aid work in insecure enviroment

8/9/2019 Aid Work in Insecure Enviroment

1/12

Introduction

In 2008, 260 humanitarian aid workers werekilled, kidnapped or seriously injured in violentattacks. This toll is the highest of the 12 yearsthat our study has tracked these incidents. Theabsolute number of attacks against aid workershas risen steeply over the past three years, withan annual average almost three times higherthan the previous nine years. Relative rates ofattacks per numbers of aid workers in the fieldhave also increased by 61%. The 2008 fatalityrate for international aid workers exceeds thatof UN peacekeeping troops.

This HPG Policy Brief updates the findings fromthe 2006 report Providing Aid in InsecureEnvironments: Trends in Policy and Operations.Its analysis follows on from that report, provid-ing the global incident data for the last threeyears. It identifies new trends and highlights

issues in the three most violent contexts for aidworkers at present: Sudan (Darfur), Afghani-

stan, and Somalia.

Methodology1

Like the 2006 report, this paper is based ondata from the Aid Worker Security Database(AWSD), created as part of an independentresearch project jointly conducted by theOverseas Development Institute in Londonand the Center on International Cooperation inNew York, and kept current since then. Theresearch team also conducted a new series ofinterviews with humanitarian programme and

security professionals and drew upon recentadditions to the literature. Starting in 2006,

Providing aid ininsecureenvironments:

2009 UpdateTrends in violence against aid workersand the operational response

HPG Policy Brief 34hpgHumanitarianPolicy Group April 2009

Abby Stoddard, PhD, is a Partner with

Humanitarian Outcomes and a Non-

Resident Fellow with the Center on

International Cooperation (CIC), New

York University.

Adele Harmer, MSc, is a Partner with

Humanitarian Outcomes and a

Research Associate with the

Humanitarian Policy Group at the

Overseas Development Institute.

Victoria DiDomenico, MS, is a Project

Officer in the International Security

Institutions programme with CIC, New

York University.

Overseas Development Institute111 Westminster Bridge Road

London SE1 7JD

United Kingdom

Tel. +44 (0) 20 7922 0300

Fax. +44 (0) 20 7922 0399

Websites: www.odi.org.uk/hpg

and www.odihpn.org

Key messages

Attacks against aid workers haveincreased sharply since 2006, with a

particular upswing in kidnapping. Surges in attack rates were seen especially

for NGO international (expatriate) staff andUN local contractors.

The three most violent contexts for aidwork Sudan (Darfur), Afghanistan andSomalia accounted for more than 60%of violent incidents and aid workervictims.

Attacks on aid workers in the mostinsecure contexts were increasingly

politically motivated, reflecting a broadtargeting of the aid enterprise as awhole.

Despite some major strides in securitymanagement, aid organisations facesignificant dilemmas in certain threatenvironments, with short-termadaptations often compromising longer-term security.

Abby Stoddard, Adele Harmer and Victoria DiDomenico

Overseas Development

Institute

1 An extensive methodology detailing the definitionsand parameters of the study can be found in the 2006report, available at http://www.odi.org.uk/hpg/aid_insecure_environments.html.


2/12


3/12


4/12

4

hpg Policy Brief 34

Looking at the attack rates of individualorganisations does not reveal strong patterns thatwould suggest that certain profiles or approaches

are more likely to be targeted. From a preliminaryreview of individual agency rates, it does notappear that organisations which we might expectto be more popular targets faith-based agencies,vocal advocacy actors or US-based organisations,for instance were experiencing a higher rate ofattacks compared to the rest of the community.4

Analysis

Concentration of incidents in a few highviolence contextsThree-quarters of all aid worker attacks over thepast three years took place in just six countries, allwith ongoing armed conflicts (in descending order:Sudan, Afghanistan, Somalia, Sri Lanka, Chad, Iraqand Pakistan). A closer examination of incidentrates reveals that the spike over the past three yearswas driven by violence in just three contexts: Sudan(Darfur), Afghanistan and Somalia. This is a morepronounced clustering of incidents in a smallergroup of countries than seen in previous years.

