july 2005 hpg humanitarian policy group background paper · 2016. 5. 4. · between the political...

Through a livelihoods lensA case study on theimpact of humanitarian assistance in Bosnia-Herzegovina

Elizabeth Stites and Sue Lautze

July 2005

Overseas Development Institute111 Westminster Bridge RoadLondon SE1 7JDUnited Kingdom

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

E-mail: [email protected]: www.odi.org.uk/hpgand www.odihpn.org

Britain’s leading independent

think-tank on international development

and humanitarian issues

About HPGThe Humanitarian Policy Group at theOverseas Development Institute is dedicated to improving humanitarian policy and practice. It conducts independent research, provides specialistadvice and promotes informed debate.

Background PaperHPGHumanitarian Policy Group

Contents

Chapter 1 Introduction 11.1 Livelihood frameworks and emergencies 2

1.1.1 Save’s household economy approach 31.1.2 CARE’s Coping Strategy Index 31.1.3 Oxfam-GB’s livelihood approach 4

1.2 Study methodology 4

Chapter 2 Humanitarian assistance and livelihood patterns in Bosnia 72.1 Livelihood patterns 7

2.1.1 Food: sources, consumption, emergency assistance 8 2.1.2 Food aid: questions of targeting 92.1.3 Shelter assistance: emergency and reconstruction 102.1.4 Shelter assistance: questions of impact 112.1.5 Remarkable impact: protection 122.1.6 The nature of protection 13

Chapter 3 Discussion of findings 153.1 Dependency: fact or fiction? 153.2 The importance of protection: implications 16 3.3 Discussion of the model 17

3.3.1 Positive aspects 173.3.2 Challenges and possible modifications 17

3.4 Conclusion 19

References 21

ii

HPG BACKGROUND PAPER

1

Measuring the impact of aidHPG BACKGROUND PAPER

The amount, in dollar terms, of international humanitarianassistance has grown remarkably since the end of the ColdWar. With the increase in the value of humanitarianassistance, the interest in being able to measure the effect andimpact of the aid provided has risen accordingly. Donors havebecome increasingly involved in the processes of programmedesign, implementation and evaluation, tasks that were oncethe sole province of implementing agencies. For example,both the Office of US Foreign Disaster Assistance (OFDA) andthe European Union (EU) run training workshops forimplementing partners on the design of projects for fundingproposals.1 Political actors have also become more involvedin programme management.The emphasis on demonstratingtangible results also permeates UN agencies, NGOs andcommunity and private giving campaigns. Some recipientgovernments are taking a greater interest in the impact ofinternational humanitarian funds within their borders.2

The combination of the growing emphasis on results and theincrease in financial resources for humanitarian assistancehas led to a growing number of evaluations of programmes,projects, strategies and expenditures conducted by imple-menting agencies, donors, academics or external consultantsin order to assess the impact of assistance. In some cases, acritical study has had direct influence.3 All studies are meant,in theory, to improve the impact of humanitarian aid, but theprimary focus often is on upwards accountability to thedonor. For reasons both of power and practicality, it remainsdifficult to capture the impact of humanitarian assistance inthe lives of disaster-affected people.

This paper uses a livelihoods framework in an effort tounderstand the relative importance of humanitarianassistance during and after the war in Bosnia-Herzegovina.Accessing and utilising humanitarian assistance can be oneaspect of a household’s larger livelihood strategy; rarely is itthe sole form of livelihood, even for populations living inrefugee camps where livelihood options are constrained.4

Where assistance is perceived by external actors to comprise

the bulk of disaster-affected people’s livelihoods, this isgiven the derogatory label of ‘dependency’, rather thanbeing welcomed as evidence of the relevance of humani-tarian assistance in the lives of people whose systems ofmanaging risk and vulnerability are under stress.5

Assessing the impact of humanitarian assistance requires athorough understanding of the range of householdstrategies and activities over time and place, a task rarelymastered even by anthropologists conducting long-termfield research in relatively peaceful settings. In disasterenvironments, dominated by the rapid generation ofinformation, comprehensive knowledge is often a scarcecommodity, while the privileging of select forms ofknowledge can dominate the disaster discourse.6 Donors,NGOs and UN agencies turn at times to academics tobridge the gap between information and knowledge, aprocess that requires compromise both by the academics(in terms of the methodological ‘short-cuts’ such workentails) and by the humanitarian organisations (in termsof the often high costs in time and money hiringacademics involves). In the end, what is often produced isa step above information, but a significant step belowcomprehensive knowledge.7

This paper summarises one such effort to bringacademics and practitioners together in an attempt toproduce a form of knowledge about the impact ofhumanitarian assistance. OFDA funded a team from TuftsUniversity’s Feinstein International Famine Center (FIFC)and Mercy Corps International (MCI) in Bosnia todevelop and implement a survey to measure changes inhousehold livelihoods over time in 394 households in sixheavily conflict-affected villages in rural Bosnia.8

Household livelihood systems were mapped over threedistinct time periods: the end of the Cold War (1989),the height of the Bosnian war (as identified by eachhousehold) and late 2004. This paper discusses themethods and findings of this survey, and some of theissues pertaining to using livelihoods analysis to measurethe impact of humanitarian assistance. The householdstudy sought to address three key questions:

Chapter 1Introduction

5 Harvey, P. and J. Lind, Dependency and Humanitarian Relief: A Critical Analysis, HPGReport 19 (London: Overseas Development Institute, 2005).

6 Hendrie, B. (1997) ‘Knowledge and Power: A Critique of an InternationalRelief Operation’, Disasters, 21, 57–76.

7 See Jok, J. M. (1996) ‘Information Exchange in the Disaster Zone: Interactionbetween Aid Workers and Recipients in South Sudan’, Disasters, 20, 206–15;Keen, D. and J. Ryle (1996) ‘The Fate of Information in the Disaster Zone’,Disasters, 20, 169–72.

8 Stites, E. et al. (2005) ‘Coping with War, Coping with Peace: LivelihoodAdaptation in Bosnia, 1989-2004’, Washington DC: US Agency forInternational Development. The full report is available atwww.famine.tufts.edu.

1 Personal communication, Dr. Angela Raven Roberts, 2 May 2005.2 For instance, in early April 2005 President Karzai of Afghanistan announced

that the Afghan government had ‘a responsibility towards the Afghan people,as well as the taxpayers in the donor countries, to stop NGOs that are corrupt,wasteful and unaccountable’. In November 2004, the Afghan governmentreported that thousands of NGOs had failed to deliver effective assistance tothe Afghan people. Legislation prevents NGOs from bidding for contracts thatare sponsored by the Afghan government. See Maitra and Ramtanu,‘The Party’sOver for Afghan NGOs’, Asia Times Online, 21 April 2005.

3 Examples include Erikkson, J. et al. (1996), The International Response to Conflict andGenocide: Lessons from the Rwanda Experience: Synthesis Report, Steering Committee for theJoint Evaluation of Emergency Assistance to Rwanda; and Karim, A. et al.(1996) Operation Lifeline Sudan: A Review. Geneva: Department of HumanitarianAffairs.

4 Kibreab, G. (1993) ‘The Myth of Dependency Among Camp Refugees inSomalia 1979-1989’, Journal of Refugee Studies, 6, 321–49.

2

A case study on BosniaHPG BACKGROUND PAPER

1. How have livelihoods changed over time in Bosnia?2. What can we learn about the impact of humanitarian

assistance using retrospective livelihood analysis?3. What are the main challenges facing rural communities

in Bosnia today?

The study of Bosnian households sought to examine howthey adapted their livelihood strategies to cope with radicalchanges in the political, social, economic and naturalenvironment. Household livelihood systems in Bosnia havebeen shaped by three dramatic transformations since1989: the transition from a socialist economy to a market-based system; armed conflict (1992–95); and thereorganisation of society in the post-conflict period.Within households, these transitions brought majorchanges in available assets. For example, death, illness andforced migration greatly damaged households’ humancapital; displacement, ethnic divisions and the fracturingof families eroded social capital among relatives,neighbours, business associates and friends; the collapse ofthe economy, prolonged war and widespread unemploy-ment drained financial capital; and violence destroyedphysical assets and left natural assets such as forests tooinsecure to be of use. Households responded to theseevents by using both short-term coping and long-termlivelihood adaptation strategies, including changes inhousehold composition and location. The research soughtto situate the role of humanitarian assistance within short-and long-term strategies.

The study examined humanitarian assistance, inparticular food aid and shelter, only when the surveyinstrument was able to capture the role that aid hadplayed in the livelihoods of the study population. Basedon the known logistical challenges of providing aid tovulnerable populations in Bosnia during the war, and theactual per capita receipts (after taxes, diversions, thefts,dilutions, etc.) of humanitarian assistance recorded inother emergencies elsewhere (see e.g. de Waal 1997), itwas not expected that assistance provided to Bosnianhouseholds would play an overly important role in theirsurvival strategies. As it turned out, while the targeting offood aid during the conflict was notably impartial, noform of direct assistance had a greater impact onhouseholds than protection interventions, in particularthe facilitation of third-country refuge. Protectioninterventions apparently saved more lives and had morelong-term impacts in terms of economic security thanother forms of humanitarian assistance. These findingsare important given the erosion of international supportfor protection and asylum.

