngara. They can enforce sector coordina-tion, as when OFDA and ECHO mandated that all agenciesreceiving food aid participate in CFA coordination or risk acutoff in assistance. They can also influence power relationsbetween other actors, as in their decision to favor NGOs overgovernment ministries in Sierra Leone or to help forceUNHCR’s hand in Ngara in 1996 by insisting that Rwandanrefugee asylum in Tanzania end soon, setting the stage for themilitary-led repatriation just before Christmas.

In these instances, the U.S. government (OFDA in SierraLeone and the U.S. State Department’s Bureau for Population,Refugees, and Migration in Ngara) and the European Com-munity worked together well. One donor official reflected thatin Sierra Leone “[we] actually compare notes on what we’llfund for the same NGO. I’ve never done this before in othercountries. The NGOs often hate this, as in Liberia where wediscovered that we were double-funding” the same NGOs tocarry out the same activity.

An indicator of donor power is the hesitancy of otheractors to criticize them. Sierra Leonean officials challengedinternational NGOs but not the donors who had favoredNGOs over the local government. Criticism for the Ngararepatriation was directed at UNHCR and the Tanzanian au-thorities, not at the U.S. and the EU, which had announced animminent cutoff of aid for the refugees. Although it is hardlysurprising that recipients of largesse are reluctant to bite thehand that feeds them, the results of donor decisions receive farmore attention than the decisions themselves, even when theyhave significant humanitarian fallout.

The Roles of the National Government Authorities

How should international agencies involve host govern-ments in humanitarian work? The coordination structuresreviewed here intentionally restricted host government con-tributions. Indeed, the two cases suggest that internationalactors viewed humanitarian coordination as working bestwhen national governments are marginalized from the pro-cess. Host governments that are low-profile and pliant ratherthan fully engaged—such as the Tanzanian authorities letting

102

UNHCR decide which NGOs would be admitted—are seen asmaking for better coordination.

This approach is shortsighted in two respects. First, itprovides international humanitarian action with a fragile localbase. Since national governments are the ultimate arbiter ofactions on their territory, they can reassert themselves when-ever they wish. That became crystal clear in Ngara. Whileinitial coordination arrangements were without doubt facili-tated by the Tanzanian government’s limited participation,the languishing repatriation effort was simply taken over bythe Tanzanian government and military in late 1996, whoorchestrated the refoulement on their terms. In Sierra Leone, theKabbah government’s simmering resentment against “JuntaNGOs,” whom it accused of secretly supporting the RUF,eventually boiled over. In January 1999, it expelled the ICRCfor assisting the RUF, a charge the ICRC strongly denied. Ittook months for the ICRC to negotiate its return and resumeoperations.

Second, relief agencies tend to focus on short-term hu-manitarian needs instead of longer-term reconstruction ordevelopment work. Even though they are often preoccupiedwith limiting the dependence of displaced populations onforeign aid, they are far less concerned about the dependentand even subordinate relations they often create with nationalgovernments. Struggling to provide food to Sierra Leoneancitizens, CFA members felt that including government offi-cials would hamper and politicize their work. Yet despite thelow priority on capacity building, the CFA will ultimatelyhave to hand over its responsibilities to the Sierra Leoneangovernment.

Indeed, the urge towards pragmatism in humanitarianaction—to empower those actors with the best potential to getthings done—carries a heavy cost. In the Rwandan refugeecamps, collaboration with indigenous leaders had major po-litical and humanitarian complications. In Sierra Leone, thedecision by donors to fund international NGOs, creating whatthe authorities view as a parallel system with government,may create serious governance problems in the future. Acoordination apparatus that functions smoothly by virtue ofsidelining the national and local authorities may reap short-

103

term gains but undermine national government authority andlimit its capacity.

In the Sierra Leone and Ngara cases alike, the roles playedby governments—the donors of resources and the host gov-ernments in the crises—form the backdrop for the coordina-tion efforts of humanitarian organizations. The followingconclusions are arranged according to the three focal points ofcoordination analyzed in the study: coordination among orga-nizations (Chapter 3), among functions (Chapter 4), and withinprograms (Chapter 5).

Coordination Among Organizations

In Sierra Leone, UNOMSIL stood as the leader of the UNfamily, UNDP’s head served as humanitarian coordinator,and HACU carried out a variety of coordination operations. Itwas a confusing and dysfunctional combination.

UNDP was a lightening rod for coordination controversy,largely because it resolutely retained development principlesand objectives while coordinating humanitarian activities. Itsclose association with the Sierra Leonean government ex-tended UNDP’s prewar development policies into the conflictyears. By inflaming humanitarian actors seeking to operate asneutral providers of humanitarian assistance in governmentand rebel-controlled areas alike, UNDP undermined its role ascoordinator of humanitarian action. This situation also illumi-nated a longstanding weakness in humanitarian coordination:the fact that the UN and NGO agencies lack any set of formalarrangements for humanitarian coordination. Though a re-cent report highlighted the problem with regard to UNHCR inrefugee settings, the same problem plagues UN-NGO rela-tions inside countries overwhelmed by humanitarian crises:“The formal authority enshrined in the lead [UN] agencydesignation does not extend to the NGOs.”1

HACU had more success coordinating with an array ofinternational and national actors, particularly at the sectorlevel. But the Consolidated Appeals Process exposed its diffi-culties vis-à-vis other UN agencies. HACU had the responsi-bility of assembling the CAP but lacked the authority to ensurethat the UN activities presented in the appeals were coordi-nated among individual agencies.

104

In Ngara, by contrast, the two major donors funneled theirsupport through UNHCR. The Tanzanian government del-egated to UNHCR the authority to select all NGO partners thatwould work with Rwandan refugees there. UNHCR in Ngaraseized this rare opportunity to centralize control and it createdan impressive coordination structure. The unusual arrange-ment worked out both for donors and for the Tanzaniangovernment and it illuminated three important coordinationissues. First, it shed light on the difficulties of working withpowerful but not necessarily popular Rwandan refugee campleaders who earned their “intimidators” nickname. AlthoughUNHCR may have initially had little choice but to work withthem, it did little to alter its relationship with these leadersover time or to cultivate others for leadership positions, andthis eventually hurt its credibility with the refugee population.Second, the Ngara arrangements illuminated the importanceof the national government’s role in coordination. Though theTanzanian authorities ceded some powers to UNHCR duringthe initial stages of the refugee crisis, the repatriation showedhow provisional UNHCR’s powers ultimately were. Third,the Ngara experience shed light on the role personalities playin coordination. In UNHCR Sub-Office Head MaureenConnelly and Deputy Head Francois Franquin, UNHCRinitiated its work with two extremely capable officials. Thoughthe contributions of others were crucial—the Tanzanians al-lowed, donors reinforced, and NGOs accepted the coordina-tion structure—UNHCR’s forceful and focused leaders cre-ated an unusually well-coordinated humanitarian response.

By contrast, the attacks on Elizabeth Lwanga, head ofUNDP in Sierra Leone, which were particularly strident fromNGO members, ultimately personalized an institutional prob-lem. There was a built-in contradiction between a coordinatorof humanitarian action premised on neutrality who was alsocharged with maintaining close relationships with the govern-ment authorities in the midst of a civil war. While Ms. Lwanga’smanagement style may have antagonized NGOs and exacer-bated coordination, UNDP’s presence at the coordinationhelm proved to be a bad fit.

105

Coordination Among Functions

The three specific functions reviewed were strategic plan-ning, resource mobilization, and security information man-agement.