Figure 5: Highest-incident countries20062008

When controlling for aid worker victims in thesecontexts, the long-term overall major attack ratefor humanitarians is actually declining, with anaverage of 2.4 aid worker victims per 10,000 overthe past three years, down from 2.7 in theprevious three-year period. Across the rest of theworld, then, it would seem that the securitysituation for humanitarians is improving, albeitonly slightly.

This finding represents both good news and badfor the international aid community. On the one

hand, it suggests that improved security aware-ness and management may have helped arrest andpossibly reverse the long-term general rise in

casualty rates that was evident in 2005. At thesame time, however, it highlights the dearth ofviable options to keep staff secure in the most

volatile contexts, where humanitarian aid is mostneeded.

Figure 6: Global rates excludingAfghanistan, Somalia, and Sudan

Although complete, reliable numbers for thehumanitarian aid worker population are notavailable for all three of these specific cases, weare able to ascertain that the attack rates areinordinately high relative to other settings. InSudan (Darfur) in 20062008, the annual average

attack rate was 27/10,000 (it has been decreasingfrom a high in 2006 of 66/10,000).

In Somalia, looking at just UN workers, the attackrate in 2008 was 40.9/1,000 in total, and46.7/1,000 for Somali nationals alone (equivalentto a staggering 409 and 467/10,000). InAfghanistan, the increasing number of incidentsand victims, combined with the fact thatorganisations are reporting a general retrench-ment to provincial capitals and a shrinking of theoverall field presence, suggest that rates arelikewise far higher than average. These countries

have in common a setting of active conflict, withbroad swathes of territory where attacks can beperpetrated with impunity. While most attacks inSudan are attributed to common banditry, inAfghanistan and Somalia criminality has colludedwith political forces pursuing national (and in thecase of al-Qaeda, global) aims. To perpetrators inthese areas, targeting aid organisations can gainthem access to economic resources, remove aperceived threat to control over a local area and/ormake a potent political statement.

Tactics

Kidnapping of aid workers has increased by over350% in the past three years, a greater rise thanany other tactic or method of violence. Kidnappersfavour international staff over nationals as victims,because they are both more valuable in terms ofransom and make for a more visible political

0 10 20 30 40 50 60 70 80 90 100

Separate incidents of major violence

Pakistan

Iraq

Chad

Sri Lanka

Somalia

Afganistan

Sudan

4 A planned study looking at country-specific attack ratesfor organisations, factoring in details of their programmingand presence, will be required to generate more definitivefindings.

97 98 99 00 01 02 03 04 05 06 07 08

4

3.5

3

2.5

2

1.5

1

0.5

0

Victims

per10,000inthefield


5/12

5

statement. This has led in some instances, notablyAfghanistan, to cooperation between criminalelements and political groups/militias, where a

kidnapping will be perpetrated by opportunisticcriminals who then seek to barter the hostage to agroup seeking to advance a political or militaryagenda.

Six aid workers were killed in suicide bombingsover the past three years, a tactic that did not affectaid workers before 2003. In Afghanistan and Iraq,improvised explosive devices (IEDs) were used forthe first time in attacks on aid workers.

The most dangerous location for aid workers in20062008 remained the road, with ambushes(including carjacking, banditry and other vehicle-based attacks) by far the most common context forviolence. The majority of kidnaps took place whilethe victim was travelling in a vehicle.

Figure 7: Methods and tactics of violenceagainst aid workers 20062008

MotivesTaking into account the contexts where incidentrates have been rising, the apparent targeting ofinternationals and the means by which violence isbeing perpetrated (i.e., the emergence of suicidebombings and targeted IEDs and the steep rise inkidnappings, which link criminal and politicalactors), it is reasonable to conclude that theincrease in violence against aid workers seen during

the past three years is at least partly politicallyoriented. For many incidents it is difficult toascertain a motive; in 55% of the incidents recordedin the AWSD for 2008, for example, the motive islabelled as undetermined. For the remainder,however, reasonable determination can be made

based on incident reports and the judgments of theoriginal reporting entities.5 Of these incidents, theanalysis reveals that political motivations have

increased relative to incidents that were purelyeconomically motivated, or in which the victims roleas an aid worker was incidental to the violence.Politically motivated incidents rose from 29% of theknown total in 2003 to 49% in 2008.