In ranking exercises, food aid proved to be the mostimportant source of food for nearly half of the totalhouseholds in the study population when facing the worstof the conflict; in some villages, the proportion was two-

thirds or higher. Over two-thirds of the households thatreceived food aid were in the lowest category of economicsecurity, and 90% of households in the very poorestcategory received food aid. In other words, food aid waswell-targeted and appears to have had a high impact on thehouseholds that received food assistance at a time ofmaximum vulnerability.

External shelter assistance was less important than food aidin household livelihood systems during the war, but it stillplayed an important role both during and after the war.During the conflict, households within the studypopulation relied most heavily upon their own resources,such as receiving assistance from friends and neighbours,renting or squatting. When available, external shelterassistance usually came from the government, in the formof emergency shelter assistance for the displaced. Similarly,after the war the majority of households within the studypopulation primarily relied on their own income or theirown production to secure their shelter needs, althoughwell over one-half received external assistance, and morethan one-third relied on humanitarian organisations astheir primary source of shelter. Post-war shelter assistanceserved an important role in household livelihoods,allowing people to return to their pre-war villages andrebuild homes. The qualitative interviews underscoredthat, for many, returning home was their desiredlivelihood outcome throughout much of the period ofconflict-induced displacement.

1.1 Livelihood frameworks and emergencies

The research conducted in late 2004 was principallyconcerned with the nature of the changes in livelihoodstrategy in rural Bosnia. The particular relationshipbetween the political economy of war and livelihoodssystems has been explored in other HPG papers, so only abrief introduction to a livelihoods framework is presentedhere.9

Both internal and external factors influence householdlivelihood strategies, including the degree and nature ofownership and access to household assets. Householdlivelihood strategies are strongly shaped by the house-hold’s asset base and the restrictions and opportunitiespresented by the policies, institutions and processes – orgovernance environment – within society.10 Livelihoodanalysis considers a range of social issues over time andspace, and how these factors in turn help to shape the waysin which households use their assets. While there is an

9 See, for example, Collinson, S. et al. (2003) Power,Livelihoods and Conflict:Case Studiesin Political Economy Analysis for Humanitarian Action, London: Overseas DevelopmentInstitute.

10 This governance environment has been referred to as ‘Policies, Institutionsand Processes’ (PIPS), ‘Transforming Structures and Processes’ or ‘SocialRelations, Institutions, Trends and Shocks’. See Scoones (1998); Collinson(2000); and Ellis (2000).

3


abundance of livelihoods models,11 Figure 1 shows asimplified version of a livelihoods framework designed byLautze and Raven-Roberts.12

By using livelihood analysis to investigate the impact ofhumanitarian assistance, it is theoretically possible to seethe relative role of humanitarian assistance within a largerset of household strategies. The livelihoods approach takeshouseholds and their efforts to manage risk and vulne-rability as the point of departure, rather than using thehumanitarian intervention itself as the unit of analysis, asis often done in sectoral evaluations. The livelihoodsapproach leaves open the possibility that the aid providedmay be irrelevant for the household, or is used in ways thatrelief agencies could never have imagined. A livelihoodsapproach also makes it possible to examine a range ofassistance entering households, such as food aid, shelterassistance or health care. Measuring the impact of thisassistance can, however, be difficult because the aim ofhumanitarian assistance is ‘to avert negative change’ (suchas death) and to provide intangibles, such as dignity andprotection (Hoffman et al. 2004). Even positive changes orthe more visible effects within a household, such asimproved health and better economic security, are difficultto attribute to one particular input, such as an emergencyhealth intervention or a food for work project. As a

livelihood analysis underscores, households continuouslyadapt their livelihood strategies in response to changes intheir asset bases and the broader governance environment.When received, humanitarian assistance makes up one partof the asset base, and is not necessarily the main or even aprime factor in the adaptive strategies the householdpursues.

1.1.1 Save’s household economy approachSave the Children (UK) developed the household economyapproach based on fieldwork dating back to the Ethiopianfamine of 1983–85. It is a tool for examining food accessand the effects of shocks upon the future food supply atthe household level (SCF 2000; Boudreau 1998). Ahousehold food economy analysis allows practitioners toidentify population groups that are vulnerable to foodinsecurity. The model explores the ways in which foodsources are likely to change in the event of a shock (suchas a crop failure or an increase in the price of staple food).Baseline data is collected, including information on cropprices, food strategies, expenditures, the availability ofwild foods and asset sales. Food, cash income andexpenditures are then converted into ‘food equivalent’units. After establishing a baseline, researchers analyse thelikely effects of a shock upon a household’s ability tomaintain normal consumption patterns from establishedfood sources. This method has been used in a variety ofsettings, including refugee camps in Kenya (Lawrence,Boudreau and King 1996), Rwanda (Lawrence and King1997; King and Fielding 1997), Kosovo (Holt and King2000), Guinea (Lawrence, Reed and Madougou 2000),and Tajikistan (Lawrence and King 2001). The World Food Programme, UNHCR and other UN and inter-national agencies make frequent use of this approach ormodifications thereof in their food security assessmentsand planning exercises.

1.1.2 CARE’s Coping Strategy IndexCARE developed the Coping Strategy Index (CSI) inconjunction with WFP. This model seeks to measurehousehold food security in a simple and relatively rapidmanner (CARE/WFP 2003).The analysis is based on a short

11 For example, Scoones, I. (1998) ‘Sustainable Rural Livelihoods: A Frameworkfor Analysis’, Institute of Development Studies Working Paper 72; Ellis, F. andH. A. Freeman (2004) ‘Rural Livelihoods and Poverty Reduction Strategies inFour African Countries’, Journal of Development Studies, 40, 1–30; Schafer, J. (2002)‘Supporting Livelihoods in Situations of Chronic Conflict and PoliticalInstability: Overview of Conceptual Issues’, London: Overseas DevelopmentInstitute.

12 See, for example, Lautze, S. and A. Raven-Roberts (forthcoming) ‘Violenceand Complex Humanitarian Emergencies: Implications for LivelihoodsModels’, Disasters; Young, H., Osman, A. M., Aklilu, Y., Dale, R. and Badri, B.(2005) ‘Darfur: Livelihoods Under Siege (Draft) Executive Summary andRecommendations’, Feinstein International Famine Center and the AhfadUniversity for Women; Lautze, S. and E. Stites (2003) ‘More than Seeds andTools: An Overview of OFDA Livelihood Intervention 1964–2002’, TheFeinstein International Famine Center; Lautze, S.,Y Aklilu, A. Raven-Roberts,H.Young, G. Kebede and J. Leaning (2003) ‘Risk and Vulnerability in Ethiopia:Learning from the Past, Responding to the Present, Preparing for the Future’,United States Agency for International Development; Lautze, S., E. Stites, N.Nojumi and F. Najimi (2002) ‘Qaht-E-Pool: “A Cash Famine”: FoodInsecurity in Afghanistan 1999-2002’, United States Agency for InternationalDevelopment.

Figure 1: Simplified livelihoods framework

Influence and access Livelihood

outcomesLivelihoodstrategies

Policies,institutions

and processes

Assets/liabilities

4


series of straightforward questions regarding strategies tocope with food insecurity, and the frequency and severity ofthese strategies. Four general categories of coping areexamined, all relating to consumption: dietary change;increasing short-term food access through borrowing, gifts,wild foods etc; decreasing household size throughmigration, temporary fostering etc; and rationing strategiessuch as skipping meals, limiting meal size, and prioritisingindividuals within the household. Field researchers apply aweighted formula to generate a single numeric scoreindicating the degree of food insecurity within thehousehold: a higher score equals greater food insecurity. Apilot study of the CSI from Eritrea shows that the CSImeasure correlates to characteristics normally associatedwith livelihood status, such as food consumption, the valueof household assets and livestock, the householddependency ratio and the area of cultivated rainfed land.Thisimplies that the CSI can be used to rapidly measurehousehold livelihood conditions, as well as tracking longer-term change in household food security (Shoham 2004).