The examples of strategic planning cited were innovativeand worthy of adaptation to other contexts. In Sierra Leone,the CFA, apart from its weak links to the government authori-ties, represents an impressive innovation for food sector coor-dination. Indeed, the coordination framework was adaptableto other sectors as well. In Tanzania, the value of a cleardelegation of authority by the government to UNHCR and ofa reduced number of NGO actors were noteworthy innova-tions.

In the Ngara case, however, a cautionary lesson is alsoinvolved. The value of UNHCR doing more to integrate theTanzanian authorities into the coordination scheme and ofNGOs doing more to respect the Tanzanian government’sroles is clear. Indeed, the experience as it played out was suchthat the Tanzanian authorities, faced with a similar emergencyin the future, might not again grant such broad powers tointernational humanitarian organizations.

With respect to resource mobilization, it is distressing thatECHO viewed coordinated funding with the U.S. StateDepartment’s funding in Ngara—by its own account, positivein its outcomes—as an exception to its customary direct fund-ing of NGOs. The impressive ECHO-OFDA coordination inthe case of Sierra Leone should be attempted in other settingsrather than viewed as exceptional. BPRM’s funding ofUNHCR’s work in Ngara made a positive contribution on thecoordination front.

Donor skepticism of the CAP process, as evidenced inSierra Leone, and the general view of UNHCR officials bysome donor agencies also signals a major problem betweenUN agencies and donors that requires attention. The consoli-dated appeals for Sierra Leone lacked both cohesion amongUN agencies and with non-UN aid groups working in thesame sectors. While HACU was more successful in reachingoutside the UN system, it lacked the authority within the UNto manage coordination efficiently.

106

As for security information management, in both Ngaraand Sierra Leone successful management of security informa-tion took place only after UN regulations aimed at preventingUN agencies from sharing information and coordinating secu-rity with NGOs were circumvented.

The politicization of information management regardingshared security problems in Sierra Leone, which became espe-cially difficult during the junta period, highlighted both thesignificance of this issue and the seemingly intractable task ofcoordinating humanitarian work with actors divided betweenpolitical partisanship and neutrality. In Conakry, no institu-tion provided the necessary leadership to resolve the disputebetween these two sets of actors. The refusal by many UN andinternational institutions to take seriously security informa-tion provided by the ICRC and NGOs based inside SierraLeone greatly hampered coordination efforts and underminedeveryone’s credibility. Ironically, the dispute was resolvablein that the two sets of findings were largely complementary:the ICRC-NGO group mostly reported on conditions outsideof Freetown while the Conakry group concentrated on activi-ties in Freetown.

Following the return of aid agencies to Sierra Leone,UNOMSIL’s involvement in coordination issues, otherwisemostly peripheral, attracted considerable attention for its rolein managing security information. UNOMSIL made little ef-fort to work with humanitarian actors, particularly NGOs. Itsdisinterest in coordination exacerbated problems in resolvingsecurity problems and was unnecessary and counterproduc-tive. The situation highlighted a condition that underminedcoordination—the tendency of humanitarian actors to go theirown way rather than resolve differences. UNOMSIL, in short,carried on military observer and human rights monitoringwork that promised to affect the security of humanitarianactors, yet failed to coordinate with them.

In Ngara, overlooking security issues at the outset of therefugee crisis was a mistake, if an understandable one, giventhe magnitude and urgency of the situation at the time. TheGatete incident, however, revealed how humanitarian actorscan learn from such mistakes. The experience suggests thatworking swiftly at the onset of emergencies to separate actual

107

or suspected fugitives from justice from bona fide refugee orIDP populations.

Coordinating Within Programs

The seeds of UNHCR’s problems in the unfortunaterefoulement fiasco of Rwandan refugees from Tanzania in late1996 started with difficulties in repatriating refugees in timelyfashion and coordinating work with refugee leaders. In addi-tion, more active collaboration earlier on with the Tanzanianauthorities might have helped avoid or minimize the ensuringdifficulties.

UNHCR’s options by November 1996 were seriously con-strained by donors and the Rwandan and Tanzanian govern-ments. The refugee leaders with whom the agency tried tonegotiate exercised a constraint as well, especially after theydisbursed half a million Rwandans into the Tanzanian inte-rior. At that point, UNHCR muted criticism of the Tanzaniangovernment for its rough handling of the refugees. Althoughthis approach was regrettable, UNHCR should not be singledout. Western governments who funded the Rwandan refugeecamps also shied away from criticizing Tanzania’s treatmentof the refugees. From the point of view of the Tanzanian andRwandan authorities, the repatriation effort was a success.UNHCR and its partner NGOs in Ngara coped with an ex-tremely difficult situation well. On balance, the assessment ofa U.S. official is probably about right: for a forced repatriation,it was well run.

Coordination of the schooling mounted by the IRC, whichinvolved educators among the refugees, UNHCR, and thegovernments of Guinea, Sierra Leone, and Liberia, was gener-ally successful and innovative. IRC negotiations with the threegovernments in developing a standardized curriculum andfacilitating student participation in regional exams made apositive contribution to the stabilization of the refugees’ situ-ation. The approach bears replication elsewhere if two prob-lems are addressed. First, refugee schools should be engagedonly if they are not being used to promote violence, as was thecase in Goma, Zaire. Second, IRC’s negotiation of teachercertification and salaries did not align with those existing in

108

the refugees’ country of origin. Coordination with educationministries in the refugees’ home countries should begin earlyso that there is greater consonance on matters of certificationand salaries between camps outside and the situations backhome.

Recommendations

The Sierra Leone and Ngara cases illuminate the impor-tance of a command element in humanitarian coordination. Inboth contexts, strong authority represented a necessary ingre-dient in successful coordination. UNHCR’s dominant role inNgara demonstrated this best: the agency received the author-ity to create a coordination structure and generally used thatauthority effectively. Conversely, HACU’s attempts to de-velop consensus among UN agencies for consolidated appealsfailed because HACU lacked the authority to enforce coordi-nation. The CFA in Sierra Leone, born out of a shared agencyconviction that arrangements beyond business-as-usual wereimperative, benefited from the requirement by donors thatagencies receiving food aid actively participate in the coordi-nation process.

These two cases also highlight the importance of establish-ing collegial and effective working relationships with nationalauthorities. In the Tanzania case, an effective delegation ofauthority to UNHCR by the Tanzanian government was ulti-mately rescinded. In Sierra Leone, the weak relations that manyhumanitarian actors had with the national government under-mined linkages with development initiatives and nurtureddiscontent among Sierra Leonean government officials anddisregard among many international humanitarian officials.

Each of these recommendations—the need to incorporatea command element and to engage the national politicalauthorities in the coordination framework—has major impli-cations for each of the sets of humanitarian actors listed below.

Donor Governments

• Higher levels of confidence are needed vis-à-vis UNagencies. The serious doubts expressed by donors

109

regarding the consolidated appeal process and thesecurity of personnel should be discussed and re-solved. Important innovations such as the joint ECHO-U.S. government funding of UNHCR in Ngara wouldthen become more replicable, and ECHO’s hesitancyto work routinely with UNHCR and other UN agen-cies would be addressed.

• Efforts should also be made to replicate the excellentfield coordination between OFDA and ECHO in Si-erra Leone.

National Actors

• The limited participation of national governments inhumanitarian coordination in these two settings callsfor careful reconsideration of the urge among interna-tional organizations to “get things done.” Nationalgovernments need a well-defined, appropriate and,most important, recognized coordination role. Con-cerns over limited capacity or corruption should bediscussed openly. Greater involvement may help dem-onstrate some of the synergies between relief anddevelopment and reduce the likelihood that nationalgovernments will disrupt humanitarian action. TheCFA should be adapted for replication elsewhere,provided its blind spot on indigenous involvement isaddressed.

• National NGOs played an insignificant role in humani-tarian action and coordination in Sierra Leone andNgara. Local capacity-building was defined by inter-national humanitarian organizations mainly as thehiring and training of local personnel to work forthem. International actors should expand their sup-port of civil society by working with national NGOs toidentify and expand their roles.

United Nations Agencies

• Although UNDP has an important role to play in hu-manitarian emergencies, the opportunity costs of theUNDP resident representative serving as humanitar-

110

ian coordinator are too high. Since UNDP’s institu-tional orientation is a poor starting point for creatingcohesive coordination arrangements with humanitar-ian actors seeking positions of neutrality, positioningthe Resrep as humanitarian coordinator should beavoided.

OCHA is the logical choice for humanitarian coor-dination responsibilities, particularly if, as OCHA’sHACU office did in Sierra Leone, it works to includea wide range of humanitarian actors in coordinationstructures outside as well as inside the UN. To suc-ceed, however, OCHA should be given the authorityto persuade, and, if necessary, compel, UN agencies tocoordinate their activities. UNDP could then apply itsexpertise to support the national government andhelp reconstruct war-torn countries, but only afterOCHA identifies the appropriate time for startingsuch activities.

• Since the lack of formal coordination arrangementsbetween the UN and NGOs undermines the effective-ness of humanitarian activities, a generic memoran-dum of understanding (MOU) between the UN andNGOs should be developed. This MOU should beapplicable early in humanitarian emergencies whenagencies are setting out their relationships, and couldbe adapted to the specifics of each setting. It shouldincorporate the reform of security coordination toinclude NGOs in information-sharing measures.

• The consolidated appeal process is a weak coordinationmechanism, reflecting the coordination weaknessesof the UN system itself. Although the concept is sound,the Sierra Leone experience highlights the problemsinherent in OCHA’s lack of authority to coordinateother UN agencies, some of which did not wish to becoordinated. Reforming the process both to ensureinteragency discipline and to connect UN actions withnon-UN actors would constitute a significant coordi-nation advance. Until this and other institutional dis-connections within the UN system are solved—in-cluding the conceptual confusion over how humani-

111

tarian action relates to development and peacekeep-ing on the ground—coordination will continue to befrustrated.

NGOs

• Of all major humanitarian actors, the international NGOs’disregard of national government actors in both theSierra Leone and Ngara cases is a cause for particularconcern. This approach is counterproductive for bothhumanitarian action and development linkages. Allactors, but especially international NGOs, need todramatically improve their working relationships withnational government counterparts and appreciate thesovereignty concerns that national governments natu-rally and appropriately maintain.

• Narrowing the number of NGOs in a particular humani-tarian theater, as demonstrated by the Ngara case, isan innovation that should be widely replicated. Yetthis runs the risk of making the larger, more estab-lished NGOs even more dominant than they are now.As a safeguard, NGOs should develop, together withUN agencies and donor governments, a system forrotating NGO participation in humanitarian emer-gencies. Precautions would ensure that institutionalcapacity and specialization are recognized and used.In such a system, an NGO already responsible for thewater and sanitation sector in one humanitarian emer-gency would not be eligible to do so in the the next,provided another NGO with similar expertise couldbe identified. Limited NGO participation in each emer-gency would thus not limit the worldwide contribu-tions of NGOs more broadly.

Changes in humanitarian coordination along the linesoutlined above would position the international communityto respond more dynamically to future emergencies.

113

NOTES

Chapter 1

1. Jonathan Benthall, Disasters, Relief and the Media (London andNew York: I.B. Tauris & Co. Ltd., 1993): 135.

2. Antonio Donini, The Policies of Mercy: UN Coordination inAfghanistan, Mozambique, and Rwanda, Occasional Paper #22 (Provi-dence, R.I.: Watson Institute, 1996): 97.

3. Random House, Random House Dictionary of the English Lan-guage, Jess Stein, ed. (New York: Random House, 1981): 321-322.

4. Larry Minear, U.B.P. Chelliah, Jeff Crisp, John Mackinlay, andThomas G. Weiss, United Nations Coordination of the InternationalHumanitarian Response to the Gulf Crisis, 1990-1992, Occasional Paper#13, (Providence, R.I.: Watson Institute, 1992): 3.

Chapter 2

1. Integrated Regional Information Network for Africa (IRIN)Weekly Round-up, March 4, 2000.

2. Robert D. Kaplan, “The Coming Anarchy,” Atlantic Monthly(February, 1994): 44-76.

3. Though often questioned by observers, this figure has beencited, among others, by United Nations Children’s Fund (UNICEF) inAfrica Research Bulletin, (December 1-31, 1996): 12514; Rod Mac-Johnson,“The Healing Begins,” in African Comebacks: Strategies for Survival(February 1997): 10; and in the Economist Intelligence Unit (EIU) Coun-try Report for Guinea, Liberia and Sierra Leone (April-June 1998): 29.

4. A report on RUF amputations in Human Rights Watch, “SierraLeone: Sowing Terror; Atrocities against Civilians in Sierra Leone,”Human Rights Watch Report 10, no. 3A (July, 1998) is long on detail butshort on context, but it documents amputation cases well.

5. For more information, see especially Paul Richards, Fighting forthe Rain Forest: War, Youth and Resources in Sierra Leone (London: JamesCurrey, and Portsmouth, N.H.: Heineman, 1996), still the most exten-sive and significant research on the RUF and the war. See also JanGoodwin’s “Sierra Leone is No Place to Be Young,” New York TimesMagazine, February 14, 1999, Section 6, 48-51; Marc Sommers’ “Fight-ing children in Sierra Leone,” Providence Journal, January 15, 1999, B6;Marc Sommers, “Beware the Children’s War,” St. Petersburg Times,August 10, 1997, 6D; and Marc Sommers, The Children’s War: TowardsPeace in Sierra Leone (New York: Women’s Commission for RefugeeWomen and Children, 1997).

6. Economist Intelligence Unit, Country Profile: Guinea, Liberia,

114

Sierra Leone 1997-98 (London: Economist Intelligence Unit, 1997): 44.7. Economist Intelligence Unit, Country Profile, 45.8. Mark Bradbury, “Rebels Without a Cause? An Exploratory

Report for CARE Britain on the Conflict in Sierra Leone,” 22.9. Richards, Fighting for the Rain Forest, 4.10. See Sommers, The Children’s War.11. Richards, Fighting for the Rain Forest, 1.12. Sommers, The Children’s War, 4.13. See, for example, “West Africa according to Mr. Taylor,” Africa

Confidential 40, no. 2 (January 1999): 2.14. From “Sierra Leone: A Brief Overview,” an unpublished

document by Joseph Opala, Freetown, April, 1996.15. Norimitsu Onishi, “Pacts Reached on Congo and Sierra Leone,”

New York Times, July 8, 1999, A8.16. See, for example, African Rights, Rwanda: Death, despair and

defiance (London: African Rights, 1994/95); Basil Davidson, “OnRwanda,” London Review of Books, August 18, 1994; A. Des Forges,Leave None to Tell the Story: Genocide in Rwanda (New York, Washing-ton, London, Brussels: Human Rights Watch, and Paris: InternationalFederation of Human Rights, 1999), A. Destexhe, Rwanda and genocidein the twentieth century (London and East Haven, Conn: Pluto Press,1995); F. Keane, Season of Blood: A Rwandan Journey (London, NewYork, Victoria, Toronto, and Auckland: Viking, 1995); P. Gourevitch,We Wish to Inform You That Tomorrow We Will be Killed With OurFamilies: Stories from Rwanda (New York: Farrar Straus & Giroux,1998); G. Prunier, The Rwandan Crisis: History of a Genocide (New York:Columbia University Press, 1995); F. Reyntjens, “Rwanda: Backgroundto a genocide,” Bulletin des seances, Academie Royale des Sciencesd’Outre-Mer, Brussels, April, 1995; and P. Watson, “Purging the Evil,”Africa Report (November-December 1994).