Figure 8: Motives behind attacks

According to the Afghanistan NGO Security Office(ANSO), a pronounced shift has occurred over thelast two years in Afghanistan. In 2007, 61% ofincidents were attributed to criminals and 39% topolitical opposition groups. The ratio switched in2008, however, with 65% of incidents believed to

be the work of armed opposition groups.6

The political targeting of aid workers by belligerentscan be either associative or direct; that is, aidorganisations may be attacked because they areperceived as collaborators with the enemy, be it agovernment, a rebel group or a foreign power; inother cases, the organisation itself may be the pri-mary target, attacked for its own actions or state-ments, or to prevent or punish the delivery of aid to

Briefing Paper 34 hpg

5 Incidents were classified as politically motivated based on acombination of the following factors: 1) first-hand determina-tions and evidence cited in the original incident report; 2)explicit statements and claims of responsibility by perpetra-

tors; 3) tactics used (e.g., bombs, suicide attacks, targetedIEDs, etc.); 4) political/military actors known to be the perpe-trators; and 5) a high degree of deliberate violence withoutapparent economic motive (i.e., aid workers killed or seriouslyinjured with vehicles/facilities burned but not robbed).6 Afghanistan NGO Security Office, ANSO Quarterly DataReport Q.4, 2008, 2008.

Raid/armedincursion;9% Ambush/attack on

the road; 30%

Individual attack/assasination

with smallarms; 19%

All other; 11%

Kidnapping; 25%

IEDsand

bombs;6%

2003 2004 2005 2006 2007 2008

Political (associative or direct targeting)

Incidental (irrelevant to the violence)

Economic (over organisations assets or contractualdispute no political motivation)

80

70

60

50

40

30

20

10

0

Attacksinwhichmotivewasdetermined


6/12

a population. The idea of the danger of associationwith certain governments or armed forces has aparticular salience with aid workers, most of whom

endeavour to be seen as separate and distinct frompolitical actors and activities, and believe that theycan increase their security by demonstrating theirorganisations independence and avoiding proximityto these actors.

The most recent evidence continues to show thateven those agencies that make considerableefforts to disassociate themselves from politicalactors and project an image of neutrality have notbeen immune from attack, suggesting and ourqualitative research with aid organisationssupports this that associative political targetingmay perhaps be less of a concern than direct orwholesale targeting. We would posit that aidorganisations are being attacked not just becausethey are perceived to be cooperating with Westernpolitical actors, but because they are perceived aswholly a part of the Western agenda. It would seemthat the undeniably Western nature and orientationof much of the international aid community is atthe root of the insecurity aid workers face incountries such as Somalia and Afghanistan. Aidworkers report that just a few years ago Afghanlocals made distinctions between organisations,

for instance between agencies that were workingwith the coalition forces Provincial ReconstructionTeams and agencies that were not. This apparentlyhas yielded to an environment where all Western-based international humanitarian organizationsare judged as partisan, save the ICRC, which, atleast in Afghanistan, seems to have effectivelystaked out a special identity and neutral space forits work.

In these highly insecure environments, it seems,the provision of aid itself justifies attack, in thatit represents an obstacle to the objectives of

belligerent groups trying to gain or maintaincontrol of the local area or to undermine centralauthority. Politically motivated violence can workin other ways as well, for example when govern-ments passively allow or actively collude withattacks on aid workers in their countries whomthey perceive as a threat to their power andcontrol.

A protective environment of acceptance for anorganisation and its programming is easier tocultivate locally, and when the threatening partieshave more limited ambitions. Mutual accommo-

dation can and has been reached between aidorganisations and governments or oppositiongroups in localised contexts, where doing so suitsthe objectives of both. For belligerents pursuing anational or even global agenda, however, such asthose groups that align with al-Qaeda, the anti-

Western target becomes increasingly broad, and anaid organisations efforts to distinguish itself or gainaccommodation for its programmes do not carry

much weight. Furthermore, in some remote areasaid workers represent the only or easiest target. Involatile situations, it is exceedingly difficult for anorganisation to assess the level of threat they facingat any given time.