1.1.3 Oxfam-GB’s livelihood approachOxfam-GB uses a livelihood approach to asses the risks tolives and livelihoods in emergencies. This approachconsiders the severity of food insecurity in regard to short-term food consumption and potential longer-term impacton livelihoods (Young et al. 2001). Field assessments useparticipatory rural approaches (PRA), rapid-appraisaltechniques and anthropometric surveys to indicate foodavailability, access to food of different groups, changes infood and income sources and the type of coping strategiesemployed. A major reduction in a main food or incomesource and declines in nutritional status imply foodinsecurity and a possible threat to livelihoods. Unlike themodels used by Save the Children and CARE, Oxfam-GBassesses the severity of food insecurity based on changes ina broad range of livelihood strategies, and can identify avariety of possible livelihood interventions as well as theneed for emergency food aid. Oxfam-GB has used thisapproach in natural disasters and conflicts (in locations suchas Colombia,Tajikistan and Sudan), but some critics feel thatthe model remains most applicable for natural disasters andneeds to be further modified for use in conflicts orprotracted emergencies (Jaspars and Shoham 2002).

The assessment approaches that CARE and Save the Childrenemploy to understand coping systems are intended toevaluate food needs and to identify vulnerable populationgroups.When conducted repeatedly at regular intervals, thesemodels are also capable of demonstrating the impacts of foodaid and assessing a population’s food security status over thelonger term. Oxfam-GB’s livelihood approach is able to takeinto account a greater range of livelihood strategies andpossible livelihood interventions, but ultimately also focuseson the severity of food insecurity.These models have provedeffective in identifying vulnerable populations or household

food needs, and may assist organisations and donors inplanning and targeting emergency food assistance. Suchapproaches, however, do not seek to provide comprehensiveinformation on household livelihood strategies or copingsystems, or to understand the role of the wide range ofhumanitarian assistance interventions and strategies withinthis broader framework.

1.2 Study methodology

The study was conducted in six villages in Bosnia and theRepublic of Srpska.The villages were purposively selected torepresent a broad cross-section of differing characteristics.They included two Bosnian Serb villages, three predom-inately Bosniac villages and one Bosnian Croat village. Thepopulations differed in economic status, access to employ-ment and in the levels of humanitarian assistance receivedduring and after the war.All of the study sites were located inzones of intense fighting during the war. In each location,government military and/or militia forces forcibly displacedall or the majority of the population. Property destructionduring the conflict was widespread and often absolute.

The research was designed to document householdlivelihood systems at three different times: 1989, eachhousehold’s self-defined ‘height of conflict’ and 2004.Thestudy drew on three main sources of primary data: 1) datagenerated from systematic, longitudinal surveys; 2) in-depth interviews with select households; and 3) in-depthsocio-economic studies on each of the six villages.Quantitative data collected through surveys regarded thenature and degree of changes in household composition,priorities for expenditures, livelihood strategies andhuman security.13 Qualitative data collected through in-depth interviews further explored shifts over time in eachof these areas. The analysis was complemented by aliterature review undertaken by the authors with supportfrom research assistants at Tufts University.14

The sample size for each village was determined based onpopulation estimates (number of households) provided bythe local council or the mayor’s office.15 Systematicsampling was used to select participant households.Sampling was representative of households at the villagelevel. A total of 394 households were included in thesample. Sample size was calculated in each communitystudy site to within ±0.10 percent precision of estimate

13 Human security prioritises a focus on individual and human security abovestate security, and includes four key aspects: human rights and personalsecurity, societal and community security, economic and resource securityand governance and political security (Newman and Richmond 2004).

14 Tufts University graduate students Elizabeth Buckius and Anna Mecagniprovided research assistance for the Bosnia report. Leah Horowitz assistedwith a literature review for this article.

15 Brezani, 22 households, 73 persons; Jakes 480 households, 1,800 persons;Krtova 292 households, 856 persons; Potocari, 200 households, 700persons; Prud 300 households, 1,000 persons; Sevarlije 305 households,1,000 persons.

5


(95% CI). The sample size for each community is asfollows: Brezani (17 households surveyed); Jakes (82households); Krtova (74 households); Potocari (66); Prud(80); and Sevarlije (75).

The longitudinal surveys were designed and field tested bystaff from Tufts University (FIFC) and Mercy Corps/Bosniaduring August and September 2004; surveys wereconducted over three weeks in September 2004. Theseresearchers trained teams of Bosnian surveyors to conductthe surveys and team leaders from FIFC and Mercy Corpsworked with survey teams in the field each day.The MercyCorp team leader re-checked the survey data, entered thedata into databases and quality-checked the data to withina 3% error rate. FIFC researchers analysed the data anddrafted the findings, which were then provided for reviewand comment to Mercy Corps/Bosnia prior to finalisationand submission to USAID OFDA.

Surveyors and team leaders conducted in-depth interviews togather qualitative data with approximately 10% of thehouseholds surveyed. These interviews were carried out inEnglish with experienced translators, and lasted approx-imately 1.5 to 3 hours.This data was collected using the samesurvey questionnaire, with follow-up questions andexplanation encouraged. The surveyors also providedqualitative data in the form of their observations, recorded indaily briefing sessions with team leaders.

One respondent was interviewed per household. Theinterviewee was any member of the household old enoughto have an adequate recollection of the pre-war period (i.e.,born in approximately 1975 or earlier). The intervieweeprovided information on every member of the householdpresent at each of the three time periods. A household wastaken to include all members in a shared accommodation,regardless of their relationship. Data was collected onhousehold demographics, occupation, employment,expenses and needs, coping strategies to meet householdneeds, shelter and utilities, food security, health, protectionand physical security. Similar questions were asked across allthree time periods to track changes within these categoriesover time. Additional qualitative data was provided throughin-depth studies of each site. Journalism students from theregion conducted the site studies in each of the six researchareas, and the Mercy Corp team leader quality-checked alldata and translated the studies.

Situating the relative role of humanitarian assistance wascentral to the study design given that the objective was toweigh the relative importance of humanitarian assistance inthe households’ overall livelihood strategies.To this end, thesurvey did not include direct questions regarding theimportance or impact of humanitarian assistance, but rathersought information about the ways in which householdsmet their basic needs in the three time periods in question.These needs were defined as access to food, access to shelterand access to medical care. Respondents ranked theimportance of the various ways in which they met theseneeds. The survey did not seek to capture the quality,quantity or precise source of the assistance received by thehousehold, rendering the analysts unable to draw directconclusions about the impact of each specific relieforganisation active in the communities during and after thewar. If humanitarian assistance was not ranked as importantby a respondent, the issue was not pursued.

The research was unavoidably biased by the academics’affiliation with an operational NGO. In each of the areasstudied, MCI had implemented assistance projects. Itshould be assumed that the respondents were aware of theinterest of humanitarian actors in the study. The extent towhich this influenced the information provided cannot bedetermined, but should be acknowledged. Both Tufts andMCI are recipients of funds from the US government but,unlike other recent studies conducted by Tufts University,USAID had little or no involvement in the formulation ofthe research questions.

Finally, the three time periods examined in the study cover15 years of significant political, economic and socialchange. For the purpose of comparison, 1989 serves as abaseline against which data from the conflict and 2004 aremeasured, but it should not be assumed that householdswere not already undergoing major changes in this period.Unlike evaluations or assessments performed in the midstof a crisis or emergency response, this study wasconducted nearly ten years after the end of the Bosnianwar. This afforded the opportunity to consider the impactof post-war assistance, and also to examine livelihoodadaptations over a longer period. In particular, the effectsof the post-war recovery and the economic implications ofthe transition to a free-market economy were moreapparent than perhaps would have been the case in theimmediate aftermath of the war.

6


7


The survey was initially designed to capture the role of foodaid and shelter assistance in household livelihood strategies;quite unintentionally (but helpfully) the survey instrumentalso documented the impact of protection interventions.Food aid was provided to Bosnia in large quantities over thecourse of the war, as well as to displaced Bosnians living inCroatia, Serbia and other countries.The delivery of food andnon-food commodities to populations within Bosnia wascomplex and required constant negotiation and renego-tiation with the warring parties to establish humanitarianaccess (Cutts 1999). These negotiations caused some toquestion the ethics of maintaining ‘neutral’ dialogue withforces perpetrating gross human rights violations uponcivilians (de Waal 1994). Others, including some whoworked in Bosnia during the conflict, categorised thehumanitarian response as an ‘alibi’ or ‘fig leaf’ that providedcover for the international community’s lack of action toreach a political solution to the conflict (Minear et al. 2004).

Regardless of possible political motives behind theprovision of food aid, external evaluations indicate thathumanitarian action, particularly by the UN, ‘contributedgreatly to feeding the people of Bosnia and preventingstarvation for two winters’ (General Accounting Office1994). Similarly, a multi-agency nutrition survey carriedout in four areas of Bosnia in 1994 concluded that, ‘ifhumanitarian food aid had not been supplied, thenutritional status of the population would have beenseriously compromised’ (Minear 2004).16

Shelter assistance took on special significance because ofthe nature of the violence in the war in Bosnia. Anestimated 2.2 million people – roughly half of Bosnia’spre-war population – were forced from their homes andvillages (International Crisis Group 1997). Forceddisplacement in Bosnia was highly functional: it was the‘overriding aim’, not just an incidental consequence, ofthe conflict, and was central to the strategy of warringparties seeking to establish ethnically homogenous states(Minear 2004). Millions sought emergency shelter inBosnia, in neighbouring countries or elsewhere in Europe,and turned to relatives, friends, strangers and outsideagencies or government bodies for help.

Bosnian households in the study population receivedshelter assistance in two periods, and from two different

sources. During the war, government bodies (local, nationaland international) were the most common providers ofshelter assistance. This changed in the post-war period,when international humanitarian organisations stepped uptheir programmes for shelter reconstruction andrehabilitation.17 Most of the shelter assistance to rural areascame after the conflict ended, as internally displaced peopleand refugees began to return home. However, the pace ofreturn was not uniform across Bosnia or for all populationgroups, and those seeking to return to areas where theywould be in the minority were less likely to returnimmediately after the end of the war. Some who returned totheir villages more than three or four years after the end ofthe conflict found that they had missed out on the bulk ofreconstruction assistance provided in the immediate post-conflict period.

2.1 Livelihood patterns

The entire study population was displaced by the conflict,and livelihood strategies underwent radical changes. Mosthouseholds adopted a range of coping mechanisms inorder to weather the hardship that, for many, continuedinto the post-war period. The economy security ofhouseholds collapsed dramatically during the conflict.18 In1989, only 9% of households had been unable to covertheir expenses, while the majority of households (56%)were putting money into savings. In contrast, by the heightof the conflict 72% of households were unable to covertheir expenses. More than half of total householdsdescribed themselves as being in the lowest bracket ofincome security, describing their household resources as‘insufficient to cover expenses; could not borrow’. Theextent of economic insecurity during the conflict varied byvillage, from a high of 98% of households from Potocariunable to cover expenses, to a low of 34% of householdsfrom Jakes.

16 The survey found no signs of protein-energy malnutrition in children underfive years and no serious micro-nutrient deficiencies in mothers and children,although weight-loss since the beginning of the war averaged 10kg for adults.

17 For instance, USAID’s Office of Foreign Disaster Assistance (OFDA)embarked on a $25 million shelter repair programme in 1996 which fundedthe rehabilitation of more than 2,500 homes in 48 frontline villages. TheEmergency Shelter Repair Program (ERSP) came under criticism forprioritising ‘same ethnic’ as opposed to minority returns that may havepromoted greater ethnic integration, and the programme only rebuilt homesin the Federation, not the Republic of Srpska. Roughly 81% of the rebuilthomes were for Bosniac families. See Fawcett and Tanner (2002).

18 Economic security was determined based on the respondents’ description oftheir household resources according to six possible income brackets: 1)sufficient to cover expenses and to save money; 2) sufficient to coverexpenses but not to save money; 3) insufficient to cover expenses, and hadto spend savings; 4) insufficient to cover expenses, and received help fromrelatives; 5) insufficient to cover expenses, and had to borrow; 6)insufficient to cover expenses, and could not borrow.

Chapter 2Humanitarian assistance and livelihood

patterns in Bosnia

8


2.1.1 Food: sources, consumption, emergency assistance Most rural households in pre-war Yugoslavia farmedkitchen gardens or smallholdings, and used the produce tosupplement their food purchases. People with full-timejobs in industry, manufacturing or the service sector spenta portion of their week working their land. Mosthouseholds in the study population (69%) relied onpurchased food as their primary food source, with theremaining third relying primarily on their own pro-duction. Both sources of food declined in importanceduring the conflict, with 27% of households primarilypurchasing food and 12% of households turning to theirown production. The decrease in the two main sources offood (markets and own production) was due to problemsof access and availability. Contributing factors includedmassive population displacement, prolonged siege oftowns and villages, widespread economic insecurity,hyperinflation, lack of inputs (including labour) andinsecurity on roads and in fields.

Reduction of food intake at the household level was acommon response to the collapse of supplies of normalfood sources. Of the total study population, 62% ofhouseholds reduced their food intake during the war, incomparison to 6% that reported reducing food intake in1989.The village of Potocari saw the greatest percentage ofits residents reducing food intake, at 89%, while only 22%of households from Jakes were compelled to reduce foodconsumption during the war. In 2004, more than one-quarter of households (28%) reported that they continuedto reduce their food intake.

At the height of the conflict, emergency food aid was theprimary source of food for a greater number ofhouseholds than relied on either purchased food or ownproduction combined. Forty-seven percent of householdsreported that food aid was their primary source of food,and an additional 17% listed food aid as their second mostimportant source.19 Table 1 shows the primary and

secondary food sources for the total study populationacross the three time periods.

As Table 1 illustrates, food aid was an important part of thesurvival strategies of households at a time when the majoritywere unable to meet their expenses and were cutting backon food consumption.The importance ascribed to food aidby beneficiary households is substantially greater than seenin some other emergencies, such as the Darfur famine of1984–85, where food aid was estimated to meet only 10%of consumption needs (de Waal 1987). It should be recalled,however, that the food aid provided in Bosnia did not meetall food needs when households were most vulnerable, asnearly two-thirds of households reported that they wereforced to reduce food consumption during the periodidentified by each household as their ‘height of conflict’.Furthermore, accessing the food aid was often difficult anddangerous and the supply was often inconsistent,particularly for households relying on humanitarianairdrops and/or aid convoys.

The percentage of households receiving food aid differedgreatly by village, as did the relative importance of food aidin household survival strategies. Food aid by village isshown in Table 2.

Table 2 indicates that food aid was a very important sourceof food for the households that received food aid duringthe war, with a majority citing this as their primary sourceof food in every village except Prud. Households fromJakes and Prud were the least likely to report receivingfood aid during the conflict (46% and 45% ofhouseholds). This is due, in part, to the fact that manypeople displaced from these villages were able to seekrefuge in Croatia or Western European countries(primarily Germany), and many were able to secureemployment and purchase their own food.

Households from Sevarlije were the most likely to receivefood aid, with 93% receiving assistance and 73% statingthat food aid was their primary source of food. This is inaccordance with the high rate (97%) of households in

Food source Source rank 1989 Height of conflict 2004

Purchased Primary 69% 27% 63%Secondary 28% 11% 31%

Own production Primary 31% 12% 34%Secondary 49% 7% 39%

Food aid Primary 0% 47% 0%Secondary 0% 17% 2%

Table 1: Primary and secondary food sources by percentage of households

19 During the conflict, households also relied on government assistance (7%),assistance from friends and relatives (3%), exchange of own products (1%),churches/mosques (1%) and ‘other’ (2%) for their primary food source.

9


Sevarlije reporting that they were unable to cover expensesduring the conflict. Potocari had the greatest number ofhouseholds unable to cover costs during the war (98%),but only 85% of households reported receiving food aid,and food aid was the primary source of food for only 62%of households.These relatively lower rates are probably dueto issues of supply, as food convoys were prevented fromreaching the Srebrenica municipality (including Potocari)for extended periods during the war.20 Much of theassistance that did reach the municipality was through theoccasional convoy and the airdrops orchestrated by NATOmember states in coordination with UNPROFOR forces.21

The importance of food aid for households in the studypopulation was related, in part, to the location of thesepopulations in some of the most war-affected areas ofBosnia. Most residents of Sevarlije, for instance, remainedin north-eastern Bosnia for the duration of the war, an areawhich was heavily contested. Similarly, many householdsfrom Krtova relocated directly across the frontline intoday’s Republic of Srpska, but although fighting was oftenless intense than on the Federation side, the war was neverfar away. In contrast, most residents of Prud and Jakes wereable to flee the country altogether, suffered fewer directimpacts from the conflict and therefore had less need offood aid.

2.1.2 Food aid: questions of targetingIf we assume that the households closest to intense fightinghad the greatest food needs, then the village-specific data onfood aid indicates that the targeting of assistance wasrelatively accurate, at least for the period of time that

corresponded to the household’s self-defined ‘height ofconflict’. Food aid was most important for those householdsclosest to the frontline (Sevarlije and Potocari), and generallydeclines in relevance for more distant populations. Someexternal reviews of humanitarian assistance to Bosnia,however, show that delivering assistance based on need wasoften extremely difficult, as the political and militaryleadership of the various sides insisted that aid be distributedbased on pre-war population estimates by ethnicity (Cutts1999).This meant, for instance, that roughly 30% of all foodaid provided by UNHCR (the lead agency) was delivered toBosnian Serb areas, as Bosnian Serbs had made up roughlyone-third of the population prior to the conflict. UNHCRinitially distributed food to four categories of beneficiaries:refugees, IDPs, returnees and the ‘war affected’. UNHCRmoved away from these categories in favour of a single‘casualties of war’ category as food needs became morewidespread and forced displacement intensified. UNHCRbegan to distribute food more broadly, and left logistics andthe registration of beneficiaries up to local municipal officials(WFP 1999).

By all accounts, need-based targeting during the war wasextremely difficult. Surprisingly, data from the surveyindicates that the poorest households were most likely toreceive food aid, but also that the poorest income bracketreceived the majority of food aid delivered (byhouseholds receiving aid).Table 3 shows the distributionof food aid by income, and the percentage of householdswithin each income bracket that received food aid duringthe conflict. Note that this table does not show theamount of food aid delivered, only the percentage ofhouseholds within each income bracket that receivedsome food assistance.

Table 3 shows that 76% of all households receiving foodaid during the conflict were in the lowest income bracket.Only 12% of households receiving food aid reported beingable to cover their expenses through their own resources.Furthermore, when looking at the percentage ofhouseholds within each income bracket who received foodaid, 90% of the poorest received food aid during theconflict. In contrast to the findings of UNHCR and WFP, our

Location Height of conflict

Receiving any food aid Food aid as primary source of food Food aid as second source of food

Brezani 82% 65% 18%

Jakes 46% 35% 10%

Krtova 84% 47% 20%

Potocari 85% 62% 18%

Prud 45% 21% 19%

Sevarlije 93% 73% 17%

Table 2: Percentage of households receiving food aid, and ranking of food aid by village (conflict)

20 This hypothesis is supported by data from WFP; the agency reported that foodaid reached roughly 43% of the civilian population in need in Srebrenica inthe first nine months of 1993, compared to 70% in the rest of the country inthe same period. WFP coverage rates were low for other besieged enclaves in1993 as well, reaching 52.4% in Zepa but only 26.4% in Gorazde. Minear etal. (2004: 14).

21 Many of these airdrops were part of Operation Provide Promise, which wasorganised by the US government in July 1992 to provide food and medicalsupplies to the besieged city of Sarajevo.The effort was supported by militaryorganisations and personnel from 21 countries and expanded to provideassistance to Srebrenica and other enclaves cut off from overland convoys. Forbibliography and references see ‘Operation Provide Promise (Bosnia)’,compiled by Melinda M. Mosley, March 2000, at http://www.au.af.mil/au/aul/bibs/opprom/opp.htm.

10


data points to remarkable efficiency in the targeting of foodassistance to those who were most in need within the studypopulation.The different findings may stem from the surveytimeframe, as these food aid statistics represent the periodthat households identified as their most vulnerable time.TheUN data, by contrast, takes a longer view. One factor behindthe high percentage of beneficiaries among the studypopulation may be relatively high rates of residency incollective centres in Bosnia, Croatia or elsewhere in Europe.Most collective centres were supplied with food fromforeign donors and aid agencies, and residents in collectivecentres received regular food assistance.

2.1.3 Shelter assistance: emergency and reconstructionThe forced displacement of nearly the entire studypopulation and the harsh Bosnian climate created pressingshelter needs during and after the conflict. Householdspursued a wide variety of shelter strategies, includingmoving to collective or refugee centres, staying withfriends or relatives, squatting in abandoned houses andseeking refuge in unconventional accommodation, such asrailway cars and garages. Many households coped throughrepeated migration, taking advantage of short-termopportunities for shelter, safety or income, and thenmoving on every six to 18 months.

In 1989, the study population was characterised by nearly100% home ownership. This changed dramatically duringthe conflict. People moved to new locations for months oryears. One-third of total households rented or leasedapartments or homes in their temporary location. Another

37% lived in housing schemes, including collective centresand refugee housing provided by aid organisations or localor foreign governments. A number of households (19% ofthe total study population) squatted in abandoned houses,most of which had been abandoned under distress by otherethnic groups.22 Rates of squatting were highest for theBosnian Serb residents of Krtova and Brezani, many of whomsought refuge in the Republic of Srpska, staying in housesvacated by Bosniacs who had fled into Federation territory.Local officials sanctioned the movement of displaced familiesinto abandoned houses, particularly in areas already clearedof minority groups. Table 4 shows the role of governmentand humanitarian shelter assistance in the conflict and post-conflict periods. Note that some households may havereceived both government and humanitarian assistance.

During the conflict, local authorities and governmentswere the main form of shelter assistance for households inthe study population. Rates of assistance received start at alow of roughly 25% of households from Krtova andSevarlije, reaching 66% of households from Jakes. Of this,32% of households in Jakes received this assistance fromthe German government.

When provided, government assistance was usually theprimary means through which households secured their

Table 3: Distribution of food aid by income category during conflict

Income category

Sufficient to cover expenses and to save money

Sufficient to cover expenses but not to save money

Insufficient to cover expenses; had to spend savings

Insufficient to cover expenses;received help from relatives

Insufficient to cover expenses; had to borrow

Insufficient to cover expenses; could not borrow

% of total HH in incomebracket during conflict (no. of HH)

12% (46)

15% (60)

2% (9)

4% (17)

7% (26)

59% (233)

% of total HH receivingfood aid, by incomebracket*

3%

9%

2%

4%

6%

76%

% of HH w/in incomebracket who receive foodaid** (no. of HH)

15% (7)

42% (25)

56% (5)

65% (11)

62% (16)

90% (209)

* Households per income bracket receiving food aid/total households receiving food aid.** Households per income bracket receiving food aid/total households per income bracket.

22 A 2004 World Bank report states that squatting remained a prevalent meansof securing housing for displaced persons in the country. ‘According to arecent estimate, almost 80% of the IDPs in Bosnia and Herzegovina remainillegally squatting in the housing of IDPs of another ethnicity or of some ofthe estimated 600,000 refugees still living outside the country.’ Holtzman andNezam, 72.

11


shelter needs in all villages. In contrast, households thatreceived humanitarian assistance during the conflict wereless likely to cite this assistance as their primary means ofsecuring shelter. Respondents that did report humanitarianassistance as their primary means of obtaining shelterduring the conflict usually resided in collective centres. Forinstance, households within the villages of Potocari,Sevarlije, and Jakes were the most likely to listhumanitarian assistance as their primary means ofobtaining shelter during the war. In these villages, 15%,8% and 5% of households, respectively, lived in collectivecentres (not shown in Table 4).

Table 4 shows that the relative roles of government andhumanitarian assistance were reversed in the post-warperiod. Government shelter assistance was received byonly 8% of households in the post-war period, while 57%of households received shelter assistance from humani-tarian agencies. Humanitarian shelter assistance alsoincreased in importance, with 36% of households relyingon it as the primary means through which they establishedtheir shelter in the post-war period.

A comparative analysis between the economic security ofvillages (based on the ability of households to coverexpenses) and the receipt of shelter assistance shows thatthere was little correlation between wealth and the receiptof shelter assistance during the conflict. In other words,unlike food aid, emergency shelter assistance does notappear to have been targeted based on economic need.Social capital proved important for finding shelter, asdisplaced households moved in with relatives or friendswho lived in more secure locations. Others were able totake advantage of their proximity to Croatia (such as manyfrom Prud and Jakes), and moved into collective centres inCroatia or went on to Germany, Austria or Switzerland.Although economic security assisted households in moving

to new locations, the correlation between economicinsecurity and shelter needs is less obvious than in the caseof food needs when households were most vulnerable.

A stronger link exists between economic security and shelterassistance in the post-war period. When analysed by village,it appears that the poorest villages contained the greatestpercentage of households receiving shelter assistance. Forexample, 82% of households in Brezani were unable to covertheir expenses in 2004 (making Brezani the poorest village),and 94% of households received shelter assistance from ahumanitarian organisation. This pattern holds roughly truefor the other five villages, with the population of Prud, thewealthiest village, reporting the least shelter assistance fromhumanitarian agencies (41% of households). An analysisbased on income bracket paints a slightly different picture, asshown in Table 5.23

Data by income bracket for 2004 indicate that humani-tarian assistance was provided relatively evenly byeconomic status. In other words, the assistance received byeach income bracket was directly proportional to thenumber of households within each income bracket. Thisfinding suggests that there may have been similardynamics driving income security and the receipt ofhumanitarian shelter assistance in the post-war period, butwe are unable to surmise what these dynamics might be.

2.1.4 Shelter assistance: questions of impactIt is difficult to gauge the impact of emergency shelterassistance provided during the conflict. Hypotheticaloutcomes in the absence of shelter assistance could includethe continued depletion of assets, the erosion of social23 Shelter assistance here refers to any assistance received to construct,

reconstruct or rehabilitate the dwelling inhabited by the household in 2004.In other words, if a household had rebuilt their home in 1996 with somehumanitarian assistance and were still living in this same home in 2004, thiswould qualify as shelter assistance in the 2004 portion of the survey.

Table 4: Households receiving shelter assistance and ranking of government or humanitarian assistance by village, conflict and 2004

Height of conflict 2004*

Government Humanitarian Government Humanitarian

Received Primary Received Primary Received Primary Received Primary any source any source any source any source

Total 38% 35% 12% 7% 8% 3% 57% 36%

Brezani 59% 59% 0% 0% 0% 0% 94% 88%

Jakes 66% 65% 16% 10% 1% 0% 63% 46%

Krtova 23% 18% 9% 0% 11% 0% 53% 34%

Potocari 38% 35% 21% 15% 17% 9% 59% 50%

Prud 30% 26% 2% 0% 4% 4% 41% 8%

Sevarlije 24% 23% 17% 13% 12% 0% 60% 35%

* Reflects assistance received to establish the current residence. This assistance was not necessarily received in 2004.

12


capital due to overcrowded accommodation with relatives,morbidity and mortality due to crowding or exposure tothe elements, a loss of dignity or repeated migration. Allhouseholds in the study population did manage to findsome form of shelter, but its quality ranged widely. Thesurvey did not provide enough data to draw conclusionsabout the impact of shelter assistance during the conflict.

The impact of shelter assistance in the post-war periodappears more tangible, with over one-third relying onhumanitarian assistance for their shelter. Further,qualitative data shows that shelter assistance made ‘returnhome’ possible for many households. Of course, therewere also many other households that returned to theirvillages even in the absence of assistance. Post-warreconstruction in the six study villages has been extensive,and most of the villages received some funds, buildingmaterials and even labour from humanitarian agencies,although rebuilding in Potocari has proceeded at a slowerpace than in the other villages.

The data shows that the majority of households in thestudy population received food and shelter assistance, andthis assistance was often integral to livelihood strategies.Such assistance was unable to counter the lasting impactsof both the war and socio-economic transition in Bosnia.In 2004, many households were still experiencing extremehardship. Although there has been much improvement inthe economic status of households in the ten years sincethe end of the conflict, many households still feel both therepercussions of a brutal war that dramatically altered thenature of Bosnian society and the continuing effects of thetransition from a socialist to a market economy. Poverty isa major concern, compounded by the high rates ofunemployment, ill-health of household members and lack

of steady income or job security. In many instances, peopleare still struggling to cover basic necessities, includingfood, shelter and medicine. In Jakes and Prud, poverty hasincreased since the end of the war.

2.1.5 Remarkable impact: protectionPerhaps the most interesting finding arising out of thesurvey data was somewhat unexpected. Food and shelterwere important interventions for households, may haveprevented further destitution and allowed families toreturn to their pre-war villages. However, protection wasthe form of assistance that had the greatest impact onlivelihood systems in both the short and longer term.Populations that were able to find safe refuge from theconflict experienced not only better physical security, butalso much better economic and food security than thosewho remained in close proximity to violence. Thesedifferences extended well into the post-conflict period,and were still apparent in 2004.

The importance of protection was readily apparent in thecollection of qualitative data. The relevance of protectionwas tested by conducting a statistical analysis of varianceon all data by household.This showed that by far the mostsignificant variables in determining the experiences,options and outcomes for households were time period andlocation. ‘Time period’ is a proxy for whether the householdwas living under war or peace, and ‘location’ is a proxy forwhether a household had a degree of protection from theconflict and was able to access assistance and economicopportunities: i.e., was able to relocate away from theconflict to a safe place.

The role of protection can be illustrated by dividing thestudy villages into two groups. The populations of Prud

Table 5: Percentage of households receiving shelter assistance by income bracket, 2004

Income bracket

Sufficient to cover expenses and to save money

Sufficient to cover expenses but not to save money

Insufficient to cover expenses; had to spend savings

Insufficient to cover expenses;received help from relatives

Insufficient to cover expenses; had to borrow

Insufficient to cover expenses; could not borrow

2004 population distribution

5%

25%

3%

19%

25%

24%

Received humanitarian shelter assistance

4%

26%

4%

18%

26%

23%

13


and Jakes had the greatest degree of protection during thewar, as many (but not all) of the households from thesevillages were able to take refuge in Croatia. Many thenmoved on to Western European nations, where they weregranted temporary protected status (TPS).The populationsof Sevarlije and Potocari are at the other end of thespectrum. Households from these villages had extremelylimited protection, access to services or options for paidemployment during the war.24 Residents of Sevarlije fledafter Bosnian Serb forces attacked the village in thesummer of 1992, but many households sought refuge intowns which subsequently experienced heavy fighting andextended siege. Potocari was part of the UN-designated‘safe area’ of Srebrenica, but remained under intense siegefor much of the war, and was eventually overrun by Serbforces and paramilitaries in July 1995.This dramatic failureof protection had both immediate and long-term impactson households in the study population.

Households from the first category of villages – Prud andJakes – were overwhelmingly better off during the conflictthan the populations of any of the other villages, though thedifferences are particularly extreme when compared toSevarlije and Potocari. These differences are the mostapparent in relation to economic security. Prud and Jakeshad the lowest rates of households unable to cover expenses(45% and 34% respectively) during the war, in comparisonto 98% of households in Potocari and 97% in Sevarlije.Thispattern continues across all indicators relating to economicsecurity: in the entire study population, the householdsfrom Prud and Jakes were by far the least likely to cut backon food consumption, educational expenses for theirchildren, the purchase of clothes and shoes and medicalcare. In contrast, nearly every household from Sevarlije andPotocari reduced these expenditures.

This pattern extended to access to basic utilities during theconflict. Ninety-five percent of households from Jakes hadaccess to sufficient water and electricity during the war.Households from Prud also had very high rates of access towater (85%) and electricity (81%). In contrast, only 13%of households from Sevarlije had sufficient electricity, and28% had access to water during the war. Potocari’shouseholds had access to both electricity and water at arate of 48%. Health status does not appear to have beenaffected by location and the protection provided.25

The significance of location across these indicators isinfluenced by several underlying factors. Most importantly,

households that were able to seek refuge in Croatia orother countries enjoyed much greater physical securitythan households that remained in Bosnia, and with it,protection from the erosion of their material, financial andsocial asset bases. Furthermore, economic opportunitieswere much greater outside of Bosnia, and thus thehouseholds that sought refuge outside the country weremore likely to be able to work for income. Thesehouseholds also had many or all of their expenses coveredin collective centres or refugee housing schemes. Incontrast, income-generating activities were rare in Bosnia.For many respondents, the constant threats of attack madefarming, market activity or travel during daylight hours aperilous undertaking. These external factors meant thathouseholds residing outside of Bosnia during the war wereable to focus their livelihood strategies on saving moneyfor the future, maintaining family networks or educatingtheir children. In contrast, households that remainedwithin Bosnia focused their livelihood strategies, for themost part, on staying alive.

There continues to be an important relationship betweenlocation and economic security in the post-war period, butthis relationship is not as significant as it was during theconflict. Prud remains much better off than the othervillages because the largely Bosnian Croat populationcontinues to take advantage of their ties to Croatia to seekemployment outside of Bosnia. Overall, however, thereturn of households to rural villages within Bosnia hashad a levelling effect on indicators of economic security.The poorest villages have experienced economic growth,while some wealthier villages have become poorer. Forinstance, the populations of Prud and Jakes wereconsistently worse off in 2004 than during the conflict innearly every indicator related to economic securitydiscussed above. Exceptions to this occur in two instancesin Prud, where fewer households reported cutting back onfood intake and health care expenses in 2004 than duringthe conflict. The trend has been particularly pronouncedfor Jakes, which was comparable to Sevarlije across mostindicators in 2004. At the other end of the spectrum,households from Sevarlije and Potocari have seenimprovements across the board. In 2004, the village ofBrezani fared the worst on most indicators due, in largepart, to its limited economic opportunities and isolation inthe mountains high above Srebrenica.

2.1.6 The nature of protection The nature of protection in complex political emergenciesis extremely varied, and this was apparent in the Bosniastudy. In some instances, protection was effective, in othersit was not. The consequences varied accordingly.Households from the study population had access to arange of different types and degrees of protection,including protection in the form of shelter offered byfriends and relatives, protection from war by crossing

24 The populations of the Bosnian Serb villages of Brezani and Krtova aresomewhere in between these two extremes. Many households from thesevillages fled to Serbia or today’s Republic of Srpska, where they were morelikely to be housed with relatives and experienced fewer direct effects of theconflict, though they still suffered economic difficulties.

25 For instance, households from Prud reported worse health status than thosefrom Potocari at the height of conflict, and households from Jakes reportedworse health status than those from Krtova.

14


political boundaries, evacuation,26 international mechanismssuch as temporary protected status, protection through thepresence of international peacekeepers and protective coverin the form of the threat of military force. Much of thepopulation undertook their own protection directly andjoined the armed forces. During the war, active duty inarmies and militias increased from five jobs in 1989 (or0.32% of the 1989 population) to 229 jobs (or 15.71% ofthe conflict population), including 24 women, seven menover 60 years of age and two male youths. During theconflict, fighting in the army or militia was the primaryoccupation for men in the study population in every villageexcept Prud. For many, fighting also provided the mostimportant source of income.

Refuge offered by friends and relatives was beneficial forhouseholds as long as the relocation sites remained free frominsecurity and attack. Thus households from Sevarlije thatfound accommodation with relatives in Bosnian towns thatlater came under siege did not fare as well as householdsfrom Krtova or Brezani who lived with relatives in Serbia orthe Republic of Srpska. All used social or family ties to findtemporary protection, but location was the determiningfactor in the quality of the protection they received.Access tosafe locations was influenced by a combination of factors,including timing (whether or not families decided to fleebefore being attacked) and international policies regardingshort- and long-term asylum.

The protection offered by political borders also varied bylocation. Many of those in the study population who fled toCroatia were able to find accommodation in collectivecentres, and were removed from the day-to-day violence ofthe war. Those who moved to Serbia also experienced littleviolence, but had less access to assistance or accommodationprovided by international humanitarian agencies. Economicopportunities were limited in both Croatia and Serbia, and

many who lived in Croatia complained that they wereunable to generate income or even pass the time in thecrowded collective centres. In contrast, those grantedtemporary protected status in Western European countriesoften found work (both legal and informal) and were ableto save money for their eventual return to Bosnia.

The international community also failed to protect Bosnia’scivilians in more overt ways, the most notorious and well-documented being the fall of the safe haven of Srebrenicato Bosnian Serb forces in July 1995. Qualitative data fromhouseholds in Potocari (part of Srebrenica municipality)provide a particularly vivid example of the impact ofprotection – or the lack thereof – on the lives andlivelihoods of conflict-affected populations in both theshort and long term. In the case of Srebrenica, protectionwas one aspect of the assistance provided to civilianhouseholds, and people incorporated this ‘protection’ intotheir livelihood strategies just as they might have includedfood aid or emergency medical assistance. Before and afterthe UN safe area designation, more than 20,000 peoplesought refuge in Srebrenica from towns and villages acrosseastern Bosnia that had already fallen to Bosnian Serbforces. Although living conditions were extremely poorand Serb forces prevented aid convoys from reaching thebeleaguered population, many residents of Srebrenicabelieved that the international community would providethe promised protection. Surveyed households in Potocari,unable to turn to migration, adapted their livelihoodstrategies to cope with the insecurity.They farmed at nightto avoid sniper fire and grenade attacks and, whenpossible, went to the market to barter goods such as soapfrom airdropped humanitarian parcels, or socks madefrom yarn acquired by unravelling sweaters provided inhumanitarian clothing distributions.

The collapse of the protection regime in July 1995 broughtan abrupt end to any attempt at normal life in Srebrenica.The population fled as Bosnian Serb troops advanced, andUN and NATO protection failed to materialise. Potocari wasthe village with the largest number of individual deathswithin the study population, with 18% of all householdmembers dying during the conflict, mostly directly killed inthe war. These losses constitute a small portion of theestimated 8,500 men and boys massacred as they tried toreach government-held territory.

26 No member of the study population reported that they had been evacuated,but evacuations did occur from contested or front line areas. Theseevacuations were often arranged by the local Red Cross societies.While manyof these evacuations did increase security for those involved, questions havebeen raised about the motives of the national Red Cross societies and theethics of helping to contribute to ethnically pure territory (Fawcett andTanner, 1999). UNHCR also helped to organise evacuations, at times workingwith ICRC. UNHCR was aware of the inherent moral dilemma of evacuations.Mooney (1995) quotes the High Commissioner as saying: ‘If we take them[for evacuations], we are accomplices to ethnic cleansing. If we don’t takethem, we are accomplices to murder’.

15


3.1 Dependency: fact or fiction?

The link between humanitarian assistance and a‘dependency mentality’ has been raised in conjunction withlong-running assistance programmes, and particularly inreference to refugee and displacement camps (Harrell-Bond1986; ICVA/UNHCR 1985). Several authors have called thispresumption into question, including Kibreab (1993), whoargues that a range of factors create the impression ofrefugee dependency. These factors are mostly based on theimpressions of external actors who – due largely to theirown biases – interpret behaviour in the camps as lethargy orlaziness. In fact, argues Kibreab, displaced people may wishto work, but are unable to do so because of unfavourableeconomic, social and environmental conditions.Furthermore, it is incorrect to assume that basic needs (asprovided through humanitarian assistance) are equal to thebasic desires of a population. This means that householdswill often seek to increase their income or commoditieseven after basic needs have been met through freedistributions. By assuming ‘dependency’, assistanceproviders assume that the coping systems, traditions andsocial structures of a displaced population have disappeared,when in fact these systems and structures remain central tothe ways that communities or households adapt followingtheir displacement.

Hoddinott (2003) questions the link between food aid anddependency in a study on the effects of food aid onEthiopian households. Hoddinott generated two sets ofresults on household and intra-household behaviour andthe receipt of food aid.The first model included weak (i.e.,no) controls for possible alternative causal factors, andfound that households that received food aid spent lesstime working on their own crops or pursuing non-agrarian occupations. On average, the households thatreceived food aid worked five times less (measured indays) than those households that did not receive foodassistance. In the second model, Hoddindott includedcontrols for household characteristics such as age, sex andeducation of the household head; household size; location;and extent of landholding. Under this more rigorousmodel, all negative effects of food aid disappeared (exceptfor a trivial reduction in the likelihood of growing ensetand a small reduction in agrarian labour, which was offsetby an increased amount of non-farm labour). Thesefindings contradict the presumed link between food aidand dependency in Ethiopia, and lead Hoddinott tosuggest that ‘The anecdotes and case studies upon whichdependency and disincentive claims are frequently builtare methodologically flawed’.

One could argue that it is nearly impossible for people tobecome dependent on food aid in emergencies becausethe food provided usually covers only a small portion ofhousehold needs. During the Darfur famine of 1984–85,for instance, food aid was estimated to meet only 10% offood consumption needs (de Waal 1987). De Waal (1994)describes a similar pattern in other famine-stricken areas aswell. Regardless of such evidence, the fear of creatingdependency continues to influence the planning of large-scale food distribution programmes. In Taliban-controlledAfghanistan, for instance, aid agencies shied away fromlonger-term livelihood, education or capacity-buildingprogrammes out of concern that this would providelegitimacy to the Taliban regime. Instead, programmesfocused on ‘short-term, life-saving emergency assistanceagainst the better judgment of agencies who knew that thisprogramming would create dependence, undermine localcoping mechanisms, and draw rural populations fromtheir villages to displacement camps’ (HPCR 2002).

There is little evidence, however, that food aid toAfghanistan established dependency, at least in ruralareas. Some people did migrate to camps, but this wasusually due either to insecurity or to severe watershortages after the onset of the drought in 1999. Focusgroup interviews in early 2002 indicated that many ofthe roughly 1,400 households represented had receivedsome food aid, but ‘the distributions were too limitedand too infrequent to reverse the multi-year deteriorationin food security’ (Lautze et al. 2002). In other words,they were low-impact and created no signs ofdependency. Countrywide data collected by WFP in ruralAfghanistan in 2003 indicates that households do notprioritise any form of food distribution or food-for-workinterventions, even in food-insecure areas (FIFC 2004).These findings are inconsistent with a population that isor has been ‘dependent’ on food aid. Households that doreceive food aid are likely to incorporate this assistanceinto their livelihood, coping and survival strategies, butthere is little evidence that the receipt of humanitarianassistance brings the collapse of motivation and initiativeat the household level.

One of the common misconceptions underpinning thenotion of dependency is that households and communitiesabandon their existing livelihood strategies and copingsystems when humanitarian assistance is provided over along period. A study of long-standing refugee camps inSomalia shows that, while communities evolve and adaptto changed circumstances, including the receipt ofhumanitarian aid, this adaptation does not entail

Chapter 3Discussion of findings

16


becoming dependent. In the Bosnian case study, all of thehouseholds had experienced abrupt and radical changes,and many were living in collective centres or refugeehousing, and provided with food aid and other forms ofassistance for several years. The ability of households topursue regular income-generating employment, to growfood or to purchase food decreased dramatically.Regardless of these circumstances, households maintainedmany of their traditional social networks and livelihoodskills and adapted, in one way or another, to the new set ofcircumstances. For instance, households continued to relyon family and social networks for shelter and remittancesystems following displacement and relocation. And whilethe availability of per capita occupations decreased duringthe conflict and in the post-war period, householdmembers adapted their livelihoods accordingly, with asharp rise in military service during the conflict, and anincrease in seasonal and occasional jobs in both the warand the post-war period.

There is little evidence of humanitarian assistancecontributing to dependency among the study populationin Bosnia. Food aid and shelter assistance were providedwidely and to those in need, and some householdsremained in collective centres and almost entirely relianton food aid for several years. Based on data fromqualitative interviews, households sought to leave thecollective centres and return to a self-sufficient lifestyle assoon as they were able, regardless of the food, shelter andhealth care provided in the centres.27 Some respondentswho found refuge and employment in Germany weremore reluctant to return home, but they attributed this towork opportunities, not to hand-outs provided to thosewith temporary protected status.

3.2 The importance of protection: implications

The overwhelming conclusion from the Bosnia study datais that the location of households and the nature of theprotection available in these locations were the mostimportant factors in determining overall wellbeing. Asdiscussed, there were many protection failures in Bosnia.These were partially due to the international community’sfailure to make good on its promises of military protectionfor the safe areas. More deep-rooted problems alsocontributed to the weakness of the protection regime.Thereluctance of the international community to push forpolitical solutions to the conflict was one of the mostsubstantial obstacles to ending the war or enforcingeffective protection mechanisms. Some critics felt thatWestern nations sought to use humanitarian assistance andthe presence of UN and UNPROFOR personnel to

compensate for the lack of political or militaryintervention to end the violence. For instance, describingthe role of UNHCR during the first two years of theconflict, Mooney (1995) says: ‘[I]n the absence of apolitical resolution of a conflict, humanitarian assistanceand international presence cannot by themselves provideeffective protection to victims’.

The erosion of third-country protection mechanisms forcivilians affected by conflict also prevented Bosnians fromaccessing secure locations. The study population containshouseholds which were able to reach Croatia, Serbia or thirdcountries, and to find accommodation in these locations.Many other households, however, were unable to reachthese places of relative safety and spent much of the warunder constant threat of attack.The differences experiencedby these two population groups point to the importance ofthird-country residential arrangements for asylum-seekersand refugees, for the extension and expansion of temporaryprotected status for civilians from nations affected byconflict, and for the establishment of well-run centres forrefugees or displaced persons if and when necessary andappropriate. In practice, however, countries close to states incrisis or turmoil are restricting border access, tighteningasylum regulations and creating obstacles for civiliansseeking temporary refuge. These trends will make it moredifficult for civilians to obtain the protection that, in the caseof the survey study population in Bosnia, proved essential toboth survival and post-war recovery.

The evidence from the Bosnia data on the importance ofprotection for the livelihood strategies and overallwellbeing of households has a variety of implications fornational governments and the international community. Inthe context of Bosnia, those households that were able tosecure protection were also more likely to be able to accesspaid employment, medical care and an adequate supply offood. The majority of these households would not havebeen able to remain in these protective environmentswithout the accommodation and assistance of collectivecentres and refugee housing. Much evidence, however, hasshown that camps or centres are not the ideal residentialmodels for displaced persons in many circumstances(Chambers 1985; Hansen 1982; Harrell-Bond et al. 1992;Van Damme 1995; Voutira and Harrell-Bond 1995). Inparticular, establishing collective centres or camps throughforce on the part of civilian or military authorities is likelyto result in the deterioration of physical, economic andhuman security (CSOPNU 2004; Macrae and Zwi 1994).

On the other hand, if and when civilian populations havealready been displaced, are fleeing extreme violence and arein need of shelter and humanitarian assistance, then therelevance and importance of protection systems such ascamps and centres become more apparent. Such protectionsystems can only improve the status of populations if inputs

27 All the households in the study population were returnees, and most weredriven by a strong desire to return to their home villages.This resulted in biasin the study design, as we did not interview any households who chose toremain in collective centres or similar accommodation.

such as food, medical care and physical protection areprovided impartially and appropriately, and in a transparentfashion. In Bosnia, for instance, people who had already fledto Croatia and were able to reach collective centres whichprovided shelter, food, schooling for children and medicalcare were much better off than those who could not. TheBosnia study illustrates the potentially positive role of theseresidential protection models. The larger message, however,is of the overall importance of protection and the imperativeof national governments and the international governmentsto ensure the success of protection regimes.

3.3 Discussion of the model

The concept of using a livelihood analysis to understandhousehold need and the impact of interventions is notoverly innovative, as illustrated in the models of Save theChildren (UK), CARE, and Oxfam-GB. These models havedistinct programmatic goals, such as assessing foodsecurity, and all questions and analysis are ultimatelyfocused on this one aspect. In contrast, the Bosnia studyused a livelihood analysis that sought to understand thebroader livelihood and coping systems of households.Ideally, such an analysis will illustrate the relative role ofhumanitarian assistance in these household systems. If ahousehold does not consider outside interventions to berelevant or important, they will not be picked up by alivelihood analysis, even if such interventions existed.

3.3.1 Positive aspectsBy covering three time periods, the survey demonstratedhow adaptations in livelihoods systems occur over time andin response to shifts both within the household (such aschanges in demographics, employment or health status) andin the external environment (such as armed conflict,economic transformation and forced migration). Livelihoodsystems were not assumed to be either static or gearedtowards a single goal, such as acquiring food. Rather,respondents ranked the priorities of their household in eachtime period (such as education, health, security, shelter,income, food or ‘other’). The livelihood analysis sought tounderstand both how these priorities altered over time, andhow various households adapted in order to meet theirchanging goals. Because livelihood analysis is based onhouseholds’ own articulation of their primary livelihoodoutcomes (e.g. food security, economic security, physicalsecurity), it is appropriate for use in a wide variety of settingsand contexts, such as vulnerability assessments, programmedesign and evaluation in times of famine, natural disasters,armed conflict, and post-conflict situations. In this study,qualitative and quantitative data collection was essential inorder to gather enough information to understand trendsand patterns. Follow-up focus groups were not possible inthe study due to resource constraints, but would have been auseful means of exploring issues in greater depth, ideallyafter an initial data analysis.

The survey tool used in Bosnia allowed for field work to becompleted in a relatively short period of time. The FIFCand Mercy Corps team members spent fewer than threeweeks administering the survey.The involvement of MercyCorps Bosnia greatly expedited the field work, as teammembers assisted in the study design and field testing, theidentification and training of surveyors, the selection ofstudy sites and logistics. They were extremely well versedin the context of the war and post-conflict settings.

Although no substitute for anthropological field work, theuse of a livelihood analysis to learn about the impact ofhumanitarian assistance has the potential to deepenknowledge about the role of humanitarian assistancewithin households managing risk and vulnerability indeeply challenging environments. Respondents were askedabout the role of humanitarian assistance, but only as oneof many possible sources of inputs such as food, shelter,and health care. The ranking by importance of thesesources showed the relative role of external humanitarianassistance for each household.

The livelihood analysis examined 15 years of tumultuouschange. The survey itself was conducted nearly ten yearsafter the end of the war, the period in which thehouseholds experienced the greatest upheaval and an eventwhich still influences all aspects of Bosnian society today.This time-lag likely affected the quality of information onemergency humanitarian assistance. Perhaps keyinterventions have been forgotten, or perhaps people feltmore at liberty to be honest about the relative role ofhumanitarian assistance in their overall livelihoods strategythan might have been the case when humanitarianagencies were conducting assessments in the midst ofcrisis. The longer timeframe allowed for other importantaspects of assistance and intervention to be uncovered.First, many households received shelter assistance yearsafter the war; in many cases, it was this reconstructionassistance which enabled them to fulfill their desiredlivelihood outcome and return to their villages of origin.Second, the extended timeframe provided enoughqualitative data to understand the longer-term effects ofthe war and post-war recovery. We were able to see thestark differences between communities during theconflict, and the levelling out of economic security in thepost-war period. Third, and most importantly, thecomparison between three distinct time periods and thewide variations in conditions made the importance oflocation, as a proxy for protection, apparent.

3.3.2 Challenges and possible modifications A livelihood analysis allows us to understand the relativerole of humanitarian assistance within a larger livelihoodframework, but it does not provide a great deal of specificdetails on the nature of this assistance.We do not know, forinstance, about the quality or quantity of assistance

17


18


received, where the assistance came from, whether therewere concerns of bias in how the assistance wasdistributed, or what household views were on theseimportant issues. The precise impact of the assistance isdifficult to measure because the input – assistance – wasnever discussed in isolation.We felt that households do notconsider humanitarian assistance in isolation, and ratherview it as one of a range of possible resources andstrategies that might, somehow, be used to better managerisk and vulnerability.

The design of the survey led to a degree of bias in thefindings in several ways. First, we only interviewedhouseholds that had returned to their pre-war villages,leading to bias in the degree of representation ofhouseholds during the first two time periods (1989 andheight of conflict). Not having a control group comprisedof people who have not opted to return to their pre-warhomes means that we are not fully aware of the factors thatmay have contributed to a decision to return or not,including the role of humanitarian assistance. A secondsource of bias regards the age of respondents. Householdscomprising elderly adults in 1989 and the height ofconflict, whose members did not survive to 2004, wereobviously not included in the survey in 2004. This led toan over-representation of healthier elderly adults, i.e., onlythose healthy, strong or fortunate enough to survive to2004.Third, the study sites were based on the suggestionsof Mercy Corps, our research partner in Bosnia. While thesix villages are different in terms of wealth, ethnicity andexperience during the conflict, they are all also villagesw

july 2005 hpg humanitarian policy group background paper · 2016. 5. 4. · between the political...

Documents