17. The United Nations and Rwanda, The United Nations BlueBooks Series 10 (New York: United Nations Department of PublicInformation, 1993-1996): 73.

18. Prunier, The Rwanda Crisis, 136.19. Marc Sommers, Reconciliation and Religion: Refugee churches in

Rwandan camps, Occasional Paper (Uppsala, Sweden: Life and PeaceInstitute, 1998).

20. Prunier, The Rwanda Crisis, 265.21. Prunier, The Rwanda Crisis, 257.22. Jean-Pierre Chretien, cited in Prunier, The Rwanda Crisis, 247.23. Prunier, The Rwanda Crisis.24. Sommers, Reconciliation and Religion.25. Larry Minear and Phillippe Guillot, Soldiers to the Rescue:

Humanitarian Lessons from Rwanda (Providence, R.I.: Watson Institute,and Paris: Organization of Economic Co-operation and Development,1996): 175.

115

Chapter 3

1. The UN Under Secretary-General for Humanitarian Affairssigns off on all humanitarian coordinator appointments, which arealso vetted by the UN Inter-Agency Standing Committee.

2. The assumption of the UNDP resident coordinator as humani-tarian coordinator has been the approach taken by the UN to thecoordination task in many major emergencies. However, in somecountries, the major operational agency (e.g. UNHCR, UNICEF, orWFP) may be given humanitarian coordination responsibilities. Thevarious models and their effectiveness is currently being reviewed byOCHA.

3. Emergency Response Division, UNDP, Building Bridges Be-tween Relief and Development: A Compendium of the UNDP Record inCrisis Countries (New York: Emergency Response Division/UnitedNations Development Program, [undated]).

4. For a discussion of the structural problems faced by all UNhumanitarian organizations in preserving neutrality and access incivil wars, see Larry Minear, “Learning to Learn,” in HumanitarianCoordination: Lessons Learned, Report of a Review Summer, Stockholm,April 3-4, 1998 (New York: OCHA, 1998): 27-45, esp. 30-32.

5. Humanitarian Assistance Coordination Unit (HACU), NationalHumanitarian Coordination Structure (Sierra Leone: HACU, 1998).

6. “ECOWAS Six-Month Peace Plan for Sierra Leone,” 23 Octo-ber23, 1997-April 22, 1998.

7. “Interim report of the inter-agency mission to Sierra Leone.”UN Security Council Document S/1988/155, February 25, 1988.

8. These observations reflect comments made at an informalOFDA-ECHO dialogue on October 28, 1998, which was facilitated bythe Humanitarianism and War Project.

9. Antonio Donini, The Policies of Mercy: UN Coordination inAfghanistan, Mozambique and Rwanda, Occasional Paper #22 (Provi-dence, R.I.: Watson Institute, 1996): 14.

10. Malkki observes that UNHCR estimated that 245,000 Burndianshad crossed into Tanzania in late 1993 and early 1994, while thegovernment of Tanzaina claimed there were more than half a millionnew Burundi refugees in Tanzania. Liisa H. Malkki, Purity and Exile:Violence, Memory, and National Cosmology among Hutu Refugees inTanzania (Chicago and London: University of Chicago Press, 1995):287.

11. John Borton, Emery Brusset, Alistair Hallam, et. al., “Humani-tarian Aid and Effects, Joint Evaluation of Emergency Assistance toRwanda (JEEAR) Report: The International Response to Conflict andGenocide: Lessons from the Rwandan Experience 3” (Copenhagen:Steering Committee of the Joint Evaluation of Emergency Assistance

116

to Rwanda, March 1996): 29.12. Borton, et al., “Humanitarian Aid and Effects,” 29.13. Borton, et al., “Humanitarian Aid and Effects,” 130.14. Borton, et al., “Humanitarian Aid and Effects, 130.15. Borton, et al., “Humanitarian Aid and Effects,” 29.16. Borton, et al., “Humanitarian Aid and Effects,” 29.17. “UNHCR Mission Report: Ngara, Tanzania,” April 15-July 17,

1994.18. Maureen Connelly, NGO coordinator, UNHCR; formerly

UNHCR sub-office head for Ngara.19. “UNHCR Mission Report,” 3, 1.20. “World Disasters Report: 1997,” International Federation of

Red Cross and Red Crescent Societies (Oxford: Oxford UniversityPress, 1997): 13.

21. “UNHCR Mission Report.”22. “World Disasters Report: 1997,” 13.23. “Humanitarian Aid and Effects, (JEEAR) Report,” 128.24. “Humanitarian Aid and Effects, (JEEAR) Report.”25. For more information, see Marc Sommers. Reconciliation and

Religion: Refugee churches in Rwandan camps, Occasional Paper (Uppsala,Sweden: Life and Peace Institute, 1997).

26. See also Bonaventure Rutinwa, “Beyond Durable Solutions:An Appraisal of the New Proposals for Prevention and Solution of theRefugee Crisis in the Great Lakes Region,” Journal of Refugee Studies 9,no. 3 (1996): 312-325.

Chapter 4

1. Based on the CFA’s “Memorandum from the National Techni-cal Committee on Food Aid to Members of the Committee on FoodAid,” September 29, 1998, Freetown, Sierra Leone.

2. Lance Clark, Key Issues in Post-Emergency Refugee Assistance inEastern and Southern Africa (Washington, D.C.: Refugee Policy Group,September 1987): 14.

3. Maureen Connelly Senior Emergency Officer, Emergency Re-sponse Mission, Kagera, Tanzania, 15 April-17 July 1994, UNHCR MissionReport, No. 94/006, UNHCR, unpublished document, 1.

4. Tony Waters, “The Relief Effort Winds Up, October 1993-October 1994,” unpublished draft of Bureaucratizing the Good Samari-tan (Boulder, Colo.: WestviewPress, forthcoming): Chapter 2.

5. Tony Waters, “The Relief Effort Winds Up,” 3.6. Nan Buzard, “Information-sharing and coordination between

NGOs working in the refugee camps of Ngara and Kibondo, Tanza-nia, 1994-1998,” unpublished manuscript prepared for the MIT-Mellon

117

Program on Forced Migration and NGOs, 1999: 6.7. John Borton, “An Account of Co-ordination Mechanisms for

Humanitarian Assistance During the International Response to the1994 Crisis in Rwanda,” Disasters 20, no. 4 (1996): 312.

8. Borton has noted that “World Vision distributed food andblankets to refugees who had just crossed the border en route to thecamps. This was contrary to the policy adopted by UNHCR and the 12agencies as it raised the possibility that refugees would settle near theborder away from the main camps.” Borton, “An Account of Co-ordination,” 322.

9. United Nations, “United Nations Consolidated Inter-AgencyAppeal for Humanitarian Assistance to Sierra Leone: March 1- De-cember 31, 1998,” June 1998, 4.

10. United Nations, “United Nations Consolidated Inter-AgencyFlash Appeal for Humanitarian Assistance to Sierra Leone: March 1-May 31, 1998,” March 1998: 10.

11. United Nations, “United Nations Consolidated Inter-AgencyAppeal,” 8.

12.Tony Waters, “The Relief Effort Winds Up,” 4.13. United Nations Field Security Handbook: System-Wide Arrange-

ments for the Protection of United Nations Personnel and Property in theField; Effective 1 January 1995 (New York: United Nations, 1995): 6.

14. A similar situation emerged in Albania during the Kosovarrefugee crisis in 1999, when NATO troops provided security to thosein refugee camps. See Larry Minear, Ted van Baarda, and MarcSommers, NATO and Humanitarian Action in the Kosovo Crisis Occa-sional Paper #36 (Providence, R.I.: Watson Institute, 2000).

15. G. Prunier, The Rwandan Crisis: History of a Genocide (NewYork: Columbia University Press, 1995): 13.

16. Tony Waters, “The Relief Effort Winds Up,” 7.17. For a discussion of the problems of insecurity in the refugee

camps and the Zairian Camp Security Contingent of Zairian troopsassembled by UNHCR, see Mark Frohardt, Diane Paul, and LarryMinear, Protecting Human Rights: The Challenge to Humanitarian Orga-nizations, Occasional Paper # 35 (Providence, R.I., Watson Institute,1999): 69.

Chapter 5

1. Mark Frohardt, Diane Paul, and Larry Minear, ProtectingHuman Rights: The Challenge to Humanitarian Organizations, OccasionalPaper # 35 (Providence, R.I., Watson Institute, 1999): 62.

2. See Elif Kaban, “Rwandan refugees shun Carter ‘go home’video,” Reuters North American Wire, May 15, 1996, which described

118

how some refugees “laughed when Carter told them: ‘I think I canassure the refugees that they will be well-treated when they go backto Rwanda.’“

3. For more information on the repatriation attempts, see JohanPotier, “Relief and Repatriation: Views by Rwandan Refugees; Les-sons for Humanitarian Aid Workers,” African Affairs 95, no. 380 (July1996): 403-429 and Marc Sommers, “Reconciliation and Religion:Refugee Churches in the Rwandan Camps,” Occasional Paper(Uppsala, Sweden: Life and Peace Institute, 1998).

4. “UNHCR notes increase in Rwandan refugee returns fromTanzania,” Agence France Presse, December 4, 1996.

5. Victor Antonie, “Tanzania warns Rwandan refugees of briefstay only,” Reuters World Service, January 20, 1996.

6. Hugh Nevill, “US, Europeans suggest new approaches torefugee problem,” Agence France Presse, April 3, 1996.

7. A UNHCR official involved in the Ngara operation, however,privately questioned the prevailing assessment. In retrospect, theofficial felt that if UNHCR had called the donors’ bluff, donors wouldhave continued to fund the Rwandan camps in Tanzania. Whether theTanzanians would have extended the Rwandans’ stay beyond 1996,however, appears much less likely.

8. Maureen Connelly, “Refugees or Hostages? Towards Under-standing the Crisis in the Great Lakes Region,” paper presented atOxford University, February 1, 1997.

9. “President Mkapa wants Rwandan refugees to go home,”Radio Tanzania External Service, in BBC Summary of World Broad-casts, Nov. 27, 1996.

10. “Tanzania welcomes return home of Rwandan refugees fromZaire,” Agence France Presse, November 26, 1996.

11. “Tanzania and U.N. Tell Rwandans to Go Home,” Reuters,New York Times, December 6, 1996, A15.

12. Joseph Kithama, “Refugees to be Sent Home,” AssociatedPress, The Dayton Daily News, December 6, 1996, 013A.

13. Kithama, “Refugees to be Sent Home.”14. Peter Smerdon, “Amnesty Slams UN over Rwandan Refugee

Repatriation,” Canadian Financial Report, December 7, 1996.15. Mark Dodd, “Rwandan refugees abandon six camps in Tan-

zania,” Reuters North American Wire, December 12, 1996.16. “Rwanda government delegation in Tanzania to discuss

refugee repatriation,” Radio Rwanda in French, December 13, 1996,reported in BBC Summary of World Broadcasts, 1115 gmt, December16, 1996.

17. “Refugees stay put, despite Tanzanian efforts to send themhome,” Deutsche Presse-Agentur, December 14, 1996.

18. David Fox, “Babies born on roadside amidst exodus from

119

Tanzania,” Reuters North American Wire, December 17, 1996.19. “Rwandan refugees crossing from Tanzania at 10,000 an

hour,” Agence France Presse, December 16, 1996.20. “Refugees stay put, despite Tanzanian efforts to send them

home,” Deutsche Presse-Agentur, December 14, 1996.21. James C. McKinley, Jr., “Tide of Rwanda Refugees Flows Back

to Tanzania Camps,” New York Times, December 14, 1996, Section 1, 6.22. Chris McGreal, “Tanzanian Keeps up Pressure on Hutu

Refugees,” The Guardian, December 14, 1996.23. Louise Tunbridge, “Aid teams kept away as Hutus file home,”

Daily Telegraph, December 17, 1996.24. Tunbridge, “Aid teams kept away.”25. David Fox, “Tanzania drives 200,000 refugees toward

Rwanda,” Canadian Financial Report, December 15, 1996.26. “East Africa: WFP Great Lakes Situation Report No. 38,” in

Africa News, December 16, 1996.27. Christophe Parayre, “Rwandan refugees crossing into Tanza-

nia at 10,000 an hour,” Agence France Presse, December 16, 1996.28. Tunbridge, “Aid teams kept away.”29. “Tanzania deports priests opposed to repatriation of refu-

gees,” Agence Frane Presse, December 17, 1996.30. Michela Wrong, “Army drives Hutu masses in Rwanda:

Refugee camps closed in act of forced repatriation,” Financial Times,December 16, 1996.

31. “Rwandan Refugees Return from Tanzania,” National PublicRadio (Morning Edition), December 16, 1996.

32. Human Rights Watch, “East Africa: Amnesty: Tanzania andUNHCR Must Respect International Law,” in Africa News, December17, 1996.

33. Human Rights Watch, “Tanzania: In the Name of Security:Forced Round-Ups of Refugees,” Human rights Watch Africa 11, no. 4(Tanzania), in a section titled “Background: From Offers of Citizen-ship to Increased Hostility” (July 1999): 2.

34. Human Rights Watch, “Tanzania: In the Name of Security.”35. Dennis McNamara, director, UNHCR Division of Interna-

tional Protection, to the House Committee on International Relations,Sub-Committee on International Operations and Human Rights; Hear-ing on “Rwanda: Genocide and the Continuing Cycle of Violence,”May 5, 1998, Washington, D.C.

36. Profile of Repatriation Activities, UNHCR Sub-Office, Ngara,Tanzania, March 15, 1996.

37. Johan Pottier, “Relief and Repatriation: Views by RwandanRefugees; Lessons for Humanitarian Aid Workers,” African Affairs 95(1996): 403-429, esp. 423.

38. Another example involved UNHCR’s attempt to activate the

120

“Tripartite Agreement on the Voluntary Repatriation of RwandeseRefugees from Tanzania,” signed by UNHCR and the Tanzanian andRwandan governments in April, 1995. The agreement stated that“voluntary repatriation, where feasible, constitutes the best durablesolution for the refugee problem” and contained pledges of supportfrom the two governments to guarantee refugee protection duringand after the return. Circulating the document in 1996, the UNHCRhead of sub-office invited NGO “comments and suggestions on whatmay be done” to speed the repatriation process “in light of the currentsituation in Rwanda.” (“Discussions/Consultations Concerning Vol-untary Repatriation,” UNHCR Memo NGA/HSO/404/96, October 2,1996, Ngara.)

39. Connelly, “Refugees or Hostages?”40. In reality, this was the Muslim refugees’ second attempt to

repatriate — they had tried in 1995 as well, but that effort, for reasonsthat are unclear (those present in Ngara at the time variously blamedUNHCR or the Rwandan government) never got off the ground.

41. For more information, Human Rights Watch, “Tanzania: Inthe Name of Security.”See also Marc Sommers, “A Child’s Nightmare:Burundian Children at Risk” (New York: Women’s Commission forRefugee Women and Children, 1998).

42. Ellen Lange, “The International Rescue Committee EducationProgramme for Refugees in Guinea, 1991-98: A Best Practice Study,”(Geneva: UNHCR/PTSS, 1998).

43. “It is striking to note that many requests by internationaldonors for educational assistance have been received from refugeegroups and populations undergoing crises. These requests have tendedto be ignored by large-scale donors. Instead, relief has been limited toimmediate survival needs, even though the provision of educationaland training programmes could be linked quite easily to relief pro-cesses.” Emily Vargas-Baron and Maureen McClure, “The New Hero-ics of Generational Commitment: Education in Nations with ChronicCrises,” in Gonzalo Retamal and Ruth Aedo-Richmond, eds., Educa-tion as a Humanitarian Response (London and Herndon, Va.: Cassell,1998).

44. If donors are reluctant to support emergency education forrefugees, they may be even more reluctant to support education forIDPs. This has certainly been the case in Sierra Leone, where thequality of education in IDP camp schools and across the country wasgenerally very low while in counterpart schools in refugee camps wascomparatively high.

45. A similar challenge awaited the Adventist Development andRelief Agency (ADRA), which was in charge of education for Liberianrefugees in the Ivory Coast.

46. See, for example, Pilar Aguilar and Gonzalo Retamal, “Rapid

121

Educational Response in Complex Emergencies: A discussion docu-ment” (Geneva: International Bureau of Education, 1998); Pilar Aguilarand Mark Richmond, “Emergency Educational Response in theRwandan Crisis,” in Gonzalo Retamal and Ruth Aedo-Richmond,eds., Education as a Humanitarian Response (London and Herndon, Va.:Cassel, 1998): 119-141; Jo Boyden and Paul Ryder, “The Provision ofEducation to Children Affected by Armed Conflict,” Oxford, UK(June 1996); and Graca Machel, “The Impact of Armed Conflict onChildren: Report of the Expert of the Secretary-General,” submittedpursuant to the resolution adopted by the General Assembly A/RES/48/157 to the Fifty-first Session of the United Nations General Assem-bly (Advance Copy), 1996. An assessment of Teacher EmergencyPackage’s (TEP) advantages and challenges can be found in MarcSommers, “Emergency Education for Children,” Working Paper,Center for International Studies, Massachusetts Institute of Technol-ogy, Mellon-MIT Inter-University Program on NGOs and ForcedMigration, Cambridge, MA, 1999.

47. The UNESCO-PEER office is based in Nairobi, Kenya. PEERstands for Programme for Education for Emergencies and Reconstruc-tion.

48. UNESCO-PEER’s TEP is an educational model that has beengenerically labeled a “school kit,” and is similar to UNICEF’s School-in-a-Box. School kits are actually only the best known component of alarger, more integrated educational approach, beginning with a recre-ational kit and followed by the TEP and then the reintroduction of theappropriate curriculum into schools. The most detailed description ofthis larger model is Pilar Aguilar and Gonzalo Retamal’s “RapidEducational Response in Complex Emergencies: A discussion docu-ment,” Geneva, International Bureau of Education, 1998.

49. For more information on this subject, see Sommers, “Emer-gency Education for Children.”

Chapter 6

1. Astri Suhrke, et al., The Kosovo refugee crisis: An independentevaluation of UNHCR’s emergency preparedness and response (Geneva:UNHCR [Pre-Publication Edition], February 2000): 73.

123

APPENDIX I

ACRONYMS

ACF Action Contre la FaimADRA Adventist Development and Relief AgencyAFRC Armed Forces Revolutionary Council

(Sierra Leone)BPRM Bureau of Population, Refugees

and Migration (U.S. State Department)CAP Consolidated Appeals ProcessCFA Committee on Food Aid (Sierra Leone)CRS Catholic Relief ServicesDHA Department of Humanitarian Affairs (OCHA)

[before January 1998]DRA Disaster Relief Agency (Tanzania)ECHO European Community Humanitarian OfficeECOMOG Economic Community of West African States

Monitoring GroupECOWAS Economic Community of West African StatesEPMU Education Program Management Unit

(Tanzania)ERD Emergency Response Division (UNDP)ERT Emergency Response Team (UNHCR)EU European UnionFAO Food and Agriculture Organization (UN)GDP Gross domestic productGNP Gross national productHACU Humanitarian Assistance Coordination Unit

(OCHA)HC Humanitarian Coordinator (UN)ICRC International Committee of the Red CrossIDPs Internally displaced personsIRC International Rescue CommitteeJEEAR Joint Evaluation of Emergency Assistance

to RwandaLWF Lutheran World FederationMOU Memorandum of understandingMSF Médecins sans Frontières

124

NCRRR National Commission for Reconstruction,Resettlement, and Rehabilitation(Sierra Leone)

NGO Nongovernmental organizationNPFL National Patriotic Front of LiberiaNTC National Technical Committee (Sierra Leone)OCHA Office for the Coordination of

Humanitarian Affairs (UN)OFDA Office of Foreign Disaster Assistance (USAID)PEER Programme for Education for Emergencies

andReconstruction (UNESCO)RC Resident Coordinator (UNDP)ResRep Resident Representative (UNDP)RPF Rwandan Patriotic FrontRTC Regional Technical Committee (Sierra Leone)RUF Revolutionary United Front (Sierra Leone)SMT Security Management Team (UN)SRSG Special Representative of the Secretary-GeneralTCRS Tanganyika Christian Refugee ServiceTEP Teacher Emergency Package (UNESCO-PEER)UAC Unaccompanied childrenUN United NationsUNDP United Nations Development ProgramUNHCHR United Nations High Commissioner for

Human RightsUNHCR United Nations High Commissioner for RefugeesUNICEF United Nations Children’s FundUNOMSIL United Nations Observer Mission in Sierra LeoneUSAID United States Agency for

International DevelopmentWAEC West African Exam CouncilWCRWC Women’s Commission for Refugee Women

and ChildrenWFP World Food ProgrammeWHO World Health Organization

125

APPENDIX II

CHRONOLOGY OF MAJOR EVENTS

Ngara/Rwanda

Oct. 1, 1990 Rwandan Patriotic Front troops enterByumba Prefecture, Rwanda, fromUganda, commencing a civil war.

Aug. 1, 1992 A cease fire and preliminary peace accordis signed between the Rwandan govern-ment and the RPF.

Aug. 4, 1993 After months of negotiations, the ArushaAccords are signed in Arusha, Tanzania,between the Rwandan government andthe RPF.

Oct. 21, 1993 Melchior Ndadaye, Burundi’s recentlyelected president, is assassinated. In theensuing ethnic violence, 50,000 Burundianpeople are killed, nearly all civilians.

December 1993 With estimates of a million or moreBurundians displaced from their homes,Burundi refugees pour into Rwanda andTanzania. The camps in Tanzania are re-mote and difficult to reach, the humani-tarian response is underfunded andpoorly organized. Estimates of the refu-gee population range from 245,000 to morethan 500,000.

February 1994 Facing disease and starvation, most of theBurundi refugees return to Burundi.About 40,000 remain in Tanzania. Mostare eventually transferred to Lukole Campin Ngara District.

126

April 6, 1994 The airplane carrying presidentsNtaryamira of Burundi and Habyarimanaof Rwanda is shot down near the Kigaliairport. Rwanda’s genocide begins laterin the day.

April 28, 1994 About 250,000 Rwandans cross the Rus-umo Bridge at the Tanzanian borderin 24 hours. They are placed at a sitecalled Benaco in Ngara District. TheRwandan refugee population will even-tually total more than a half million.

July 1994 Rémy Gatete, former KibungoBourgmestre, suspected genocidaire, andleader in the Rwandan refugee camps,organizes a demonstration after UNHCRasks him to report to its office in Benacorefugee camp. He is subsequently arrestedand removed from Ngara District.

April 1996 European Union Commissioner for Hu-manitarian Aid Emma Bonino and USAIDAdministrator Brian Atwood visit theRwandan refugee camps in Ngara Dis-trict and insist that the Rwandan's stay inTanzania “must be temporary.”

November 1996 About 600,000 Rwandan refugees returnto their country from Zaire in the wake ofclashes near the refugee camps involvingthe Zairian national army and formerRwandan government soldiers againstZairian insurgents led by Laurent Kabilaand the Rwandan national army.

Dec. 5, 1996 Tanzanian government and UNHCR of-ficials announce a plan to repatriate theRwandan refugees by year’s end. Theplan involves coordination with theRwandan government.

127

Dec. 12, 1996 More than 500,000 Rwandan refugees inNgara District abandon the refugee campsand head into the Tanzanian interior. TheTanzanian army acts swiftly to reversetheir direction and forcibly turns themtowards Rwanda. As the Tanzanian andRwandan governments had wished butin a fashion that marginalized the inter-national humanitarian community, nearlyall the refugees were repatriated toRwanda before the end of the month.

Sierra Leone

March 1991 The RUF enters Sierra Leone from Liberiawith Foday Sankoh at the helm, thus be-ginning the RUF’s insurgency against thegovernment.

April 1992 Sierra Leonean Army Captain ValentineStrasser overthrows President JosephSaidu Momoh in a military coup and as-sumes the presidency.

November 1994 The UN declares Sierra Leone a humani-tarian emergency months after some UNand international NGO agencies havebegun to shift their operations from de-velopment to emergency work.

January 1996 Strasser is overthrown by Brigadier JuliusMaada Bio in a military coup.

February 1996 Multi-party parliamentary and presiden-tial elections end four years of militaryrule.

March 1996 Ahmed Tejan Kabbah, a career UNDPofficial, becomes president in a runoffelection.

128

November 1996 Kabbah and Sankoh sign the first peaceagreement between the government andthe RUF.

May 1997 A military coup led by Major Johnny PaulKoroma and the Armed Forces Revolut-tionary Council forces Kabbah into exile.

June 1997 Koroma invites the RUF to join the juntagovernment in Freetown. Sankoh, underarrest in March 1997 in Nigeria on anarms charge, is made Koroma’s vice-presi-dent in absentia.

October 1997 A second peace treaty is signed betweenKabbah and the AFRC-RUF coalition.

February 1998 The Nigerian-led ECOMOG force chasesthe AFRC and RUF from Freetown, re-turning them to their forest hideaways.Kabbah is reinstated as president.

October 1998 Foday Sankoh is sentenced to death bythe High Court in Sierra Leone. Kabbahcalls for rebel forces to surrender andoffers amnesty.

January 1999 The RUF, with AFRC support, invadesFreetown a second time, leaving waves ofatrocities and mutilations in their wake,but ECOMOG eventually chases themfrom the capital.

July 1999 The third peace treaty between Kabbahand the RUF is negotiated, awarding theRUF key positions in the government.

May 1, 2000 Five hundred UN peacekeeper troops aretaken hostage by the RUF. Most are soonreleased, yet the peace treaty unravelsand insecurity plagues the countryside.

129

APPENDIX III

AGENCIES CONSULTED

In addition to independent regional and humanitarian ex-perts, refugees, and IDPs, officials from the following organi-zations were interviewed.

Action Contre la FaimCARECatholic Relief ServicesConcern UniversalEuropean Community Humanitarian OfficeEconomic Community of West African States

Monitoring GroupFeinstein International Famine Center, Tufts UniversityInternational Committee of the Red CrossInternational Federation of Red Cross

and Red Crescent SocietiesInternational Crisis GroupInternational Rescue CommitteeLutheran World FederationMerlinMédecins sans Frontières-BelgiumMédecins sans Frontières-HollandOxfam Great BritainProgramme for Education for

Emergencies and ReconstructionThe Sierra Leonean and Tanzanian governmentsTanganyika Christian Refugee ServiceUSAID Office of Foreign Disaster Assistance

and Office of Transition InitiativesUnited Nations Development ProgramUnited Nations Children's FundUnited Nations High Commissioner for RefugeesUnited Nations High Commissioner for Human RightsUnited Nations Office for the Coordination of Humanitarian

Affairs, including Humanitarian AssistanceCoordination Unit in Sierra Leone

United Nations Observer Mission in Sierra LeoneUnited Nations Security Office

130

U.S. State Department Bureau of Population, Refugeesand Migration

World Food ProgrammeWomen’s Commission for Refugee Women and ChildrenWorld BankWorld Vision

131

APPENDIX IV

ABOUT THE AUTHOR AND RESEARCH INSTITUTION

Marc Sommers has worked on a number of humanitarian,human rights and security issues, including urban refugees,conflict resolution, emergency and peace education, the im-pact of war on children and youths, the right to work andcivilian-military relations. A research fellow at BostonUniversity’s African Studies Center, he has served as a con-sultant for a wide range of agencies, including the Academyfor Educational Development, CARE, Conflict ManagementGroup, Oxfam America, UNESCO, USAID’s Office of ForeignDisaster Assistance, the U.S. Department of Defense and theWomen’s Commission for Refugee Women and Children. Dr.Sommers has carried out fieldwork in a number of African,Latin American, and European countries, including SierraLeone in 1997 and 1998 and Ngara District, Tanzania, in 1994,1996, and 1998. His research has been supported by the Ford,H.F. Guggenheim, Mellon and Rotary Foundations and theSocial Science Research Council. A forthcoming book, entitledFear in Bongoland: Burundi Refugees in Urban Tanzania, will bepublished by Berghahn Books in 2001.

The Humanitarianism & War Project is an independentpolicy research initiative underwritten by some 50 UN agen-cies, governments, NGOs, and foundations. Since its inceptionin 1991, it has conducted thousands of interviews on complexemergencies around the world, producing an array of casestudies, training materials, books, articles, and opinion piecesfor a diverse audience.

The project is currently examining the process of institu-tional learning and change among humanitarian organiza-tions in the post-Cold War period and highlighting innovativepractices devised by individual agencies to address specificchallenges. Current research builds on case studies, bothgeographical (the Persian Gulf, Central America and the Car-ibbean, Cambodia, the former Yugoslavia, the Great LakesRegion, and the Caucasus) and thematic (the interface be-tween humanitarian action and peacekeeping, and the roles ofthe media and the military in the humanitarian sphere). Re-

132

search is tailored to the expressed needs of humanitarianorganizations, the primary constituency of the project.

Intergovernmental organizations that have contributed tothe project are the European Community Humanitarian Office(ECHO), International Organization for Migration (IOM),OECD Development Centre, United Nations Disaster ReliefOrganization (UNDRO), United Nations DevelopmentProgramme (UNDP), United Nations Institute for Trainingand Research (UNITAR), DHA, OCHA, UNHCR, UNICEF,the UN Special Emergency Program for the Horn of Africa,UN Staff College, UN University, UN Volunteers, WFP, andWHO.

NGO contributors are the American Red Cross, CARE-US, Catholic Relief Services, Danish Refugee Council, Interna-tional Center for Human Rights and Democratic Development(Canada), International Federation of Red Cross and RedCrescent Societies, International Orthodox Christian Chari-ties, International Rescue Committee, Lutheran World Fed-eration, Lutheran World Relief, Mennonite Central Commit-tee (U.S.), Mennonite Central Committee (Canada), MercyCorps International, the Nordic Red Cross Societies, Norwe-gian Refugee Council, Oxfam-UK, Save the Children-UK,Save the Children-US, Trócaire, and World Vision-US.

Project donors also include the governments of Australia,Canada, France, the Netherlands, Sweden, the United King-dom, and the United States. Generous support has come aswell from the Arias Foundation, Ford Foundation, FourthFreedom Forum, Howard Gilman Foundation, MacArthurFoundation, McKnight Foundation, Andrew W. Mellon Foun-dation, Pew Charitable Trusts, John D. and Catherine T.Rockefeller Foundation, and the U.S. Institute of Peace.

The Project is an activity of Brown University’s WatsonInstitute for International Studies, which was established in1986 to facilitate analysis of global problems and to developinitiatives to address them. Additional information about theInstitute and the Project may be found at <http://www.brown.edu/Departments/Watson_Institute>. As ofSeptember 2000, the Project will be located at the FeinsteinInternational Famine Center at Tufts University’s School ofNutrition and Science Policy. More information will be avail-able from <http://www.hwproject.tufts.edu>.

OCCASIONAL PAPERS OF THE WATSON INSTITUTE

#1: North-South/East-West: Establishing a Common Agenda by Harry G. Barnes Jr., YuriA. Krasin, and Enriqué Iglesias#2: The United States, the Soviet Union and Security in the Third World by MichaelClough, Meryl A. Kessler, Robert Levgold, S. Neil MacFarlane, Augustus RichardNorton, Celeste A. Wallander, and Thomas G. Weiss#3: Human Sources of Global Change: Priority Research Initiatives for 1990-1995 by RobertW. Kates, William C. Clark, Vicki Norberg-Bohm, and B. L. Turner II#4: Superpowers and Regional Conflict in a Post-Cold War World: The Caribbean Basin andSouthern Africa by James G. Blight, Jorge I. Dominguez, Fen Osler Hampson, Jo L.Husbands, Gillian Gunn, Lloyd Searwar, Newell M. Stultz, Wayne S. Smith, andThomas G. Weiss#5: Toward Collective Security: Two Views by Sir Brian Urquhart and Robert S.McNamara#6: Brazil and Mexico: Contrasting Models of Media and Democratization by Ilya Adler,Elizabeth Mahan, Joseph Straubhaar, and Nelson H. Vieira; introduction by ThomasE. Skidmore#7: Working Together by Eduard A. Shevardnadze#8: Humanitarianism and War: Learning the Lessons from Recent Armed Conflicts byLarry Minear, Thomas G. Weiss, and Kurt M. Campbell#9: “This Will Not Be Another Vietnam:”George Bush and the Persian Gulf War byRichard A. Melanson#10: Collective Security in a Changing World by Lincoln P. Bloomfield, Tom J. Farer,Leon Gordenker, Ernst B. Haas, John Mackinlay, Oscar Schachter, James S. Sutterlin,and Thomas G. Weiss#11: Competing Gods: Religious Pluralism in Latin America by Gerard Béhague, David J.Hess, Marc Belanger, and Anani Dzidzienyo; introduction by Thomas E. Skidmore#12: Continuity and Change: Women at the Close of the Twentieth Century by ReginaCortina, Eleanor Doumato, Marida Hollos, Prema Kurien, and MarilynRueschemeyer#13: United Nations Coordination of the International Humanitarian Response to the GulfCrisis 1990-1992 by Larry Minear, U. B. P. Chelliah, Jeff Crisp, John Mackinlay, andThomas G. Weiss#14: Humanitarian Challenges in Central America: Learning the Lessons of Recent ArmedConflicts by Cristina Eguizábal, David Lewis, Larry Minear, Peter Sollis, and ThomasG. Weiss#15: United Nations Authority in Cambodia by Jarat Chopra#16: United Nations Financing Problems and the New Generation ofPeacekeeping and Peace Enforcement by Anthony McDermott#17: German Big Business and Europe in the Twentieth Century by Volker R. Berghahn,Reinhard Neebe, and Jeffrey J. Anderson#18: Humanitarian Action in the Former Yugoslavia: The UN’s Role, 1991-1993 by LarryMinear (team leader), Jeffrey Clark, Roberta Cohen, Dennis Gallagher, Iain Guest,and Thomas G. Weiss

#19: Mexico: The Artist is a Woman by Lucretia Giese, Carmen Boullosa, MarjorieAgosín, Sandra Berler, Elena Gascón-Vera, Laura Riesco, and Margo Glantz;edited by Regina Cortina#20: Humanitarian Action and Security in Liberia 1989-1994 by Colin Scott, incollaboration with Larry Minear and Thomas G. Weiss#21: Armed Conflict in Georgia: A Case Study in Humanitarian Action and Peacekeepingby S. Neil MacFarlane, Larry Minear, and Stephen Shenfield#22: The Policies of Mercy: UN Coordination in Afghanistan, Mozambique, and Rwandaby Antonio Donini#23: Haiti Held Hostage: International Responses to the Quest for Nationhood 1986-1996by Robert Maguire (team leader), Edwige Balutansky, Jacques Fomerand, LarryMinear, William G. O’Neill, Thomas G. Weiss, and Sarah Zaidi#24: Imagining a Free Cuba: Carlos Manuel de Céspedes and José Martí by Cathy L.Jrade, George Monteiro, Nelson R. Orringer, Louis A. Pérez, Jr., Ivan A. Schulman,Thomas E. Skidmore, and Wayne S. Smith; edited by José Amor y Vázquez#25: Humanitarian Action and Politics: The Case of Nagorno-Karabakh by S. NeilMacFarlane and Larry Minear#26: War and Humanitarian Action in Chechnya by Greg Hansen and Robert Seely#27: National Self-Determination: Approaches and Case Studies by Galina Starovoitova#28: An Overview and Assessment of 1989-1996 Peace Operations Publications byCindy Collins and Thomas G. Weiss#29: Humanitarian Impacts of Economic Sanctions in Burundi by Eric Hoskins andSamantha Nutt#30: Integration and Disintegration in the Former Soviet Union: Implications forRegional and Global Security by P. Terrence Hopmann, Stephen D. Shenfield, andDominique Arel#31: Toward More Humane and Effective Sanctions Management: Enhancing theCapacity of the United Nations System by Larry Minear, David Cortright, JuliaWagler, George A. Lopez, and Thomas G. Weiss#32: Humanitarian Action in the Caucasus: A Guide for Practitioners by Greg Hansen#33: Relief and Development: The Struggle for Synergy by Ian Smillie#34: A Humanitarian Practitioner's Guide to International Human Rights Law byWilliam G. O'Neill#35: Protecting Human Rights: The Challenge To Humanitarian Organizations byMark Frohardt, Diane Paul, and Larry Minear#36: NATO and Humanitarian Action in the Kosovo Crisis by Larry Minear, Ted vanBaarda, and Marc Sommers#37: Humanitarian Action: Social Science Connections by Stephen C. Lubkemann,Larry Minear, and Thomas G. Weiss#38: When Needs Are Rights: An Overview of UN Efforts to Integrate Human Rights inHumanitarian Action by Karen Kenny#39: Humanitarian Action: A Transatlantic Agenda for Operations and Research byLarry Minear and Thomas G. Weiss#40: The Dynamics of Coordination by Marc Sommers

WIIS • 2 Stimson Ave. • Brown University • Providence, RI 02912

TH

E DY

NA

MIC

S OF C

OO

RD

INA

TIO

NO

ccasional Paper #40

THOMAS J. WATSON JR. INSTITUTE FOR INTERNATIONAL STUDIES

Occasional Paper #40

THE DYNAMICS OF COORDINATION

MARC SOMMERS