Summary Evaluation Report
Acknowledgements:
We wish to express our genuine gratitude to the management and staff members of UNHCR/Burundi in Bujumbura and sub-offices in Muyinga, Ruyigi and Makamba as well as to our interlocutors at other UN, bilateral donor, and NGO offices for their effective support of this evaluation at all of its stages; their expert insight into the concepts and operations of the refugee assistance; and first and foremost, for their untiring effort of helping the dream of return come true for hundreds of thousands of Burundians. We would also like to thank the Government of Burundi officials at the national, province, and commune levels for generously sharing their time in meetings with us, and for their candid opinions about the present and future of their country. Ultimately, little could be understood about the return and reintegration without meeting and speaking with the repatriates and their neighbors. We thank them for telling their stories, and sharing their worries and hopes for a better life in peaceful Burundi.
While accuracy and clarity were the authors’ primary concern in presenting the evaluation findings, we take full and sole responsibility for any error found in this report.
IMPACT EVALUATION
OF PRM HUMANITARIAN ASSISTANCE
TO THE REPATRIATION AND REINTEGRATION OF BURUNDI REFUGEES
(2003-08)
Summary Evaluation Report
March 2009
The Zero-Sum Game:
Analyses, Findings, and Conclusions of the Evaluation Study
Terra P Group, Inc.
Bujumbura, Burundi – Gaithersburg, MD
Prime Contract #: | S-AQMMA-08-C-0182 | ||
Contract Title: | Impact Evaluation of PRM Humanitarian Assistance to the Repatriation and Reintegration of Burundi Refugees | ||
Sponsoring unit(s): | Bureau of Population, Refugees and Migration (PRM): Office of Assistance for Africa | ||
Contractor name: | Terra P Group, Inc. | ||
Authors and contributors: | Alexander Telyukov, Ph.D. and Mary Paterson, Ph.D., M.S.N., with contributions from: ‘Community Leadership Center’; Joseph Ndayegamiye, Ph.D., Joseph Bigirumwami, Ph.D. (all - Bujumbura, Burundi), Irina A. Telyukova, Ph.D., Department of Economics, UC San Diego, La Jolla, CA; Ludo Visschers, Ph.D., Department of Economics, Simon Fraser University, Vancouver, BC, Canada | ||
Recommended citation: | Telyukov, Alexander; Paterson, Mary. Impact Evaluation of PRM Humanitarian Assistance to the Repatriation and Reintegration of Burundi Refugees. Evaluation Report, Vol.1: The Zero-sum Game: Analyses, Findings and Conclusions of the Evaluation Study. Gaithersburg, MD: Terra P Group, Inc., 2008. x + 282 pp. | ||
Language: | English | Release date: | February 15th, 2009 |
Abstract: | The first volume of the evaluation report presents analyses, findings and conclusions, developed in a six-month study of humanitarian assistance that was provided to Burundian refugees during their return to, and reintegration in their home country. The evaluation had a twofold institutional focus: (i) multi-sectoral programs of refugee assistance, led by UNHCR; and (ii) assistance provided under bilateral grants, financed by the Bureau of Population, Refugee and Migration (PRM) of the U.S. Department of State. The study relied on pre-existing and new evidence from multiple sources, including the UNHCR Statistical Database, UNHCR and PRM program documentation; a household survey, a survey of water sources, stakeholder interviews, and visits to 17 assistance sites in three provinces of high return. Analyses were based mostly on concepts and methodologies developed for this evaluation. The reported study answered five evaluation questions: Did Burundi refugees return? Did refugees return in safety and dignity? Did returnees achieve reintegration? Can it be attributed to humanitarian assistance? and, Will return be durable? – Four rounds of analysis produced and validated the main conclusion: Near-complete repatriation and reintegration are facts of life and were achieved with an indispensable contribution from the program of assisted return. Risks to sustainable reintegration are discussed and recommendations provided on how to control some of them. | ||
Keywords: | Burundi, refugees, humanitarian assistance, repatriation, reintegration, Government of Burundi; UN agencies, donor agencies, implementing partners; Bureau of Population, Refugees and Migration (PRM); UNHCR, UNICEF, PARESI, IRC, World Relief, shelter assistance, food assistance, non-food assistance, housing, land-related conflict, protection, water and sanitation, environmental health, education, peace villages; relief and development, sustainability, durable return. | ||
Contact information: | A. Telyukov, M. Paterson; tel: (+1)240-683-6823; e-mail: partners@terra-p-group.net | ||
List of Acronyms
ACF | Action Contre la Faim [Action Against Hunger] |
AHA | African Humanitarian Action |
ALRIs | Acute lower respiratory infections |
BCG | Bacille Calmette-Guérin [vaccine against tuberculosis] |
BFr | Burundian Franc |
CA | Cooperative Agreement |
CADI | Centre pour l'Autosuffisance et le Développement Intégré [Center for Self-Reliance and Integrated Development – Burundian NGO] |
CAP | Consolidated Appeals Process |
CFR | Code of Federal Regulations (U.S.) |
CNTB | Commission Nationale des Terres et Autres Biens [National Commission for Land and Other Assets] |
CO(O)PEC | Coopérative d’Epargne et de Crédit [Loans and Savings Cooperative, Burundi] |
CORD | Christian Outreach Relief and Development |
DTP | Diphtheria, tetanus and pertussis vaccine |
EC | The European Commission |
ECHO | European Commission’s Humanitarian Office |
EMRA | Emergency Refugee and Migration Assistance |
EU | European Union |
FTS | (OCHA) The Financial Tracking Service |
FY | Fiscal Year |
GBV | Gender-based violence |
GoB | Government of Burundi |
GTZ | Gesellschaft für Technische Zusammenarbeit [Society (Company) for Technical Cooperation, Germany] |
HC(s) | Health Center(s) |
Hep B | Hepatitis B vaccine |
Hib | Haemophilus influenzae type vaccine |
HIV/AIDS | Human immunodeficiency virus / Acquired immunodeficiency syndrome |
IASC | The Inter-Agency Standing Committee |
ID | Identification (National ID Card) |
IMC | International Medical Corps |
IMF | International Monetary Fund |
IRC | International Rescue Committee |
ISTEEBU | l’Institut de Statistiques et d’Etudes Economiques du Burundi [Institute of Statistics and Economic Research, Burundi] |
JAM | Joint Assessment Mission |
Kcal | Kilocalories |
M | Million |
MRA | Migration and Refugee Assistance |
N | Number of respondents (in a household survey) |
na | Not applicable |
NGO | Non-governmental organization |
NIDC | National Identification Card |
NTU | Nephelometric Turbidity Units |
OCHA | United Nations Office for the Coordination of Humanitarian Affairs |
OPV | Oral polio vaccine |
PARESI | [the Government of Burundi Project for the Reintegration of War-affected Persons] |
PRGF | Poverty Reduction and Growth Facility |
PRM | Bureau of Population, Refugees and Migration of the U.S. Department of State |
Q | Quarter |
RC(s) | Reception Committee(s) |
RCE | Régie communale de l’eau [Commune Water Administration, Burundi] |
R&R | Return (Repatriation) and Reintegration |
TB DOTS | Directly Observed Treatment, Short-course (WHO -recommended strategy of treating tuberculosis) |
UN | United Nations |
UNDP | United Nations Development Programme |
| UNHCR | United Nations High Commissioner for Refugees |
UNICEF | The United Nations Children’s Fund |
USD | US Dollars |
USG | U.S. Government |
VIF | Variance Inflation Factor |
VRF | Voluntary Return Form |
Watsan | Water and Sanitation [as assistance sector] |
WFP | United Nations World Food Programme |
WHO | World Health Organization |
Table of Contents
Executive Summary.......................................................................................................... 7
1. Evaluation Agenda...................................................................................................... 9
2. Evaluation Program and Implementation.................................................. 10
3. Technical Findings and Conclusions............................................................... 12
3.1 Focus Areas of PRM-funded Assistance and Expected Program Outcomes...... 12
3.2 Achievement of Repatriation..................................................................................... 14
3.3 Quality of Repatriation............................................................................................. 15
3.3.1 Achievement of Voluntary Return............................................................................... 153.3.2 Assurance of Safe Return............................................................................................ 17
3.3.3 Protection of Returnees’ Legal and Economic Interests.............................................. 19
3.4 Achievement of Reintegration.................................................................................. 19
3.4.1 Integrated Assessment of Reintegration...................................................................... 203.4.2 Assessment of Reintegration by Welfare Component................................................... 21
3.4.3 Assessment of Reintegration as Perceived by Returnees............................................. 24
3.5 Attribution of Reintegration to Assistance.......................................................... 25
3.5.1 Food Assistance.......................................................................................................... 253.5.2 Cash Support.............................................................................................................. 26
3.5.3 Land Conflict Management........................................................................................ 26
3.5.4 Housing Construction................................................................................................. 27
3.5.5 Water and Sanitation.................................................................................................. 29
3.5.6 Education................................................................................................................... 32
3.5.7 Health......................................................................................................................... 33
3.5.8 Protection of Safety and Human Rights...................................................................... 34
3.6. Prospects for Durable Return................................................................................. 34
4. Summary Conclusions............................................................................................... 37
5. Recommendations....................................................................................................... 3
Executive Summary
This reported evaluation of the UNHCR and PRM-funded programs of assisted repatriation to Burundi (2003-8) was conducted in June – November 2008. The evaluation focused on program outcomes and impact. Pre-existing and new evidence was analyzed to answer the following five questions: (i) Did Burundian refugees return? (ii) Did they return in safety and dignity? (iii) Did returnees achieve reintegration? (iv) Was reintegration assistance provided and if yes, did it help reintegration? (v) Will return be durable?
a. Study Design, and Implementation:
A set of methodologies, designed for the evaluation, included as follows: (i) A reintegration measurement framework, based on a composite welfare index; (ii) Two cross-sectional statistical models to estimate welfare variance between the repatriate and incumbent (non-repatriate) families; and among the repatriates only; (iii) A household survey of the repatriate and the incumbent households, based on a 3-stage stratified random sample; (iv) A water testing survey and a plan for visual assessment of watsan facilities; (iv) Health and school survey questionnaire and visual assessment plan; (v) A focus group guide to assess community-based organizations; (vi) Assistance site selection plan.
The evaluation’s field stage included 38 interviews and discussions in Bujumbura and 17 assistance sites in Muyinga, Ruyigi and Makamba provinces. Representatives of the following 5 stakeholder groups shared their opinions and information: (i) the Government of Burundi at the national, province and commune levels; (ii) UNHCR and other UN agencies; (iii) donor organizations, (iv) implementing partners of UNHCR and PRM, and (v) aid beneficiaries: returnees and staff of constructed/rehabilitated social services facilities. Primary field evidence about the assisted return was produced during visits to a boarder checkpoint in Tanzania, 4 transit centers, and a temporary lodging facility. Primary field evidence about reintegration assistance was produced from: (i) a household survey of 204 rugos in 19 collines of the three high-return provinces; (ii) professional testing of 27 water sources; (iii) interview-based surveys and visual assessment of 2 primary schools, 4 health centers, 2 gravity-fed water lines, 7 sanitation facilities, serving schools and health centers.
Prior to and following the field stage, these assessments were conducted: (i) a large-scale review of program documentation and statistics; (ii) a content analysis of transcribed interviews by arranging primary material in 33 subject nodes and subdividing it by 5 stakeholder groups; (iii) a case abstract analysis of 265 land-related disputes, settled out of court in Q2 2008; (iv) econometric model testing of household survey-based data; (v) analytic review of digital picture gallery created during the site visits and household survey in Burundi.
b. Select Findings and Conclusions:
Repatriation:
The benefits provided under the assisted return program in Burundi have ensured comprehensive physical, legal, and economic protection of Burundian refugees during their repatriation to the home country. These protections have been well designed, sensitively adjusted to the evolving need, and rigorously executed. The beginning of the assisted return in 2003-4 was afflicted with organizational problems, due, primarily, to the delayed deployment of UNHCR resources in the areas of high return. In 2005-8, the assisted return program attained its maturity and consistently delivered on expectations of the repatriates, the Government of Burundi, and UNHCR’s donors alike. If there is a succinct way to define the UNHCR-led program of assisted return to Burundi, it would be ‘evidence-based pragmatism’ – a set of conceptual and operational features that puts the program in the range of best-practice responses to refugee crises and humanitarian crises in general.Reintegration:
Reintegration was evaluated on a set of technical evidence (levels 1 to 3) and based on perception of its primary beneficiaries – the returnees (level 4).The technical side of reintegration has been examined at three levels: (i) econometric estimation of the difference among the repatriate and incumbent families by an integrated measure of socio-economic achievement – the welfare index; (ii) estimation of variance and similarity between the two population groups in each welfare component; (iii) estimation of the likelihood of reintegration based on the amount and reintegration effects of the post-return assistance. The econometric model produced an unambiguous finding that the repatriates and the incumbents are indistinguishable by welfare level at the present time. The ‘present time’ in the context of this analysis is defined by the average post-return experience of 4.2 years. If there was a welfare gap immediately after return, it disappeared by now -- 4.2 years after return. The second round of analysis examined the comparative standings of the repatriates and the incumbents by their current assets (consumption and savings), their durable physical assets (housing), their human assets (education and health), and their safety. The finding from this study corroborates the econometric output: the repatriates and the incumbents possess assets at very similar levels. There are factors, for example family size, that determine living standards of Burundian households at a statistically significant level. Ex-refugee status is not among such factors. The near-term purpose of refugee assistance – to put returnees on a par with the resident population of Burundi – has been achieved for an average repatriate family as of September 2008. The analysis of reintegration benefits in 7 sectors of assistance – food, cash, land, shelter, water and sanitation, education and health – leads to the conclusion that assistance played an important role in the achievement of reintegration. Reintegration assistance was designed, scaled, timed, and delivered in response to key material and social needs and effectively enough to make the repatriates themselves to acknowledge: ‘Yes, we benefited from our returnee status in Burundi’.
The importance of the material aspects of reintegration has been put in perspective by the repatriates themselves. They expressed strong satisfaction with their decision to return despite the fact that their current well-being is below pre-return expectations, let alone their pre-return and pre-displacement living standards. The reviewed evidence from the household survey suggests that the return and reintegration have been a self-motivated process in Burundi.
Sustainability:
Given the returnees’ deeply-felt desire to be back to their home country despite all the post-return hardship, can UNHCR and PRM conclude that their efforts are bound to succeed in Burundi, if minimally effective? – For the short term, the likely answer is ‘Yes’. Long-term, there is not enough evidence to conclude with confidence that the near-complete repatriation marks an irreversible end to forced displacement. There are factors that support sustainability yet have not been put to work at full strength. There are also negative factors that have not been entirely curbed. The main areas of concern include a latent discontent over uncompensated redistribution of land; a divisive role of aid when it is excessively targeted to repatriates beyond the first two-three years; a partial loss of momentum in the relief-to-development transition; and lack of confident projections of post-election policy trends.The land issue is central to the near- and long-term prospects for sustainable reintegration. The overcrowding on the currently farmed land, compounded by the legacy of mismanaged land ownership rights and lack of progress toward overdue reforms in the agrarian sector makes it difficult to endow repatriates with land – a source of complications at the current stage of the assisted return program and, potentially, a stumbling block on the road to sustainable reintegration.
c. Recommendations in Support of Sustainable Reintegration and Durable Return:
The land issue is the most important challenge to sustainable reintegration. A strategic response to this challenge requires a system of compensation, accessible to any of the two contending parties in land-related disputes. Compensation should be high enough to incentivize an opt-out solution to a dispute. Either claimant or defendant should be encouraged to relocate and/or take on a path to non-agricultural livelihood. Compensations, therefore, should be set at levels that would provide the recipient family with cash to buy land elsewhere, or with start-up capital for self-employment or business activity in the current or alternatively chosen place of residence. Public land distribution is a strategy that supplements or partially substitutes for cash compensations.
The following steps will help create a much-needed system of land-related compensations:
· Tapping into a major line of loans to Burundi provided by IMF under the PRGF facility. “PRGF-supported programs are designed to cover only areas within the primary responsibility of the IMF, unless a particular measure is judged to have a direct, critical macroeconomic impact” (IMF, 2007). Areas typically covered by the IMF include advising on prudent macroeconomic and financial policies and related structural reforms such as exchange rate and tax policy, fiscal management, budget execution, fiscal transparency, and tax and customs administration. It is important, therefore, to prepare a position paper on behalf of the donor community, explaining to IMF the macroeconomic drag that landless returnees or dispossessed incumbents will create in terms of GDP foregone, wage levels across the economy driven down, budget revenues curtailed, and budget outlays escalated.
· To establish compensations at economically viable levels, a system of land valuation is important. Its design should be integrated into the current work on the new land registry. Comparative valuation of land, either by tracking prices of previous transactions or directly appraising land productivity by combination of fertility and location will be important for future land banking, in addition to a system of fair compensations.
· Physical availability of land is needed for the system of compensations to function effectively. Further consolidation of land in a public inventory and its distribution on the principles of accountability and fairness should be supported by the current and future aid programs.Mainstreaming repatriation assistance is a way to reduce the divisive effect of international aid. Given that PRM’s funding vehicles are earmarked for refugee assistance, PRM cannot spend on the general population except by producing public goods in the water, sanitation, and construction/rehabilitation of social service facilities in the communities of high return. Even under the existing system of earmarks, geographic targeting can be refined. Initial simulations performed on data from the UNHCR/Burundi Statistical Database suggest a multi-variable targeting formula with the important new variable: ‘Average Post-return Experience in Months’ by commune. This variable must be used in conjunction with several others: ‘Number of unassisted returnees’, ‘Aid Delivery Lag’ (in months of program effort); ‘Presence of Other Donors’. The previously used variables of ‘Number of Returnees’ and ‘Percent of Returnees in the Local Population’ should remain but be statistically de-emphasized.
While expecting the development stage to take over in the next few years, PRM may want to consider a ‘maintenance grant’ to slow down or stop the normal process of deterioration of previous aid results. This grant can be used as a pilot for developing professional workforce and private business start-up in Burundi in the critically important areas of maintenance, construction, and community strengthening.
1. