Commissioned and published by the Humanitarian Practice Network at ODI

Number 51June 2005

Network Paper

About HPNThe Humanitarian Practice Network at theOverseas Development Institute is an independent forum where field workers, managers and policymakers in the humanitariansector share information, analysis and experience.The views and opinions expressed in HPN’s publications do not necessarily state or reflectthose of the Humanitarian Policy Group or theOverseas Development Institute.

Overseas Development Institute111 Westminster Bridge RoadLondon SE1 7JDUnited Kingdom

Tel. +44 (0) 20 7922 0300Fax. +44 (0) 20 7922 0399

HPN e-mail: hpn@odi.org.ukHPN website: www.odihpn.org

Britain’s leading independent think-tank on international developmentand humanitarian issues

HPNHumanitarian Practice Network

Managed by

Humanitarian Policy Group

Max P. Glaser

Humanitarian engagement with non-state armed actorsThe parameters of negotiated access

In brief

•This paper addresses the question ofhumanitarian engagement with the non-state armed groups that increasingly populate the zones in which humanitarianaction takes place. In particular, it seeks to understand why some combatants react positively and consistently to humanitarian demands to meet access preconditions, while others respond erratically, decline to respond or are hostile.

•The paper looks less at how to negotiate with such groups, and more at the varioustypes of non-state armed group with whichnegotiations are likely to be conducted.

•The ultimate objective of this paper is to determine the parameters of responsiblehumanitarian engagement – that is, to investigate the scope of successful engagement, one which maintains minimaloperational preconditions, such as security for aid workers, and to explore the availablemodalities of engagement.

NP 51 cover 2nd 16/6/05 10:21 am Page 3

Humanitarian Practice Network (HPN)Overseas Development Institute111 Westminster Bridge RoadLondon, SE1 7JDUnited Kingdom

Tel: +44 (0)20 7922 0331/74Fax: +44 (0)20 7922 0399 Email: hpn@odi.org.ukWebsite: www.odihpn.org

Printed and bound in the UK

ISBN: 0 85003 760 3Price per copy: £4.00 (excluding postage and packing).© Overseas Development Institute, London, 2005.

Photocopies of all or part of this publication may be made providing that the source is acknowledged. Requestsfor the commercial reproduction of HPN material should be directed to the ODI as copyright holders. The NetworkCoordinator would appreciate receiving details of the use of any of this material in training, research or pro-gramme design, implementation or evaluation.

Acknowledgements

This Network Paper is based on the research paper Negotiated Access – Humanitarian Engagement withArmed Non-State Actors, published in May 2004 by the Carr Center for Human Rights Policy, Kennedy Schoolof Government, Harvard University.

About the author

Max Glaser is an independent consultant for humanitarian operations. He can be contacted at: mglaser@pax-consultancy.com.

NP 51 cover 2nd 16/6/05 10:21 am Page 4

Humanitarian Practice Network

The Humanitarian Practice Network (HPN) is an independent forum where field workers, managersand policymakers in the humanitarian sector share information, analysis and experience.

HPN’s aim is to improve the performance of humanitarian action by contributing to individual and institutional learning.

HPN’s activities include:

• A series of specialist publications: Good Practice Reviews, Network Papers and HumanitarianExchange magazine.

• A resource website at www.odihpn.org.• Occasional seminars and workshops to bring together practitioners, policymakers and analysts.

HPN’s members and audience comprise individuals and organisations engaged in humanitarianaction. They are in 80 countries worldwide, working in northern and southern NGOs, the UN andother multilateral agencies, governments and donors, academic institutions and consultancies.HPN’s publications are written by a similarly wide range of contributors.

HPN’s institutional location is the Humanitarian Policy Group (HPG) at the Overseas DevelopmentInstitute (ODI), an independent think tank on humanitarian and development policy. HPN’s publicationsare researched and written by a wide range of individuals and organisations, and are published by HPNin order to encourage and facilitate knowledge-sharing within the sector. The views and opinionsexpressed in HPN’s publications do not necessarily state or reflect those of the Humanitarian PolicyGroup or the Overseas Development Institute.

Funding support is provided by institutional donors (AusAID, CIDA, DANIDA, DFID, DevelopmentCooperation Ireland, MFA Netherlands, SIDA, USAID), non-governmental organisations (British RedCross, CAFOD, Christian Aid, Concern, International Rescue Committee, MSF, Oxfam, Save the Children(UK), World Vision) and UN agencies (WFP).

To join HPN, complete and submit the form at www.odihpn.org or contact the Membership Administrator at:

Humanitarian Practice Network (HPN)Overseas Development Institute

111 Westminster Bridge RoadLondon, SE1 7JDUnited Kingdom

Tel: +44 (0)20 7922 0331/74Fax: +44 (0)20 7922 0399

Email: hpn@odi.org.ukWebsite: www.odihpn.org

© Overseas Development Institute, London, 2005.

NP 51 cover 2nd 16/6/05 10:21 am Page 2

110 Room for Improvement: the Management and Support ofRelief Workers by R. Macnair (1995)

11 Cash-for-Work and Food Insecurity in Koisha, SouthernEthiopia by P. Jenden (1995)

12 Dilemmas of ‘Post’-Conflict Transition: Lessons from theHealth Sector by J. Macrae (1995)

13 Getting On-Line in Emergencies: A Guide and Directory to theInternet for Agencies involved in Relief and Rehabilitation byL. Aris, P. Gee and M. Perkins (1996)

14 The Impact of War and Atrocity on Civilian Populations: BasicPrinciples for NGO Interventions and a Critique ofPsychosocial Trauma Projects by D. Summerfield (1996)

15 Cost-effectiveness Analysis: A Useful Tool for the Assessmentand Evaluation of Relief Operations? by A. Hallam (1996)

16 The Joint Evaluation of Emergency Assistance to Rwanda:Study III ed. J. Borton (1996)

17 Monetisation: Linkages to Food Security? by J. Cekan, A.MacNeil and S. Loegering (1996)

18 Beyond Working in Conflict: Understanding Conflict andBuilding Peace (The CODEP Workshop Report), by J. Bennett and M. Kayitesi Blewitt (1996)

19 Human Rights and International Legal Standards: what reliefworkers need to know by J. Darcy (1997)

20 People in Aid Code of Best Practice in the Management andSupport of Aid Personnel ed. S. Davidson (1997)

21 Humanitarian Principles: The Southern Sudan Experience byI. Levine (1997)

22 The War Economy in Liberia: A Political Analysis by P.Atkinson (1997)

23 The Coordination of Humanitarian Action: the case of SriLanka by K. Van Brabant (1997)

24 Reproductive Health for Displaced Populations by C.Palmer (1998)

25 Humanitarian Action in Protracted Crises: the new relief‘agenda’ and its limits by D. Hendrickson (1998)

26 The Food Economy Approach: a framework for under-standing rural livelihoods by T. Boudreau (1998)

27 Between Relief and Development: targeting food aid fordisaster prevention in Ethiopia by K. Sharp (1998)

28 North Korea: The Politics of Food Aid by J. Bennett (1999)29 Participatory Review in Chronic Instability: The Experience

of the IKAFE Refugee Settlement Programme, Uganda byK. Neefjes (1999)

30 Protection in Practice: Field Level Strategies for ProtectingCivilians from Deliberate Harm by D. Paul (1999)

31 The Impact of Economic Sanctions on Health and Well-being by R. Garfield (1999)

32 Humanitarian Mine Action: The First Decade of a NewSector in Humanitarian Aid by C. Horwood (2000)

33 The Political Economy of War: What Relief Agencies Needto Know by P. Le Billon (2000)

34 NGO Responses to Hurricane Mitch: Evaluations forAccountability and Learning by F. Grunewald, V. deGeoffroy & S. Lister (2000)

35 Cash Transfers in Emergencies: Evaluating Benefits andAssessing Risks by D. Peppiatt, J. Mitchell and P. Holzmann (2001)

36 Food-security Assessments in Emergencies: A LivelihoodsApproach by H. Young, S. Jaspars, R. Brown, J. Frize andH. Khogali (2001)

37 A Bridge Too Far: Aid Agencies and the Military in Humanitarian Response by J. Barry with A. Jefferys (2002)

38 HIV/AIDS and Emergencies: Analysis and Recommend-ations for Practice by A. Smith (2002)

39 Reconsidering the tools of war: small arms and humanitarian action by R. Muggah with M. Griffiths(2002)

40 Drought, Livestock and Livelihoods: Lessons from the1999-2001 Emergency Response in the Pastoral Sector inKenya by Yacob Aklilu and Mike Wekesa (2002)

41 Politically Informed Humanitarian Programming: Using aPolitical Economy Approach by Sarah Collinson (2002)

42 The Role of Education in Protecting Children in Conflict bySusan Nicolai and Carl Triplehorn (2003)

43 Housing Reconstruction after Conflict and Disaster bySultan Barakat (2003)

44 Livelihoods and Protection: Displacement and VulnerableCommunities in Kismaayo, Southern Somalia by SimonNarbeth and Calum McLean (2003)

45 Reproductive Health for Conflict-affected People: Policies,Research and Programmes by Therese McGinn et al. (2004)

46 Humanitarian futures: practical policy perspectives byRandolph Kent (2004)

47 Missing the point: an analysis of food security interven-tions in the Great Lakes by S Levine and C Chastre with SNtububa, J MacAskill, S LeJeune, Y Guluma, J Acidri and AKirkwood

48 Community-based therapeutic care: a new paradigm forselective feeding in nutritional crises by Steve Collins

49 Disaster preparedness programmes in India: a cost bene-fit analysis by Courtenay Cabot Venton and Paul Venton(2004)

50 Cash relief in a contested area: lessons from Somalia byDegan Ali, Fanta Toure, Tilleke Kiewied (2005)

Network Papers 1995–2004Network Papers are contributions on specific experiences or issues prepared either by HPN members

or contributing specialists.

Good Practice ReviewsGood Practice Reviews are major, peer-reviewed contributions to humanitarian practice. They are produced periodically.

1 Water and Sanitation in Emergencies by A. Chalinder (1994)2 Emergency Supplementary Feeding Programmes by J.

Shoham (1994)3 General Food Distribution in Emergencies: from Nutritional

Needs to Political Priorities by S. Jaspars and H. Young (1996)4 Seed Provision During and After Emergencies by the ODI

Seeds and Biodiversity Programme (1996)5 Counting and Identification of Beneficiary Populations in

Emergency Operations: Registration and its Alternatives byJ. Telford (1997)

6 Temporary Human Settlement Planning for DisplacedPopulations in Emergencies by A. Chalinder (1998)

7 The Evaluation of Humanitarian Assistance Programmes inComplex Emergencies by A. Hallam (1998)

8 Operational Security Management in ViolentEnvironments by K. Van Brabant (2000)

9 Disaster Risk Reduction: Mitigation and Preparedness inDevelopment and Emergency Programming by JohnTwigg (2004)

A full list of HPN publications is available at the HPN website: www.odihpn.org. To order HPN publications,contact hpn@odi.org.uk.

NP 51 cover 2nd 16/6/05 10:21 am Page 5

i

Chapter 1 Introduction 1

Aims and scope 2

Chapter 2 The changing conflict context and the rise of new non-state armed groups 5

Spoilers 5

The ‘learning belligerent’ and the civilian connection 5

Trinitarian and non-trinitarian warfare 6

Chapter 3 Characteristics and classifications of ANSAs 7

Command structure 7

Independence from state control 7

The use of violence for political purposes 8

Excluded categories 8

Effective control over population and territory 8

Classifying ANSAs in economic terms 9

The need for fighters 9

External relations: recognition and legitimacy 10

The civilian position: in- and out-group membership 10

Classifying ANSAs in political terms 11

Chapter 4 Engaging ANSAs: effectiveness and reliability 13

Objectives and tactical position 13

Organisational structure and internal discipline 14

Analysing the risks and benefits of humanitarian engagement 14

Mitigating contextual factors: security strategies and promoting IHL 16

Chapter 5 Modalities, levels and interlocutors 19

Direct or indirect engagement? 19

High or low? 19

Open or confidential? 19

The choice of interlocutor 20

Phases of engagement 20

Chapter 6 Conclusions and recommendations 23

Unwarranted humanitarian engagement 23

Recommendations 23

Best practice guidelines 23

Contents

NP 51 2nd 16/6/05 10:24 am Page i

Humanitarian engagement with non-state armed actors: the parameters of negotiated access

Annex 1 Recommended reading 25

Annex 2 Bibliography 25

List of boxes

Box 1: Negotiated access: principled process or accommodation of interests? 3

Box 2: Trinitarian and non-trinitarian warfare 6

Box 3: Persuasion: promoting IHL to change combatants’ behaviour 16

List of figures and tables

Figure 1: The function of civilians in conflict 3

Figure 2: Trinitarian warfare 6

Figure 3: Non-trinitarian warfare 6

Table 1: ANSA classification 11

Table 2: Willingness and capability 13

Table 3: Objectives and structure 13

Table 4: Risk–benefit analysis according to the classification of ANSAs 15

Table 5: General risks–benefits from humanitarian engagement 15

Figure 4: Choice of interlocutor: direct and indirect contacts 20

ii

NP 51 2nd 16/6/05 10:24 am Page ii

Over the last several decades, non-state armed groups have become acommon feature of civil conflicts andinternal wars. These Armed Non-State Actors (ANSAs) go by manynames, including liberation move-ments, rebel groups, paramilitaries,insurgents and warlords, mercena-ries and private military and securitycompanies. The category could nowalso include transnational terroristorganisations such as Al-Qaeda. Theproliferation of armed non-stategroups mirrors the proliferation ofinternal conflicts across the globe.International wars such as the US-led invasions of Afghanistan andIraq are now very much theexception: virtually all of today’sconflicts are internal to states, notinternational.

The proliferation of ANSAs hascomplicated humanitarian accessbecause it has contributed to adeterioration in the security conditions for aid workers inconflict zones. These armed groups may also compromisethe impartiality of aid and the status of civilians by co-optingthem for logistical or political support, blurring thedistinction between combatant and non-combatant.Civilians may be recruited as fighters, whether voluntarily orthrough coercion, and civilian environments may be used toprovide tactical cover. Humanitarian assistance may beblocked if its delivery is deemed not in the armed group’sinterest. The provisions of International Humanitarian Law(IHL) may be breached, and humanitarian accesschallenged, as a deliberate strategy of war.

The difficulties ANSAs raise for humanitarian agenciesmean that negotiating with them for access to vulnerablepopulations has become an inevitable – and very complex– part of the humanitarian experience in many contexts.The overall objective of humanitarian aid is to provide life-saving assistance and alleviate suffering. In achieving thisobjective, safe access to vulnerable and needy people inwar zones is essential. Hence, the direct aim ofhumanitarian engagement with ANSAs is to secure twopreconditions as minimal operational criteria. The first isensuring security guarantees for aid workers. The secondis to secure the ANSA’s respect for the rules of IHL. Withrespect to the latter, recognising the special status ofcivilians and their right of access to impartial humanitarianassistance is crucial. Over the past decade, a third, morecontentious, aspect of engagement has emerged, namelythe need to protect civilians as such.1 Negotiating access

with ANSAs suspected of breaching human rights orcommitting crimes against humanity has raised questionsas to whether access agreements with such groups serveor undermine the protection of civilians, as agreementsmay accord undue legitimacy to these ANSAs. They mayalso jeopardise any ‘common front’ intended to isolateabusive ANSAs, as these groups may play onehumanitarian actor off against another. Here, the issue of‘responsible engagement’ comes into focus, meaningaccess negotiations promoting ‘humanitarian space’ (ageneral respect for IHL principles), as opposed to accessnegotiations that create ‘agency space’, meaningagreements that pertain only to some humanitarian actorsbut not others, or which fail to establish general securityguarantees, or to prevent continued breaches of IHL.

To be sure, ANSAs are not the only forces abusing civilians orbreaching IHL; ‘regular’ armies violate these rules as well.However, this paper concentrates on ANSAs because of theirprevalence, their tremendous variety, the very differentconflict contexts in which they operate and the diverseaccess difficulties they present. The paper also focuses on aspecific subset of the humanitarian enterprise, namely non-governmental humanitarian agencies, referred to here asNGHAs, meaning those non-governmental agenciesinvolved in the provision of emergency life-saving assistancein the context of conflict and war.2This focus reflects the factthat, as the international community has progressivelywithdrawn from so-called non-strategic states (Rwanda,Somalia or the Democratic Republic of Congo, for instance),

1

Chapter 1

Introduction

A former Janjaweed fighter guards a humanitarian NGO vehicle, West Darfur, August 2004

©M

ax P. Glaser

NP 51 2nd 16/6/05 10:24 am Page 1

Humanitarian engagement with non-state armed actors: the parameters of negotiated access

NGHAs are, more than ever before, being forced to engagewith ANSAs to negotiate access conditions on their own.At the other extreme, mainly as a result of the ‘War onTerror’, conflicts such as those in Afghanistan and Iraq aresaturated with third states’ strategic interests, politicisingthe context and compromising humanitarian accessconditions. This dual international response has meantthat the negotiation of access progressively takes placebetween non-state actors: NGHAs on the one hand, andANSAs on the other. Although other agencies involved inhumanitarian action, such as UN bodies and the ICRC, alsoengage with ANSAs, these organisations are moreformally configured within the framework of internationalconventions and diplomacy. This does not, in itself, alterthe challenge of humanitarian engagement that theseagencies face. For NGHAs, however, it underlines theimportance of enhancing knowledge of the intricacies ofhumanitarian engagement and access negotiation withANSAs, as they increasingly deal with these groupsbilaterally.

Aims and scope

Aid workers in conflict zones encounter armed men on adaily basis, at roadblocks for example. While dealing witharmed actors certainly calls for negotiation skills, thispaper seeks to address some of the broader aspects ofhumanitarian engagement with non-state groups. Inparticular, it seeks to understand why some combatantsreact positively and consistently in response tohumanitarian demands to meet access preconditions,while others respond erratically, decline to respond or arehostile. This problem in turn raises other immediatequestions. What level of interlocutor – the leadership, amid-level commander, or a field combatant – is likely toproduce the most effective result? Should one seek directcontact, or work through a civilian go-between? Is it betterto conduct negotiations openly, or should they beconfidential?

Questions like these underscore the importance of asystematic approach to assessing the likelihood that aparticular armed non-state group will respect the termsof a negotiated access agreement. In other words, howreliable is the access agreement likely to be? To addressthese questions, this paper looks less at how tonegotiate, and more at the various types of non-statearmed group with which this negotiation is likely to beconducted, with the proviso that the former is aprerequisite to the latter. The ultimate objective of thispaper is to determine the parameters of responsiblehumanitarian engagement with ANSAs – that is, toinvestigate the scope of successful engagement, onewhich maintains minimal operational preconditions, suchas security for aid workers, and to explore the availablemodalities of engagement with ANSAs.

The objective of humanitarian engagement with ANSAs isto ‘negotiate’ the minimal preconditions necessary foraccess. These are security of aid workers, and respect for

the principles and rules of IHL. The term ‘negotiate’ isdeliberately placed in quotation marks here, since what isnegotiated, and with whom, is as unclear as how tonegotiate. Arguably, even the term ‘negotiation’ isinadequate, since the principles of IHL as conceived inthe Geneva Conventions are by definition non-negotiable.Agreements on access between ANSAs and humanitarianagencies do not express compromise on a contestedissue, but rather the balancing of pragmatic interests.Agreement on access expresses the coincidence of eachside’s internal goals; it does not necessarily reflectagreement on shared principles of IHL. In this sense,negotiation is not so much principle-driven, but rather adynamic bartering process formed and fed by itsparticipants: ANSAs and NGHAs. In practical operationalterms, negotiation can be deemed successful if itchanges the behaviour of the non-state armed actor in itstreatment of civilians and its respect for aid workers.Thus, rather than judging humanitarian engagement andaccess negotiations on legal, moral or ethical grounds,this paper analyses these issues pragmatically, in termsof the underlying interests of the humanitarianorganisation and the non-state actor.

In order to understand the dynamics of negotiated access,three dimensions are important:

• the specific relationship between the ANSA andcivilians;

• the ANSA’s internal organisation and commandstructure; and

• the ANSA’s external relations and supporting actors.

Of these three aspects, the first is of primary importance,since civilians serve as the link between the humanitarianNGO and the ANSA: access to civilians is the motive forseeking engagement with an armed non-state group.Civilians are not simply inactive ‘bystanders’ in conflict:they can fulfil diverse, crucial roles, as fighters, politicalsupporters, labourers, messengers or proxy targets.Often, these roles are not picked by choice, but rather area consequence of the manner in which civilians areimplicated and configured in conflict by warring parties,be they ANSAs, their opponents, competing non-stategroups or a state’s military forces. Civilians may activelyside with an ANSA to extract advantage, or they may becoerced into doing so. The precise outcome of accessnegotiations (the balance of interests that is reached)depends mainly on the attitude of the ANSA towardscivilians (supportive or indifferent?); the degree to whichthe ANSA is dependent on civilians (supportiveconstituency or opposed community?); and the ANSA’smode of control over civilians (protective or oppressive?).These relationships appear to have a strong influence onANSA attitudes towards humanitarian access andpresence, depending on how this presence dovetails withtheir interests.

One of the first steps in assessing the reliability of aparticular armed group is understanding the context of

2

NP 51 2nd 16/6/05 10:24 am Page 2

negotiation, and the distinctive characteristics of thegroup or groups in question, their underlying interestsand the nature of their relationship with the civiliansunder their control. To that end, this paper offers twotypologies by which ANSAs can be grouped. The first,based on a socio-economic model of ANSA–civilianrelations, distinguishes four types of relationship:symbiotic, parasitic, independent and predatory.4 Thistypology describes specific ‘terms of exchange’ betweenANSAs and civilians, in particular as they relate to so-called war economies, in which, for example, civiliansreceive protection from an armed group in exchange forservices or support, or are compelled to provide theseservices under threat of abuse. The second model uses apolitical, rather than economic, classification. It identifiesfour categories of ANSA–civilian relations: protective,competitive, antagonistic and sectarian. Both typologiescan be used to identify and determine the risks andbenefits of humanitarian engagement with ANSAs inmore objective terms.

The approach outlined in this paper is of course open todebate. It is hoped that it will be seen and used as a way togauge and assess how NGHAs approach non-state armedgroups in the service of their humanitarian work. Perhapsthe best point of departure is to understand that not allANSAs are the same: different groups pose differentchallenges and operate in different ways. This counselscaution in regarding ANSAs as somehow monolithic. Themethodology offered here should therefore not be seen asa ‘magic bullet’, solving all the problems, but rather itshould provide NGHAs with the flexibility they need toapproach different situations and different groups.Ultimately, the methodology should encourage analysis ofthe civil relations that underlie the process of negotiatingaccess, and a move away from the view that the problem ofnon-state armed groups is solely a military or securitymatter.

Chapter 1 Introduction

3

Box 1

Negotiated access: principled process oraccommodation of interests?

The principles of IHL as stated in the Geneva Conventionsare non-negotiable. What then is ‘negotiated’ whenengaging with an ANSA for humanitarian access? What canbe bartered or traded with a non-state armed group inexchange for its granting access to a humanitarian agency?

OLS: the original model of ‘negotiated access’

Operation Lifeline Sudan was initiated in 1989 as atripartite agreement between the UN, the Sudanesegovernment and the Southern opposition rebel groupthe Sudan People’s Liberation Army (SPLA). It is seen asthe prototype model of ‘negotiated access’. By signingthe agreement, the Sudanese government temporarilyceded sovereignty of Southern Sudan to the UN, leavingthe latter with the problem of administering aid. Theagreement was supported by so-called ‘ground rules’intended to provide security guarantees to aid workers,as well as protection to civilians. In practice, however,the ground rules were used more as a tool to providesafe access than as a means of holding rebel authoritiesaccountable.3 One of the major reasons for this was thecontested legitimacy of the Southern rebels in the eyesof the government in Khartoum. Neither the UN nor theNGHAs in Southern Sudan had any choice but to workwith the rebel authorities, even if they were notrecognised. The OLS agreement could not ‘solve’ theissue of legitimacy, and was in fact designed in such away that it circumvented it to facilitate humanitarianaccess and aid. This had the effect of leaving a gap inaccountability mechanisms, in particular in respect ofthe Southern rebels.

Figure 1

The function of civilians in conflict

For NGHA

Victims of war and conflict

Beneficiaries/recipients of aid

Logistics, labour force, safe harbour

‘Enemy’ (genocidal and ethnic conflict)

Target (support basis of opponents)

For ANSA

(Political) support basis

Fighters (combatants)

CIVILIANS

NP 51 2nd 16/6/05 10:24 am Page 3

4

NP 51 2nd 16/6/05 10:24 am Page 4

Over the last decade and a half, the nature of conflict haschanged dramatically. The Cold War was driven largely bythe ideological conflict between the superpowers, playedout in proxy conflicts in the developing world. In countriessuch as Ethiopia, Angola or Mozambique, ANSAs involvedin these conflicts portrayed themselves as ‘liberationmovements’, fighting a political fight for nationalliberation. Since the end of the Cold War, these ideologicalconflicts have given way to wars fought more explicitlyover resources. In countries such as Angola and SierraLeone, conflict became less a means to an end than an endin itself; ‘exploiting the fruits of insecurity and chaos to thebenefit of armed factions and militia’.5 In these conflicts,non-state armed groups like Charles Taylor’s NationalPatriotic Front of Liberia (NPFL) or the Revolutionary UnitedFront (RUF) in Sierra Leone have abused relief aid throughlooting and manipulation, and utilised it as a conflictresource.

Meanwhile, ethnically-driven conflicts in Rwanda and theBalkans, and the ongoing war in Ituri, DRC, have created afresh set of problems for the humanitarian enterprise: howto avoid partisan aid in a conflict where (mainly) one sideis being victimised. The standard humanitarian response of‘proportional assistance to both sides’ does not work, asby default it tends to serve the interests of the mostpowerful (and ruthless) party to the conflict.6 Finally, thepost-9/11 ‘War on Terror’ has seen the emergence of newstrains of transnational non-state actor, exemplified by Al-Qaeda and its affiliates, with their strong anti-Westernphilosophy and nihilistic methods. More recently,international conflict under the ‘War on Terror’ rubric hasspawned new non-state groups seeking to exploitinstability and oppose Western occupation. These groupsdemonstrate little concern for civilians or aid workers, asboth are considered proxy targets in the achievement oftheir goals.

Spoilers

The concept of ‘spoiler forces’ has emerged to describegroups that profit from instability or lawlessness, and whichaccordingly seek to prevent or ‘spoil’ the establishment ofpeace. The concept distinguishes between ‘limitedspoilers’, ‘greedy spoilers’ and ‘total spoilers’.7 As the labelsuggests, limited spoilers may be relatively easy to containas their demands are parochial, and limited to localconcerns. Greedy spoilers may be appeased once specificgrievances, usually economic, are addressed. Totalspoilers, however, pose severe challenges to peace. Al-Qaeda is perhaps the extreme example of a group thatcould also include the Islamic insurgents fighting Westernoccupying forces and their local allies in Iraq. Such groups

also pose serious challenges to humanitarian access inconflict zones; to illustrate a point already powerfully madeby the bomb attacks against the UN and the ICRC inBaghdad in 2003, one Iraqi insurgent group adopted thename Mujahideen Sans Frontières in a chillingly clearreference to the humanitarian agency Médecins SansFrontières. In Afghanistan, so-called ‘neo-Taliban’ forceshave attacked Afghans perceived as supporting the new,Western-backed regime, as well as foreign construction andaid workers. In the most notorious incidents, an ICRCdelegate was killed in late 2003, and five MSF aid workerswere murdered in May 2004.

The ‘learning belligerent’ and the civilianconnection

Through their engagement with aid agencies, ANSAs haveprogressively gained experience of, and insights into, thedynamics of humanitarian engagement in conflict, andhave adapted their approaches and tactics accordingly. Forexample, in October 1996, at the height of the Congoleserebel offensive in Kivu, the Rwandan military preventedhumanitarian agencies from reaching the refugee camps inGoma. The Rwandan leadership maintained that their ownexperience in ‘bush war’ taught them that NGOs’ logisticsand means of communication could be used to theadvantage of their opponents; denial of access waspredicated on a desire to deny that advantage to theInterahamwe and the former Rwandan army.8

The ‘learning belligerent’ makes choices on the basis ofrational considerations and pragmatic interests. It isimportant that NGHAs recognise that this adaptive processis taking place, and apply their own reciprocal, adaptivelearning, in particular for operational staff responsible forinitiating engagement with ANSAs. For NGHAs, this learningincludes acknowledging the difference between negotiationseen as the establishment of ‘humanitarian space’, andnegotiation to achieve ‘agency space’. Where the latter mayresult in the acceptance of one or more (specific) NGHAs byan ANSA to provide humanitarian assistance, the former ismuch broader, and aims at generating blanket respect forIHL by ANSAs, including the security of aid workers and theprotected status of civilian hors de combat. The eagernessof some NGHAs to gain access to civilians can easily beexploited by ANSA leaders to play one agency off againstanother, as the RUF did in Sierra Leone, for instance.9 Thisissue will be taken up again in the following chapters.

‘Control over civilians’ is a central aspect in the process ofnegotiating access with ANSAs. The nature of theANSA–civilian relationship is of crucial importance, since itinfluences the degree to which the ANSA will be willing to

5

Chapter 2The changing conflict context and the rise of new non-state armed groups

NP 51 2nd 16/6/05 10:24 am Page 5

respect access preconditions, and the likelihood that theaccess agreement will hold. In internal conflicts, control andaccess to civilians (or the denial of these things toopponents) is decisive for an ANSA’s functions andobjectives, regardless of whether they are aiming at awinning strategy, or obstructing the victory of theiropponents. Civilian populations are caught in conflict, assurrogate political agents, as (part-time) fighters, asproviders of resources and logistics, as proxy targets(‘human shields’), or, in the worst case, as the object of waritself. NGHA interests in accessing these populations caneasily clash with ANSAs, as civilians are regarded asbeneficiaries by the former, but as decisive resources by thelatter. Hence, agreement on humanitarian access willdepend mainly on the significance of the role civilians play inthe interests of the ANSA with whom access is negotiated.

Trinitarian and non-trinitarian warfare

To understand this shift in the position of civilians inwarfare, the idea of a transformation from ‘trinitarianwarfare’ to ‘non-trinitarian warfare’ is useful (see Box2).10 This clarifies that targeting of civilians in war is not anew phenomenon; however, the manner in which civiliansare configured in a belligerent’s war strategies haschanged the character of this targeting. Rather thanbeing tactical targets, civilians have moved to theforefront of warfare as objective targets in a deliberatestrategy of control. ANSA commanders, as newbelligerents, have concomitantly adapted their strategiesboth with regard to civilians and with regard to NGHAsand other humanitarian agents coming to the assistanceof civilians.

Humanitarian engagement with non-state armed actors: the parameters of negotiated access

6

Box 2

Trinitarian and non-trinitarian warfare

Carl Von Clausewitz, the classic nineteenth-century theorist of conflict, described warfare as comprising a trinity ofelements: a government, its army and the civilian population from which that army was drawn. In ‘non-trinitarian’ warfare,this relationship between government, army and civilian population is distorted. The ‘government’ may or may not exist,or at least not in the sense in which Clausewitz understood it. In internal conflict, the government is challenged by otherarmed groups, or its control over territory and population may be incomplete; in some cases, there may be no centralauthority at all. These ‘non-trinitarian’ wars tend to be about controlling a population (or denying that control to anopponent), rather than controlling a specified territory (the territory of an opponent state, for example). Here, therelationship between the non-state group, the government and the civilian population is of crucial importance.

Figure 2Trinitarian warfare

Government A

Citizens Army

Government B

Army Citizens

Figure 3Non-trinitarian warfare

Government (?)

Civilians

(competing)ANSA

Army/paramilitary

Opposition(ANSA)

NP 51 2nd 16/6/05 10:24 am Page 6

The most general definition of an armed non-state actor isas follows: ‘Groups that are armed and use force to achievetheir objectives and are not under state control’.11 Thisdefinition is useful in evaluating engagement with armedgroups, as it avoids politically-charged terms like ‘terrorist’and ‘freedom fighter’, and is not specific with regard to theconflict dynamics (whether armed groups are involved in abattle against a state, for example, or between each other).At the same time, however, such a broad definition doesnot fully take into account the wide variety of non-statearmed groups that exist, the diverse conflicts in which theyoperate and the wide range of possible relationshipsbetween the armed group and the civilian population, all ofwhich are important in assessing the reliability ofhumanitarian access agreements.

For the purposes of access negotiations, several featuresof an ANSA have been suggested as being important.These include a basic command structure; independencefrom state control; and the use of violence for politicalpurposes (as opposed to, for instance, criminalobjectives).12 Ideally, these are the minimum criteria thatan ANSA should meet if humanitarian engagement is toresult in effective change in the ANSA’s attitude towardscivilians. A fourth criterion, suggested by this author, is theexercise of effective control over a territory and apopulation. These criteria should not be interpreted toorigidly; rather, they are guidelines by which to judge themotivation and ability of a particular ANSA to respond tohumanitarian engagement, and the reliability of theagreements eventually reached.

Command structure

An armed non-state actor mustdemonstrate some degree of basiccommand structure, though this maynot necessarily be unified. Commandcan be centralised, expressing a higherdegree of organisation, or decent-ralised, as is often the case with groupsoperating in guerrilla warfare. Theefficiency of the control and commandstructures is expressed in a higher orlower degree of discipline among thecombatants. This is a crucial deter-minant in the ability of the group tofulfil security guarantees to an aidagency, and in its ability to abide by IHL.(This qualification obviously alsoapplies to ‘regular’ armed forces;undisciplined troops or an unstructuredcommand may make these formationsunreliable.)

Related, though distinct, ANSAs may sometimes collaborateand coordinate their actions, thus forming a common front.Once the reasons for such a collaboration disappear andrivalry develops, the front disintegrates, often taking accessagreements with it. In Afghanistan during the 1990s, forexample, once the common enemy disappeared with thewithdrawal of Soviet forces in 1989, the mujahideen revertedto internecine fighting. Although the various factionspossessed independent command structures, their rivalryincreased the insecurity facing NGHAs in Afghanistan, whichhad formerly depended on various factions for cross-borderaccess, but were now suspected of being in alliance withrival groups.13 Similarly, once the South Sudan rebelmovement effectively split in 1992, the parameters ofnegotiated access shifted dramatically, as the Sudanesegovernment played one faction off against another. Thefactions themselves then followed suit, playinghumanitarian actors off against each other.

Independence from state control

The level of independence from state control is crucial.Some ANSAs operate as a form of extension or proxy forcefor governments. Examples of such groups include theAutodefensas Unidades de Colombia (AUC) paramilitaries inColombia and the Janjaweed fighters in Darfur, Sudan, aswell as death squads like Arkan’s infamous paramilitaries inthe Balkans or the Interahamwe in Rwanda. Since thesegroups exist precisely to circumvent state or governmentaccountability and operate outside the law, stateinvolvement is often clear but officially denied. Such groups

7

Chapter 3Characteristics and classifications of ANSAs

Abandoned Soviet armour outside Mazar e Sharif, Afghanistan, February 2004. Once Soviet forces withdrew in 1989, the mujahideen

reverted to internecine fighting

©M

ax P. Glaser

NP 51 2nd 16/6/05 10:24 am Page 7

Humanitarian engagement with non-state armed actors: the parameters of negotiated access

frequently do not possess the independent authority todecide their actions, and so engagement for the purpose ofprotection or access will be ineffective. This does not meanthat contacts should never be contemplated with groupsoperating in collusion with the state, but rather that it maybe difficult to identify appropriate levels of effectiveleadership which will assume responsibility, and are willingand able to instruct subordinate troops.

The use of violence for political purposes

This is the hardest characteristic to pin down because of theopacity that surrounds the political agendas of many armednon-state actors. Many have no clearly defined political aims;others claim to struggle for ‘social justice’, but may not putthis into practice through ensuring the protection of civilians.ANSAs may not necessarily aim at regime change or the totaltakeover of the state. Instead, they may operate in a way thatdenies control to the adversary, for example the government,the official authorities or intervening forces, by generating orperpetuating insecurity and instability. Some observersidentify these tactics as attempts to ‘redefine the social andpolitical context by violent means’.14 The fact that a grouprepresents a non-state entity does not in itself make itnecessarily illegitimate, nor does it mean that every actiontaken by state actors (for example the military or the police)is legitimate.

There is also an issue here to do with the scale or nature ofthe violence deployed. When massacres, systematicviolence, mutilation, abduction or rape take place,perpetrated either by state or non-state actors, engagementfor the purpose of access becomes questionable. The resultsof engagement with ANSAs that gravely violate human rightsmay have detrimental effects on civilians. In the Balkans, forexample, the preconditions attached to access played intothe hands of those actors engaged in ethnic cleansing. Whileit can be argued that violence inflicted by extremist groups inIraq and Afghanistan is politically motivated, the deliberatetargeting of humanitarian staff negates any potential positiveoutcome from engaging with this type of ANSA.

Excluded categories

The characteristics of ANSAs described above exclude threecategories of actors: criminal groups, terrorist organisationsand private commercial actors. These groups do not fit thecategory of ANSA as defined here, and as a consequenceengaging with these actors will not yield any positive results,or will expose the NGHA to extreme danger.

Criminal groups are excluded not so much due to theirunlawfulness but rather because they do not aspire tocontrol territories or populations (with the possibleexception of extortion rackets run at local levels), andbecause they use violence mainly for financial, rather thanpolitical, gain. Engaging with such groups may not delivereffective results in terms of protecting civilians, whileexposing staff to risks such as abduction and/or extortion.

Likewise, ANSAs which are involved in parallel criminalactivities (for example cross-border drug and arms trading,human trafficking, abduction or extortion) raise the stakesof engagement for access purposes and for the protectionof civilians, as the NPFL in Liberia, the RUF in Sierra Leoneor UNITA in Angola have all demonstrated.

The terms ‘terror’ or ‘terrorist’ are much more ambiguous.Terror is generally seen as a tactic of warfare aimed atundermining morale. Government armies as well as ANSAsmay use this tactic; the US military did so with its ‘Shock andAwe’ campaign in Iraq. The terms ‘terrorist’ or ‘terror group’mainly serve the political purpose of de-legitimising specificgroups. This is, of course, not to deny that some ANSAsdeploy terror tactics. However, groups which apply terrorexclusively or excessively do not generally appear to beresponsive to humanitarian arguments. One exception tothis may be the Liberation Tigers of Tamil Eelam (LTTE) in SriLanka, a group which has applied terror tactics extensivelyin large-scale suicide bombings, but which has also beenrelatively reliable in its dealings with humanitarian agencies.

The last exclusion concerns private companies, usuallyoperating on behalf of contracting states. Private militaryand security companies are increasingly providingcommercial security services and military and tacticalsupport in conflict zones. Firms such as Executive Outcomesand Sandline have been active in Sierra Leone and Angola,and in Iraq and Afghanistan private companies have beencontracted by the occupying powers to secure roads andcivil and military facilities, and for de-mining. Some of theseactivities will by default also facilitate access forhumanitarian agencies.15 There is an ongoing debate as towhen private actors operating on a commercial basis qualifyas mercenaries. Some criteria to define this status havebeen developed, though the distinction is not clear-cut.16

Nonetheless, this mix of civil and military affairs constitutesa growing concern for humanitarian agencies. Ifhumanitarian access is contingent on the activities of thesecompanies, it would appear more effective to address thecontracting state or the contracting military forces overissues of humanitarian access, since accountability remainswith the contracting parties.

Effective control over population and territory

In addition to the three characteristics outlined above, ANSAsshould exercise effective control over territory andpopulation. The question of control is decisive in determiningthe ANSA’s ability and willingness to stick to accessagreements; it is highly dependent on the extent to which theANSA exercises control over civilians, the manner in which itdoes so and the degree to which it depends on civiliansupport. This is not, however, as clear-cut as it sounds.‘Effective control over territory’ is not synonymous withclearly-defined frontlines or borders; rather, the ANSAdominates a given territory by virtue of its operations andtactics. Such dominance is not necessarily achieved by

8

NP 51 2nd 16/6/05 10:24 am Page 8

Chapter 3 Characteristics and classifications of ANSAs

permanent occupation, as it can be exerted throughintermittent, hidden or remote presence, such as throughcombatants or agents embedded within the civilianpopulation. By the same token, ‘effective control overpopulation’ should not be interpreted as, or confused with,‘good treatment’, nor does it necessarily imply active supportor the identification of civilians with the ANSA’s aims. Controlcan be obtained through abuse, terror and intimidation,through repression, propaganda or intermittent hostileactions in specific areas, designed to deny control to ANSA’sopponents. There is clearly little use in engaging with anANSA for the purposes of humanitarian access if it does notexercise any control over territory or population. But moreimportant than merely establishing the fact of control isunderstanding the quality and the manner by which control isachieved: through actions sympathetic to, and supported by,civilians, or through abuse.

The support of a population for a particular armed group isfar from given, and notions such as ‘popular support’ forarmed struggle are not as straightforward as claims for it byvarious ANSAs may suggest. Civilian support or oppositioncan be invisible, latent or inaccurately expressed. Equally,lack of civilian support, criticism or opposition may remaininvisible, unexpressed or unnoticed. In Colombia and SriLanka, the public expression of opinions about armed non-state groups is stifled by the fear of being identified as asupporter or opponent of the insurgents.17 This may result inpolitical stigmatisation, the loss of economic assets, legalprosecution or, in extreme cases, physical abuse, expulsionor even execution. Considerations like these are particularlyimportant when choosing civilian interlocutors to participatein access negotiations with armed groups (this idea isdeveloped later in this paper).

Classifying ANSAs in economic terms

The analysis of civilians and armed groups in war economieshas produced a general classification of civil–militiarelations expressed in economic terms.18 In a war economy,civilians can be of utilitarian importance to an ANSA’sincome, and civilians can profit from cooperating with theANSA.19 The benefits for civilians can be economic, in theform of income or employment, either by joining the ranks ofan armed group as combatants, or by supplying logisticsupport. Civilians can also benefit from the protection of anANSA in exchange for their collaboration.

The classification outlined here defines ANSA–civilianrelations according to the following categories:

• Symbiotic economic relations: militia aim at restructuringsome social order in exchange for support and revenues,resembling and emulating the function of the state.

• Parasitic economic relations: militia offer protection tocivilians in exchange for collaboration, resemblingmafia protection rackets in their extortive character.

• Independent sources of revenue: militia are notdependent on the population for income (civilians may

participate in cross-border trade or the extraction ofprecious mineral resources).

• Predatory economic relations: militia are careless of thefate of the population, rule through fear and intimidationand prey on the population to increase their power.

These typologies suggest different opportunities forhumanitarian negotiation: in other words, the possibility ofinfluencing an ANSA’s treatment of civilians. Clearly,influencing an ANSA’s behaviour towards civilians will bemost difficult in a context of predatory economic relations,where the group has no apparent interest in improving thefate of civilians. It will be easiest where symbiotic economicrelations exist, when such an interest is assumed. However,the question remains as to what drives an ANSA to engagewith humanitarian actors to negotiate access. In otherwords, what are the advantages or rewards for an ANSA inengaging with a humanitarian agency? Conversely, what arethe disadvantages or penalties of non-engagement? Toanswer this question requires investigation of the potentialof aid or aid agencies to sustain, enhance or impair thecapabilities of the ANSA.

The need for fighters

The need for fighters should be regarded as a key issue in anANSA’s survival. Combat capacity may serve the interests ofcivilians in their search for protection and security, as well asin terms of income opportunities. A prerequisite to civiliansupport for an ANSA is that the latter truly representscivilians’ concerns, and acts accordingly. This is evidentlymore likely when ANSAs are dependent on support fromcivilians (a symbiotic relationship, according to the typologyabove), as opposed to ANSAs that act independently ofcivilian support. Access to food and income are strongincentives to recruitment. Parasitic or opportunistic groupscan utilise these factors to entice and attract fighters to theirranks. Hence, in conditions where incomes are low, andwhere there is political instability and a lack of economicopportunities, it is relatively easy for ANSAs to attractunemployed men. Consequently, being a fighter hasincreasingly become a vocation in itself. Fighters easilyswitch from one ‘employer’ to the other, as demonstrated bydefeated Interahamwe forces during the late 1990s, whodispersed and joined various ANSAs as far afield as Angolaand Congo-Brazzaville. In West Africa, fighters of variousfactions continuously cross over to competing groups.20

Loyalty is no longer expressed in terms of tribal or ethniclineage, as patronage-financing takes over as a recruitmentmechanism.21 In such conditions, personal needs and theprevailing market forces of recruitment become moreimportant than other identity definitions, such as race,ethnicity, political affiliation or geographic origin. Thedecisive role aid can play in raising fighters among refugeecommunities has been addressed extensively,22 but shouldnot be overlooked in the context of internal conflicts, whereemployment conditions offered by an ANSA to prospectivecombatants may include the benefits accruing fromhumanitarian services, such as medical aid, food and shelter.

9

NP 51 2nd 16/6/05 10:24 am Page 9

Humanitarian engagement with non-state armed actors: the parameters of negotiated access

10

External relations: recognition and legitimacy

The presence of humanitarian agencies can itself beimportant to an ANSA in terms of recognition. Not all ANSAsmay be sensitive to this – total spoiler forces in particular areunlikely to be – but others may be responsive to issues oflegitimacy and recognition. Alternatively, an ANSA may seekto delegitimise an adversary, either by demonstrating itsmisbehaviour (in terms of abuse) or by showing the‘rightness’ of its own policies and attitudes towards NGHAs.Permitting or preventing humanitarian action can be used forpropaganda purposes, as the RUF has shown in Sierra Leoneby allowing certain agencies to operate, while rejecting theUN. At a time when human rights violations are increasinglybeing addressed via international tribunals, sanctions andinternational indictments, ANSA leaders are more sensitivethan in the past to exposure to criminal charges, and showmore concern about this risk. To mitigate it, ANSA leadersmay try to ‘play the humanitarian card’ by allowinghumanitarian aid in, thereby showing their respect forinternational conventions and discrediting charges ofmisbehaviour. The degree to which this sensitivity will play arole obviously hinges on the intentions of ANSA leaders, theirpersonality, the tactical position in the conflict and theamount of (credible) pressure exerted upon them. Obviously,this sensitivity to ‘humanitarian concerns’ can be symboliconly, and lacking in true commitment: granting access tohumanitarian aid may thus serve a false function. Thesefactors need to be taken into account, especially whendealing with particularly abusive ANSAs, which are guilty orsuspected of human rights violations and crimes againsthumanity.

The civilian position: in- and out-groupmembership

In order to assess the various difficulties and (unintended)harm to civilians that may result from an NGHA negotiatingaccess with an ANSA, the concept of ‘in-group’ and ‘out-group’ membership can be a useful tool.23 ‘In-group’members are those civilians that form the ANSA’sconstituency; ‘out-groups’ are those civilians outside thatconstituency. In-groups are treated better than out-groups,as they enjoy the protective capacities of ANSAs. Out-groupsare at best ignored, and at worst deliberately targeted, as forexample in ethnic wars. In-/out-group membership can bedecided by ideology, politics, ethnicity, nationality, tribal orclan delineation, by shared economic objectives or bycommon economic grievances.24 The underlying assumptionin this approach is that the larger the in-group constituency,the more responsive the ANSA will be to attempts topersuade it to improve the lot of civilians. Consequently,civilians may possess more leverage over the ANSA tonegotiate better treatment and access opportunities forNGHAs. The larger the in-group constituency, the less likely itis that humanitarian access will be challenged: by organisingassistance for its own constituency, an armed group that isdependent on civilian support is likely to find that supportincreased. However, it is also likely that the members of the

out-group will view this aid as partisan. More importantly, inmany situations it is precisely the out-group that is likely tobe most in need of aid and protection.

Utilising in-/out-group analysis to estimate the challenge tohumanitarian access can also be a dangerous exercise, as ittends to overlook the political dynamics of conflict. It isprecisely when access is least challenged – when an ANSAhas a broad constituency – that aid is most likely to beconsidered as partial (by the ANSA’s opponents), or that itwill fuel the conflict by enhancing the ANSA’s capabilities.Adversaries may perceive assistance as partisan, and maylaunch attacks on civilians, or even on the agencies aidingthem. To guard against a partial response, even-handednessis commonly advised, so that assistance is provided to bothsides in the conflict proportional to prevailing needs.However, in conflicts that generate asymmetric needs – thatis, where one side objectively has more humanitarian needsthan the other or, worse, where only one side is in need –such an approach draws aid into the politics of conflict bydefault. The potential risks involved in such an even-handedapproach were demonstrated in Bosnia, where Serb militiainsisted on the equal distribution of aid on a fifty-fifty basis,whereas an independent needs assessment would havefavoured non-Serb Bosnians.25 Accepting these demandsenhanced support for the Serb leadership among the in-group constituency, and the Serbs sustained their capacity tocleanse the out-group. In Ituri district in the DRC, the Hemaleadership stated that it would respect humanitarian aid solong as its adversaries, the Lendu, were not aided, whereasthe humanitarian needs of the Lendu objectively were larger.This negative attitude among the Hema leadership led to theassassination of five ICRC staff in 2001. MSF had suspendedits activities in 2000 after a series of incidents and threats.26

These examples demonstrate that the treatment of the ‘out-group’ is particularly important, as this may be precisely thegroup that is targeted by the ANSA, and hence is most inneed of aid and protection. Out-group civilians can representthe ‘adversary’, but this is not necessarily synonymous with‘enemy’: civilians can be targeted by ANSAs just for beingsuspected of providing support to opponents. This is thecase in Colombia, Liberia, Sierra Leone, Sri Lanka and theDRC, where massacres frequently occur in retaliation foralleged support to opponents, or as a deterrent to suchsupport. The treatment of the out-group therefore does notfollow the same rationale as that of in-groups. Whereas thelatter may in the worst case simply be neglected by an ANSA,the former will be treated much more harshly. This isespecially the case when ANSAs are not just excluding out-group civilians, but are particularly negative and hostile. Atthe extreme, this hostility takes the form of ethnic cleansingand genocide. In-/out-group membership analysis is thus auseful tool in determining an ANSA’s dependence on in-groups, but it can be misleading as a way of establishing theterms of access to out-groups. The analysis of in-/out-groupdynamics must therefore be carefully evaluated in the light of the prevailing political context and conflict dynamics, inorder to avoid unintended negative and controversialconsequences.

NP 51 2nd 16/6/05 10:24 am Page 10

Chapter 3 Characteristics and classifications of ANSAs

11

Classifying ANSAs in political terms

The issues explored above introduce a political dynamicinto the classification of ANSAs. A classification in politicalterms expresses both the attitude of the ANSA towardscivilians (those targeted by NGHAs), and the quality of theANSA’s dependence on civilians. Such a classificationallows us to identify the potential compatibility orincompatibility between the interests of the NGHA and theinterests of the ANSA. A classification in political termsalso expresses the various combinations of civilian rolesin conflict as presented in Figure 1 (page 3). Thisclassification should not be seen necessarily as analternative to the economic typology described above –rather, it attempts to include the various categories as setout there. The categories are defined as follows, with thecorresponding categorisation of ANSA–civilian economicrelations given in brackets:

• Protective (symbiotic): the ANSA plays an active role inthe protection of civilians and the promotion of civilorganisation. The ANSA and civilians share commonvalues and interests. Civilians are not defined by in-and out-group divisions. Civilians support ANSA aimsand fight on a volunteer basis. The ANSA actively seeksrecognition and is sensitive about human rightsconcerns. Example: liberation movements.

• Competitive (predatory or parasitic): the ANSA acts in

competition with state or non-state actors, rallying thesupport of civilians, or denying that support toopponents. Civilians may be implicated with the ANSAthrough labour relations and (illicit) trade, orcontracted as fighters, but they do not necessarilyshare the ANSA’s interests. Example: factions within awar economy.

• Antagonistic (independent or predatory): the ANSA isdriven by a self-centred identity based on ethnicity orreligion. It is supported by, and raises fighters solely from,the in-group, and acts on that group’s behalf. It is highlyantagonistic towards the out-group, and insensitive to itshuman rights concerns. It seeks recognition for itsfunction of defending the in-group rights. Examples:groups engaged in genocidal war, ethnic cleansing.

• Sectarian (independent): the ANSA is driven byextremist ideology or nationalism. Civilians andfighters are mobilised through the promotion ofextreme views towards out-groups or opponents (forexample, ‘the West’). The ANSA does not seeklegitimacy or recognition, but emphasises its credibilitythrough hostile actions, and is insensitive to humanrights concerns. Example: Islamic extremist groups.

Table 1 summarises this four-fold classification. Itdescribes the type of relationship between the ANSA andcivilians that would correspond to each category, and thenature of the in-out group dynamics.

Table 1: ANSA classification

Protective

Competitive

Antagonistic

Sectarian

Dependency on civilians

The ANSA is highly dependent on civilian support,and is likely to protect civilians and be open toengagement with NGHAs.

The ANSA vies for control over, or support of, civil-ians. Depending on its tactical and politicalposition, the ANSA may be responsive to humani-tarian engagement, but can equally becomenegative if competing ANSAs are deemed to profitfrom aid, or where civilians are not supportive ofan ANSA, or of no use to it economically or asfighters.

ANSA depends on support of the in-group againstthe out-group, and in-group support will be verystrong. Humanitarian needs are likely to be veryhigh in the out-groups, and assistance to theseout-groups is likely to be strongly opposed by theANSA.

Independent from broad popular support. Recruitsfrom extreme political groups.

In-out group dynamics

Strong in-group dynamics. The broader thisconstituency, the more responsive the ASNA willbe to humanitarian engagement.

The ANSA is sensitive to in- and out-groupdynamics, responds positively to engagement foraccess purposes, but likely to abuse this for polit-ical or war- economy goals.

Extremely strong in-out group dynamics. The out-group is the major target in conflict, often definedin political, ethnic or religious terms. A positiveresponse to engagement is vulnerable to abusefor purposes of propaganda or legitimacy.

Extremely limited in-group dynamics, if at all. Out-group can be defined as 'the other' in very broadterms. Highly insensitive, engagement unlikely tohave an effect.

NP 51 2nd 16/6/05 10:24 am Page 11

12

NP 51 2nd 16/6/05 10:24 am Page 12

Key to the question of humanitarian engagement withANSAs is the reliability of the agreements reached. Willthese agreements effectively meet the two basicoperational preconditions of humanitarian action: securityfor aid workers, and respect for the principles of IHL? Thereliability of agreements depends on the willingness of theANSA to comply with the terms, and its capability to do so.This can be formulated as follows:

• Willingness: why is the group receptive to humanitariandemands?

• Capability: how will/can the group adhere to theagreement?

The model followed in this paper is based on a set of implicitassumptions. The first is that the responsiveness of a non-state armed group to demands for humanitarian access isdirectly related to the degree to which the ANSA isdependent on civilians. The greater the degree ofdependence, the more likely it is that the ANSA will respecthumanitarian access agreements. A second assumption isthat an ANSA’s behaviour towards civilians, and itsresponsiveness to humanitarian demands, is stronglyinfluenced by its dependence on external relations, and thedegree to which the group seeks legitimacy, recognition orcredibility. Thus, the more an ANSA seeks recognition,legitimacy or credibility, the more responsive it is likely to beto humanitarian access demands. But it is also important tounderstand the distinctive characteristics of the non-stategroup in question. In particular, how does its commandstructure and objectives influence its capability andwillingness to comply with access conditions?

The evaluation of the humanitarian response in the GreatLakes used the following model (Table 2) to highlightissues of willingness and capability as they related toquestions of access. The recommended approach in thiscontext was called ‘strategic coordination’, involving theUN, the local military and political actors, and rebelauthorities. It aimed to secure the acceptance of warringparties to a framework of consent for IHL and humanitarianprinciples.27

Table 2: Willingness and capability

A similar matrix (see Table 3) can be produced in line withthe analysis of ANSA–civil relations in Chapter 3. Thiscorrelates an ANSA’s structure with its objectives, toassess how challenging humanitarian engagement withthat group might be.28 Narrow objectives signify a highdegree of self-interest (for example, gains from the wareconomy); broad objectives indicate that the group seeksbenefits for the population at large (for example, landreforms), or at least for a wider group than just ANSAcombatants. Clearly, an ANSA that fits in the BroadObjectives box in this table would have greater interest incomplying with humanitarian demands for access, sincedoing so benefits the constituency that the ANSA claims torepresent. The categorisation ‘clear’ or ‘loose’ in referenceto the ANSA’s structure is an elaboration of the minimalqualifying characteristic discussed in Chapter 3: the needfor a basic command structure. Obviously, the clearer thecommand structures the better organised the command,generating a higher degree of internal discipline, andmaking it more likely that combatants will abide by theleadership’s instructions. Hence, adherence to accessstipulations will be more reliable. In a disorganised,chaotic or loosely organised group, there is less likelihoodthat members will comply, and consequently accessagreements with such groups will be unreliable.

Table 3: Objectives and structure

This model complements the willingness/capabilityframework of consent, indicating the most challengingcircumstances for access. Unsurprisingly, these occur whenan ANSA is both incapable and unwilling to comply with thepreconditions for access. However, this model allows us toshow how the ANSA’s objectives and organisationalstructure determine that capability and willingness. Theformer (structure and objectives), in other words, can beused as indicators for the latter (capability and willingness).The task now is to identify the apparent or stated objectivesof the ANSA, and ascertain its organisational structure.

Objectives and tactical position

As the foregoing analysis has shown, ANSAs have widelydivergent objectives. A sectarian ANSA, with no clear

13

Chapter 4Engaging ANSAs: effectiveness and reliability

Willing and able(Pure consent)

Willing but unable(Capacity-building required)

Unwilling but able(Pressured consent)

Unwilling and unable(Enforced consent)

Narrow objectives

Broad objectives

Most challenging

Moderately challenging

Moderately challenging

Least Challenging

Loose structure Clear structure

NP 51 2nd 16/6/05 10:24 am Page 13

Humanitarian engagement with non-state armed actors: the parameters of negotiated access

motive to win a war and thus no clear objective vis-à-viscivilians, will mostly be insensitive to humanitarianarguments, and will act quite predictably: engagement foraccess purposes is likely to elicit a negative response. Aprotective ANSA, on the other hand, is clearer in itspolitical goals, and may be likely to be more sensitive toarguments in favour of the well-being of the civilians onwhich it depends. In between, one finds the difficult cases:opportunistic and competitive ANSAs, engaged in war fortheir own benefit or in competition with each other, whileat the same time making virtuous claims for the social orpolitical objectives of their struggle (for example,Colombian guerrilla groups). An ANSA that is stronglyantagonistic towards a particular group (the out-group)will obviously align its objectives with the interests of thein-group on whose behalf it claims to operate, and onwhich it depends.

The tactical position of the ANSA has a major influence onthe stance it is likely to take towards humanitarianengagement. ANSAs which are highly dependent oncivilian support but in a weak tactical position may beinterested in a humanitarian presence, but unable toprovide security guarantees because their militarycapabilities are overstretched, or because theircommunications and command lines are malfunctioning. Incontrast, an antagonistic ANSA which finds itself in a weaktactical position may be inclined to provide assurances foraccess in an attempt to establish a stronger position forthe in-group civilians that form its constituency. Insummary, responding positively to engagement in itselfmay not mean that an ANSA is actually in a position toprovide security guarantees, either because it is unwillingto do so, or because it is unable to do so.

Organisational structure and internal discipline

Obviously, the ability of a non-state armed group to adhereto access agreements is highly dependent on its internalorganisation. As described above, command structurescan vary from a dispersed (loose) structure, where thegroup operates in independent units, to a centralised andhierarchical (tight) structure. The degree of discipline is notnecessarily directly related to these two extremes, thoughit is safe to assume that, once a centralised ANSAdemonstrates a high degree of discipline, the combatants’behaviour will be consistent. The degree of discipline in aloosely organised group is primarily dependent on thequality of the group’s commanders, and on the instructions(if any) given from the centre. Payment of combatants mayalso play a crucial role in internal discipline. More looselyorganised groups are more likely to pay combatants in lootor booty, thereby increasing the risk of misconduct. In bothtight and loosely organised ANSAs, internal discipline maybe disrupted if disputes break out between commanders,perhaps over political direction or strategy, though this riskis higher in loosely-organised ANSAs, and in coalitions ofindependent ANSAs. The position the ANSA takes on a

ceasefire or peace negotiations may also trigger seriousclashes and disputes, as has repeatedly happened inSomalia and, more recently, in Darfur. The case ofAfghanistan has already been noted. Once the anti-Sovietcommon front melted away, rivalries and inter-factionalfighting broke out, and the security environment for aidworkers deteriorated.

Analysing the risks and benefits of humanitarian engagement

Once one has investigated the particular manner in whichANSAs are configured – both internally, in terms oforganisational structure and objectives, and externally, interms of the specific ways in which civilians relate to theirinterests (in-/out-group analysis) – it is possible to perform arisk–benefit analysis of the intended engagement with theANSA for the purpose of humanitarian access. The risks andbenefits are diverse, and depend on the tactical position ofthe ANSA, the nature of the conflict and the specific conflictdynamics (is it internal conflict, as in Colombia, or a conflictstemming from an external intervention, as for example inIraq?). Hence, a specific risk–benefit of humanitarianengagement in particular cases cannot be defined here, sinceeach conflict arena is unique. Rather, some general risks andbenefits can be identified, to be applied in each and everysituation, and reassessed as time passes and humanitarianaccess is achieved. In other words, it is not sufficient toperform such analysis once: it must be done prior toengaging with the ANSA, but it is also extremely important toreassess the analysis, in particular when there arefundamental changes, such as the introduction of a newANSA, a split in the ANSA, the arrival of peacekeeping orenforcement forces, or the achievement of a ceasefire orpeace agreement.

Table 4 describes the potential risks NGHAs and civiliansface in dealing with ANSAs in each of the four categoriesset out in Chapter 3.

It is also possible to explore the positive and negativeaspects of an access agreement for each of the partiesconcerned – the NGHA, civilians and the ANSA – in a moregeneral way. These are set out in Table 5; they are notspecified for each category of ANSA as defined above.29

As stated above, it is beyond the scope of this paper tocomprehensively explore all the combinations of risks andbenefits possible in the wide variety of conflict situationsand different ANSA typologies. The precise outcome of sucha risk–benefit analysis depends on many factors, such as thenumber of ANSAs involved, their attitudes towards eachother, their relations with civilians, the eventual presence of(international) intervention forces and the position of theincumbent government and its military forces. Someexamples illustrate the wide variety of possibilities.

• In Liberia, humanitarian access in rebel-held territoriesin the 1990s was opposed by ECOMOG, the West

14

NP 51 2nd 16/6/05 10:24 am Page 14

Chapter 4 Engaging ANSAs: effectiveness and reliability

1155

African peacekeeping force that deployed to thecountry in 1990. In an extreme response, ECOMOGaircraft bombed a convoy of humanitarian vehicles enroute to rebel areas.

• In Colombia, NGHAs trying to reachcivilians in contested areas facedaccusations that they had beeninfiltrated by informants. These accu-sations came variously from guerrillagroups, paramilitaries and govern-ment forces.

• The authorities in Sudan have made itvery difficult for NGHAs to gain accessto areas controlled by rebels in Darfur.At the same time, NGHAs operating inrebel-held areas with northernSudanese staff have faced mistrustfrom rebels.

The variety of possible responses toattempts at securing humanitarian accessdemonstrates the need to properlyanalyse the relationship between all the

actors, including relations between armed groups andcivilians, to identify which actors may feel threatened byaccess or may oppose it, and which have an interest inencouraging access and so may accept it, and why.

Protective

Competitive

Antagonistic

Sectarian

Table 5: General risks and benefits from humanitarian engagement

Threats generated by type of ANSA–civil relations

Accusations against NGHA by incumbentregimes

Accusations against NGHA by competing ANSA

The (political) abuse of aidAbuse of aid to war efforts

Accusations against NGHA by internationalbodies or governments

Risks generated by humanitarian engagement

Loss of neutrality and/or impartiality of NGHAUndue legitimacy of ANSA

Retaliation on civiliansThreats against NGHAUndue recognition of ANSA

Loss of neutrality /impartialityUnintended consequences (e.g. aiding ethnic cleansing)

Insecurity of aid workers and attacks on NGHAAttacks on civilians

To civilians

To NHGA

To ANSA

Table 4: Risk–benefit analysis according to the classification of ANSAs

Positive effects (benefits)

• Access to humanitarian aid• Increased protection

• Fulfil mandate • Meet needs • Staff security

• Access to dialogue• Influence behaviour

Negative effects (risks)

• Perceived as sympathetic to ANSA• Retaliatory attacks by rivals of ANSA

• External accusations• Internal NGHA divisions• Attacks by competing ANSAs

• Acquire undue legitimacy• Increased conflict

(other ANSAs)

Sudanese government attack helicopters at Nyala airport, Darfur, December 2004. The Sudanese authorities have made it very difficult for

humanitarian agencies to gain access to areas of Darfur controlled by rebels

©M

ax P. Glaser

NP 51 2nd 16/6/05 10:24 am Page 15

Mitigating contextual factors: securitystrategies and promoting IHL

As stated at the start of this paper, the objective ofhumanitarian engagement with ANSAs is to fulfil twofundamental preconditions – the security of aid workersand respect for IHL. Two approaches have been applied byvarious humanitarian actors to mitigate the consequencesof ANSAs’ disrespect for IHL and prevailing insecurity.These are the adoption of an ‘acceptance strategy’ forsecurity, and the promotion of IHL.

Acceptance strategies

The ‘acceptance strategy’ aims to reduce agencies’vulnerability to insecurity by enhancing acceptance for thehumanitarian presence in the surrounding environment.30 Inthis way, the civilian population, which is sympathetic to thehumanitarian presence, may act as an early-warningnetwork, thereby increasing the security parameters inhostile environments. Inherent in this approach is theassumption that a viable positive connection exists betweenthe civilian population and the ANSA. As discussed above,such a connection is not always evident, as the ANSA mayalso be in a negative and hostile position with regard to thecivilians concerned. An acceptance strategy will thus onlywork in specific types of conflict; it is most suited to contextswhere the ANSA is protective, but in these environmentssuch a strategy is probably redundant. Conversely, it willdefinitely not work with sectarian ANSAs. The real challengeis in cases of antagonistic or competitive ANSAs. In thesecases, the specific conflict dynamic and the position ofcivilians vis-à-vis the armed group must be established first,in order to assess the effectiveness of the acceptancestrategy. If an ANSA has a particularly negative attitudetowards the civilians the humanitarian agency is aiding, theacceptance strategy may at best act as an early-warningmechanism, but it will not necessarily provide a securityguarantee to aid workers.

Promoting IHL

Promoting IHL principles among ANSAs to enhancehumanitarian space is a traditional activity of the ICRC, andseveral NGHAs have followed its example. However, thisstrategy has produced only limited results. ANSAs often actin breach of IHL and abuse civilians as a deliberate strategyor as a survival mechanism. It appears that the willingness ofa particular group to accept IHL rules depends on whetherdoing so coincides with its tactics and interests. Theexperiences with promoting IHL described in Box 3demonstrate that ANSA commanders can learn to use IHLterminology and language, which is perhaps at least a goodstarting-point. However, in the application of IHL rules, othertactical and pragmatic interests or sheer opportunism marktheir behaviour. Since the principles of IHL are non-negotiable, the view that promoting IHL is more comparableto ‘persuasion’, rather than negotiation, seems accurate (seethe section on choice of interlocutor in the next chapter).31

1166

Humanitarian engagement with non-state armed actors: the parameters of negotiated access

Box 3

Persuasion: promoting IHL to change combatants’ behaviour

DRC: ACF’s sensitisation campaign in South Kivu

During 2000–2001, various so-called Mai-Mai militia andother competing ANSAs were cause for increasinginsecurity in the province of South Kivu, DRC. ACFlaunched a broad campaign to disseminate IHL principlesand values to these groups, in order to make them morecompliant with humanitarian access needs, and morerespectful of the status of civilians. According to ACF, theaim was ‘to help all armed groups understand theprinciples and life-saving importance of humanitarianaid’.32 Although the campaign managed to reach manydifferent commanders and groups, and as such can beseen as a success, its results were limited. It appears thatthe commanders’ knowledge of IHL rules increased, buttheir behaviour did not alter significantly. Although ACFtemporarily enjoyed greater secure access, one of its staffmembers was later briefly abducted, casting doubt on thelasting effects of such dissemination campaigns.

Colombia: ICRC’s promotion of IHL amongst paramilitaries

The ICRC has worked to disseminate IHL principles toparamilitary forces in Colombia. A study of this work bythe Geneva-based Center for Humanitarian Dialogue(CHD) found that it had increased knowledge of IHLamong senior paramilitary commanders. It did not,however, change paramilitaries’ actual behaviour towardscivilians.33 The best that could be said was that, withoutdissemination, the situation would have been much worse(as one paramilitary member was quoted as saying).Similarly, Human Rights Watch (HRW) found thatknowledge of IHL among guerrilla leaders was relativelygood, but in practical terms IHL principles meant little.

These experiences highlight the limited effect ofdissemination beyond the integration of IHL language intothe ANSA lexicon. This can be seen as another dimension ofthe ‘learning belligerent’ phenomenon discussed in Chapter2. IHL promotion campaigns may lead ANSAs to tolerate anagency’s presence for a time, but changes in tacticalpositions or the appearance of new actors will progressivelyundermine these effects. The immediate impact of suchcampaigns is therefore to temporarily increase ‘agencyspace’, as opposed to ‘humanitarian space’, or a generalrespect for IHL. Paradoxically, the promotion of IHL is mostneeded in situations where it is least likely to be effective.This does not mean that the dissemination of IHL ispointless, as the increased use of IHL language can have along-term effect. But it does suggest that, in the mostdemanding situations, the immediate impact on accessconditions and security remains limited.

NP 51 2nd 16/6/05 10:24 am Page 16

Chapter 4 Engaging ANSAs: effectiveness and reliability

1177

The degree to which ANSA commanders may be persuadedto accept IHL is highly dependent on the degree to which anANSA relies on the civilians NGHAs are trying to reach. Theanalysis of the different types of ANSA–civil relationsdescribed in Chapter 3 is indicative of a particular group’spropensity to be persuaded: a protective ANSA will clearly bemore responsive to persuasion than antagonistic andsectarian groups.

The question of how agencies should react in operationalterms in the face of an ANSA’s disrespect for the principlesof IHL and human rights is a vexed one. In Sierra Leone, forexample, most agencies refused to work in territoriescontrolled by the RUF. This was in line with a UN-led policyof isolation. The exceptions were MSF and ACF, whichcontinued to operate in RUF areas, and were consequentlycriticised for doing so by other agencies and the UN. TheUN Secretary-General referred to this as ‘breaking acommon approach towards a misbehaving warringparty’.34 In their decision to engage with the RUF, MSF andACF were placing the imperative to help at least somecivilians above the highly doubtful contention that the

RUF’s behaviour could somehow be ‘corrected’ throughisolation.

This episode also shows that the acquisition of ‘agencyspace’, where certain agencies were allowed to operatethanks to limited concessions by the ANSA, is not thesame as the acquisition of ‘humanitarian space’, wherethe ANSA respects the general principles of IHL. Thepromotion of IHL may succeed in achieving the former,but it does not necessarily promote the latter. Someagencies may be temporarily granted access and givenassurances that access conditions will be respected,while it may well be the case that the same ANSAcontinues to breach IHL, commit human rights violationsand/or deny access to other humanitarian actors.Notwithstanding some successes, ‘correcting’ thebehaviour of an abusive ANSA is well beyond the scope,and role, of NGHAs. The protection of civilians exposed tosuch abuses should be sought in credible (international)responses, forcing warring parties, governments andnon-state armed groups to respect the status and rightsof civilians in war.

NP 51 2nd 16/6/05 10:24 am Page 17

1188

NP 51 2nd 16/6/05 10:24 am Page 18

As stated at the outset of this paper, various immediatequestions emerge when considering the modes ofhumanitarian engagement. Should one approach ANSAsdirectly or indirectly? Should the ANSA be engaged at thehighest level possible or the lowest, in the field? Shouldengagement be done openly, or confidentially?

Direct or indirect engagement?

The question of whether engagement should take place ina direct or indirect manner hinges mainly on howapproachable the ANSA in question is. Contacts cannotalways be made directly, as the ANSA may operate in asubversive fashion, or its leadership may be in a remotearea. Indirect (or for that matter confidential) contact mayalso be preferred in cases where the incumbent authoritiesor competing ANSAs are sensitive to such engagement.There may also be legal objections to direct contact fromjudicial or political institutions of the state. In these cases,indirect contact may be facilitated through parallelchannels by the ICRC or diplomatic envoys, but this is notalways a viable option.

Direct engagement takes place either with top or field-levelcommanders; in the latter case, these commanders willoperate more-or-less as mediators between the NGHA andthe higher-level leadership. Figure 4 shows a simple modelof the direct and indirect contacts between an NGHA, thecivilian community and an ANSA. The most crucialrelationship is that between the potential civilianinterlocutor and the ANSA command.

Questions that need to be asked here include: can thecivilian interlocutor ‘negotiate’ safely with ANSAcommanders, or is there the risk that they will facepersecution by the authorities or rival ANSA, or evenretaliation from fellow civilians in the community? Whatseems to enable a civilian interlocutor to influence ANSAcommanders? Does this potential influence stem from thecivilian’s privileged position, granted by ANSAcommanders? It is also important to understand thecommand structure and degree of internal disciplineprevailing in the ANSA, and the degree of independence ofthe civilian community from th