Operational adaptations anddevelopments in policy and practice

In keeping with the findings of our data analysis,agency staff interviewed for this study consideredthat the security environment in which they wereoperating had deteriorated over the past threeyears. Most interviewees cited operations inAfghanistan, Somalia and Sudan (Darfur) asexceptionally challenging. Despite significantinvestment in security management, the policy andprogrammatic frameworks that guide aid agencies inthese contexts appear inadequate to protect staffand operations from the rising number of attacksand new forms of threat. In addition, what seemedlike highly promising new directions in the UNssecurity management have stalled in the face oforganisational obstacles and the shake-up in thedepartment in the aftermath of the bombing of UN

offices in Algiers on 11 December 2007, in which 17UN employees were killed.

Security management: developments andchallengesAt the time of the publication of the 2006 report,aid agencies were grappling with a series ofchallenges regarding operational security. Theseincluded:

The adoption and effective dissemination ofsecurity policy frameworks and procedures.

A reliance on passive approaches to acceptance

whereby acceptance is assumed rather thanbrokered and maintained throughout the life of aprogramme.

The need for a more systematic means to trackand analyse security incidents, both within andbetween agencies.

Human resource challenges such as high staffturnover and inadequate training, particularlyfor national staff members.

Ad hoc approaches to security-related pro-gramming, such as remote management, andinsufficient appreciation of the increased riskssuch approaches posed for local staff and

partner agencies. A lack of criteria to determine when and how

programmes should be curtailed or halted dueto deteriorating security conditions.

Inadequate inter-agency security coordinationglobally and ad hoc approaches in the field.

6

hpg Briefing Paper 34


7/12

7


Over the last three years the security domain in aidoperations has grown, and some of the issues raisedin the 2006 report have begun to be addressed. The

sector continues to professionalise, with moresecurity posts being established at headquarters andin regional and field locations. Security budgets havereportedly increased or at a minimum have succeed-ed in reflecting stated needs for the majority of aidorganisations. Many agencies now have establishedsecurity policies and procedures, and some requirethese to be annually revised and submitted forreview. Others have invested in security audits toanalyse whether practice on the ground reflectsorganisational policies and procedures. Multi-national member organisations such as Save theChildren and Mdecins Sans Frontires, which pre-viously had separate security management for eachnational affiliate, are developing common securityframe-works, ensuring greater consistency in theapplication of security measures across the organis-ation. Agencies report that more staff, includingnational staff, have been trained in crisis and incidentmanagement, and have been given specific securitytraining related to the more exposed roles that theyundertake, such as drivers and guards. Attemptshave also been made to increase the length of stay offield staff, particularly managers, with a view toreducing turnover rates. Much of this work has been

carried out internally, although some agencies havealso solicited assistance from private securityproviders, typically for security training, risk assess-ment and management support.

Our 2006 report documented a transformativepolicy shift within the UN, whereby the newDepartment of Safety and Security (UNDSS) andthe UN aid agencies were developing aprogramme-led approach to risk management.Taking as its starting point the priorities of thehumanitarian intervention, this enabling appro-ach sought to institute the necessary security

conditions to allow UN and partner programmes tocontinue operating in insecure contexts, ratherthan scaling back or evacuating. The Security RiskAssessment (SRA) model developed by UNDSS hashowever not yet been fully institutionalised, andthere have been a series of significant setbacks. Inparticular, the findings of the Independent Panelon Safety and Security of United Nations Personnel

and Premises Worldwide, commissioned in theaftermath of the Algiers bombing in late 2007,pointed to management failures among seniorUNDSS staff; although not implicated, the Under-Secretary-General for Security resigned (at the

time of writing he was continuing in an actingcapacity until his successor was appointed).

The panel also made a series of recommendationsregarding security management, including replacingthe security phase system with the SRA; implemen-

tation of Minimum Operating Security Standards(MOSS) for UN offices in vulnerable locations; andimproving the balance between programme delivery

and security needs in some high-risk areas.7 It isunclear how quickly these recommendations willbe carried out given the uncertain leadershiptransition. In the interim, there is concern that theUN will become more risk-averse, with importantimplications for the security of its contracting staff,its partners and the wider implementingcommunity. Meanwhile, although a number of NGOshave adopted the SRA methodology or other riskassessment frameworks and are undertaking riskassessments independently, it is unclear whetherthese are considered serious decision-making toolswithin the organisation, and such independentassessments lack the broader scope andparticipation of a UN-led SRA.

Overall incident tracking is increasing, and someNGO alliances are attempting to centralise incidentmanagement reporting across national affiliates.Nonetheless under-reporting continues, and thevast majority of medium-sized and small organi-sations have no or inconsistent means to track andanalyse incidents. There is no fully functioningsingle mechanism in the UN for tracking, reportingand analysing incidents affecting the UN family and

partner agencies. NGOs have also failed to set up ashared platform for inter-agency reporting. Overall,the aid world still does not appreciate theimportance of joint incident analysis, and it isproving difficult to shift from an anecdotally driveninformation environment to one based on a morestandardised, centralised approach to document-ing and analysing security incidents. Agencies havesought to make it easier for field staff to reportincidents, including keeping reporting formats assimple as possible, and to eliminate disincentivesfor reporting. For instance, some have assigned theresponsibility of reporting to staff members

outside the security management structure, soindividuals do not feel that their jobs arethreatened if an incident occurs, or alternativelyreports have been tied to an insurance claimprocess so reporting is made more automatic(although this can create other reporting problems,particularly if insurance provisions only coverexpatriate staff).

Inter-agency security coordinationOrganisations remain wary about sharinginformation with others, and sometimes even withdifferent parts of the same organisation. That said,

there are some extremely positive examples of inter-agency security coordination in the field. The ANSOin Afghanistan, the NGO Safety Preparedness and

7 The report, entitled, Towards a Culture of Security andAccountability, was issued on 9 June 2008 (www.un.org/News/dh/infocus/terrorism/PanelOnSafetyReport.pdf).


8/12

8


Support Project in Somalia (SPAS) and the Gaza NGOSecurity Office (GANSO) have all been welcome, andhave attracted wide participation from agencies on

the ground. In other highly insecure contexts, how-ever, it has proved impossible to construct similarentities, and for the most part agencies still rely oninformal means of coordination and informationsharing. At headquarters level, the European Inter-agency Security Forum (EISF) and InteractionsSecurity Advisory Group (SAG) have facilitateddialogue amongst operational and security man-agers in Europe and the United States. Although theyare more communication than coordination mechan-isms, these arrangements have been valuable inencouraging and promoting good practice, as well assharing lessons learned and providing country-specific information in near to real time. Finally, theimportant Saving Lives Togetherinitiative, which wasdesigned to provide a policy-level framework toimprove security collaboration between the UN andNGOs, remains a worthy initiative, but has notreceived the resources it needs and uptake has beenslow. Greater donor support is required.

Trends in operational securityThe security triangle paradigm of acceptance,protection and deterrence remains the conceptualbasis for aid agencies operational security. Within

much of the aid community, the concept ofacceptance cultivating relations with local actorsand communities is still seen as an appropriateand effective approach to security, particularly forNGOs.8 In some of the most insecure contexts,however, most security managers acknowledge thatit is not a viable security strategy. This is the casewhen the threat is diffuse, such as in lawlessenvironments where banditry is pervasive, or whenit derives from belligerents pursuing national orglobal objectives, for whom the efforts and appealsof aid agencies will have little purchase. Securitymanagers point out that, when the aid community

effectively lacks the ability and interlocutors toengage in dialogue with threatening actors, as is thecase for many in Afghanistan and Somalia, accept-ance becomes impossible. This could explain whyagencies that have worked in these countries for adecade or more are now being attacked.

Lacking alternatives, agencies working in highlyviolent settings have emphasised stricter securitymanagement and protective measures, such asmoving in convoys, hardening physical securityaround their facilities and adhering to StandardOperating Procedures. After a major security incident

a vicious cycle typically ensues, whereby staff arepulled back and consolidated at provincial levels,

contracting the organisations field presence andfurther complicating efforts to build local acceptanceand goodwill.

Many agencies have also reinforced their adherenceto humanitarian principles, in particular the principleof independence. Humanitarian practitioners seeindependence as increasingly vital to their ability tonegotiate access and to their overall level of security.This is done primarily by reducing the agencysreliance on institutional funding, especially if apotential donor is negatively perceived by the hostcommunity. How effective this is in maintaining thesecurity of staff is, however, debatable. The ICRC hashad some success in regaining the acceptance thatits unique role bestows, but all manner of NGOs havesuffered increasing attacks irrespective of theirfunding and partnerships. Perhaps more promisingly,a number of agencies have expanded theiracceptance efforts from a focus on attaining localknowledge at the field level to identifying differentlevels with which to forge connections, for exampleestablishing a dialogue with influential nationals inthird countries, establishing regional cells to buildbackground contacts and knowledge on regionaltrends, or investing in media communications withinfluential foreign broadcasters such as Al Jazeera.

Although typically considered as an exceptional andshort-lived means of conducting operations, extremelow-profile approaches and modes of programmingcontinue in Iraq and Afghanistan over four years afterthey were first introduced. In this approach, agenciesoperate without branding and limit their engagementwith the host and even the beneficiary community.Most humanitarian agencies maintain that thisstance, and its opposite, the use of a highly visibledeterrent or counter-threat in the form of armedguards or escorts, is not a desirable strategy, butinsist that they have little option apart fromwithdrawing aid programming altogether. While the

use of commercially contracted armed protection,including guards and escorts, remains very much theexception and is confined to particular places, everymajor international humanitarian agency has paid forarmed security in at least one operational context. In2007, for example, approximately 22% of inter-national humanitarian agencies reported usingarmed security services.9 In countries such as Iraqand Sudan, humanitarian agencies have used privatesecurity sparingly, relying instead on minimising orsuspending operations, withdrawing staff andremotely managing their programmes.

Remote management and the transfer of riskRemote management in the sense that it is used hereis employed to ensure that aid continues to reach the8 The 2006 report noted the particular security challenges

for the UN humanitarian agencies in that they are not sim-ply aid providers but also part of a larger entity with a clearpolitical identity and role. Thus, it is inherently more diffi-cult for UN agencies to cultivate acceptance as independenthumanitarian actors.

9 Abby Stoddard, Adele Harmer and Victoria DiDomenico, TheUse of Private Security Providers and Services in HumanitarianOperations (London: ODI, 2008) http://www.odi.org.uk/resources/hpg-publications/policy-briefs/33-private-security-providers-humanitarian-operations.pdf.


9/12

9


beneficiary population when security constraintsinhibit traditional management methods. In thisscenario, the organisation withdraws or limits the

movements of its international staff, while shiftingmore responsibility to national staff or local partners.This approach is generally based on the assumptionthat local actors face a lower level of risk thaninternational entities or personnel. As explained in2006, however, this assumption is frequently false,and simply shifts the burden of risk to local staff andpartners who often have fewer security resourcesand less training.

Today, agencies are more mindful of the impli-cations of remote management, both for local staffand for engagement in a country in the long term.Remote management, though intended as a short-term expedient, can have a series of knock-oneffects which make it difficult for the agency to re-engage later through more traditional means.These effects include reduced ground-levelinformation, less credibility and lower levels oftrust in the agency, as well as increased risks forlocal implementing actors. As one NGO intervieweenoted, after a year of remote management inSomalia, threats against the agencys national staffhad increased as they became identified asdecision-makers and resource handlers.

Agencies have had to take a hard look at what theyare asking national staff and partners to do in

these insecure contexts. Most argue that theirapproach to local staff has improved, citinginvestment in security training, stress manage-

ment, counselling and support, but most alsoacknowledge that more needs to be done. Partly inthe rush to address the extreme disparity ininternational and national security provision,agencies are now keen to promote equity, and toensure that all staff are treated the same. Thepoint, however, is to differentiate. National staff,because of their job functions and their localrelationships, require specific security measuresthat are proportionate to, but not necessarily thesame as, those provided to international staff.Other issues raised include the fact that incidentsare not documented as systematically fornationals, partly because the risk to the organis-ation is different. For example, injuries to a nationalstaff member do not involve insurance issues, oradvice to embassies or to families in anothercountry. In some agencies, national hires arecovered by separate policies, particularly regardinginsurance (including medical evacuation andkidnap and ransom). Agencies have also had toconsider developing policies and extendingsecurity provisions to partner organisations andbeneficiaries, as the recipients of aid (for instanceAfghan girls receiving education) have also come

under attack. This opens up a new and even morecomplex set of challenges, and have often simplysuspended programming.

Figure 9: Contraction of aid activity/access following violent incidents

2006 2007 2008

Afghanistan Chad Somalia Sudan Others

50

45

40

35

30

25

20

15

10

5

0

Suspensions,withdrawalsandr

elocations


10/12

10


Figure 10: Declining access for UN and NGO personnel in Somalia, 2008

Access issuesAs security worsens, aid operations are often scaledback or withdrawn, affecting both the quality and

quantity of assistance beneficiaries receive. As the2006 report pointed out, measuring access is,however, a challenging pursuit. There are as yet noobjective or robust means to comprehensivelyassess claims that access is declining, and views onthis differ. While the overall footprint of theinternational assistance community might haveshrunk in a given country, some agencies may havemaintained or even increased their operationalpresence in response to the withdrawal of otheragencies. ICRC, for example, maintains that it hasincreased its operational engagement in some veryinsecure contexts, and has done so with inter-national staff and without armed escorts, armouredcars or military protection. Nonetheless, a reviewof incident reports in 2008 shows that, over thecourse of the year, at least 12 large NGO pro-grammes were suspended in six different countriesafter serious attacks, affecting an unknown numberof beneficiaries.

Of the 380 incidents in the AWSD for 20062008, 82resulted in suspension, withdrawal or relocation, in15 countries (Figure 9). While this is by no means anexhaustive survey, the available data do serve to

highlight some trends in access. Each year sawnearly a doubling of the previous number ofprogramme suspensions due to insecurity. Thelargest increase was in Somalia, with nine-foldgrowth over 2007, representing more than 40% ofincidents in 2008. There were also notable increasesin Afghanistan and Chad (both up nearly four timesbetween 2007 and 2008).

UN Office for the Coordination of HumanitarianAssistance (OCHA) has developed a tracking systemto monitor and report access constraints, and this is

currently being piloted in six insecure contexts.10 Itsreports indicate that, from January 2008 onwards,UN international staff presence in Somalia fell by41%; the NGO international staff presence declinedby 13% in south and central Somalia (SCZ), whileincreasing by 50% in Somaliland and Puntland. Forone NGO, the decision to withdraw staff and closedown its programme of therapeutic feeding affected280 severely malnourished children. The agency alsoclosed a surgical programme, providing emergencytrauma and obstetric care, which had performedapproximately 70 operations and 200 emergencyconsultations a month.11

Conclusion

Aid workers in the most dangerous settings face fewoptions. In places like Sudan (Darfur), Somalia andAfghanistan, the choice boils down to reducing orwithdrawing essential aid from needy populations, orrunning intolerable risks to the lives of staff andpartners. We do not disagree with humanitariansefforts to disassociate themselves from political andmilitary actors doing so is a sensible and necessarystep. If the greater portion of international

humanitarian aid organisations were able to achieveindependence and project an image of neutrality thiswould surely enhance operational security andbenefit humanitarian action as a whole. However, it isimportant that organisations are not misled into

Source: OCHA Somalia

10 OCHA, Access Monitoring and Reporting FrameworkSomalia Report (unpublished document).11 Ibid.

SCZ Puntland Somaliland

80

70

60

50

40

30

20

10

0SCZ Puntland Somaliland

80

70

60

50

40

30

20

10

0

January December

UN international staff presence NGO international staff presence


11/12

11


believing that this in itself will result in increasedsecurity for their staff in the most insecureenvironments, at least in the short term.

In the 2006 report we made 23 recommendations foragencies and donor governments. Few have beenfully realised, partly because many require acommonality of purpose and coordinated actionacross the humanitarian community that have yet tobe achieved in the security arena. Individually,however, some organisations have made progress ina few important areas, including increasing securitysupport to local staff, developing incen-tives toreport security incidents and participating in inter-agency dialogue. We would urge continued action inall these areas, in particular a redoubling of efforts

to work together on incident reporting, tracking andsharing, and establishing additional field-levelsecurity services akin to ANSO and SPAS.

To evaluate the nature and level of threat in conflictenvironments, aid agencies must focus theirincident analysis and assessment on identifyingwhen the aid community has become a wholesalepolitical target, and acceptance becomes ineffective.That said, while it is vital to seek security solutionsfor the most dangerous contexts, it is important thatorganisations do not let conditions in these contextsdictate security models elsewhere. In particular, anactive acceptance strategy should be emphasised asan appropriate and principled approach in themajority of aid settings.

Summary table: Aid worker attacks, 19972008

1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008

Number of

incidents 35 27 32 42 29 46 63 63 75 106 119 155

Total aid

worker victims 73 69 65 91 90 85 143 125 172 239 206 260

Total killed 39 36 30 57 27 38 87 56 54 86 79 122

Total injured 6 15 15 23 20 23 49 46 95 87 84 76

Total

kidnapped 28 18 20 11 43 24 7 23 23 66 43 62

Total natl

staff victims 40 52 40 70 62 68 116 101 158 213 171 211

Total intl staff

victims 33 17 25 21 28 17 27 24 14 26 35 49

Natl staff

killed 31 32 24 46 21 30 65 47 49 82 71 104

Natl staff

injured 4 14 14 20 17 22 45 40 91 75 74 63

Natl staff

kidnapped 5 6 2 4 24 16 6 14 18 56 26 44

Intl staff

killed 8 4 6 11 6 8 22 9 5 4 8 18

Intl staff

injured 2 1 1 3 3 1 4 6 4 12 10 13

Intl staff

kidnapped 23 12 18 7 19 8 1 9 5 10 17 18

UN victims 22 24 17 31 28 18 31 11 28 61 38 65

ICRC victims 9 26 7 9 11 7 8 1 3 10 4 4

IFRC victims 10 5 4 3 5 20 11 5 17 8 1

NGO victims 31 14 37 49 48 54 84 98 134 145 143 185

Donor/other

victims 1 2 1 4 2 6 11 4


12/12

Acknowledgments

We would like to thank the wide range of organis-

ations and individuals that continue to assist andsupport this work, by providing time for interviewsand supplying and cross-checking incident data andfield staffing figures. These individuals andorganisations made this research possible. Part-icular thanks are due to the following for providingresearch input and/or peer review: Patrick Brugger(ICRC), Nic Lee (ANSO), Johannes Luchner (ECHO),Laura Marti (WFP), Robert Painter (UNDSS), StefanoPeveri (WFP), Helene Quentrec (NCCI), Jean S.Renouf (EISF), David Richards (SPAS), ManuelRodriguez (ECHO), John Shafer (InterAction), AngelaValenza (OCHA, Somalia), Simon Wainaina (SPAS);and to a wide range of NGO security and operationalstaff, including Peter Buth (MSF), Alexandre Carle(CARE), Dominic Crowley (Concern), Jan Dalheimer(Welthungerhilfe), Pascal Daudin (CARE), MarcDubois (MSF), Dorte Hempfing (CARE), HeatherHughes (Oxfam), Randy Martin (Mercy Corps),Michael ONeill (Save the Children US), Lara Puglielli(CRS), Mike Tomkins (World Vision), and RayWhittaker (Islamic Relief). The paper also greatlybenefited from a technical peer review by ourstatistician-consultant, Professor Paul Thurman

(Columbia University). The study also draws onextensive work undertaken by our co-author in2006, Katherine Haver.

The views expressed in this paper were informed bydiscussions with our interviewees, but do notnecessarily reflect the opinions of those indi-vidualsor their organisations. Responsibility for any errorsof fact or interpretation remains with the authors.

12


References

Afghanistan NGO Safety Office (2009), ANSOQuarterly Data Report, Q4 2008, Welthungerhilfe,Germany.

J. Holmes, Statement by the Under-Secretary-General for Humanitarian Affairs, Security CouncilOpen Debate on the Protection of Civilians inArmed Conflict, 14 January 2009.

A. Stoddard, A. Harmer and V. DiDomenico, The Useof Private Security Providers and Services in

Humanitarian Operations (London: ODI, 2008).

A. Stoddard, A. Harmer and K. Haver, Providing Aidin Insecure Environments: Trends in Policy and

Operations. (London: ODI, 2006).

aid work in insecure enviroment

Documents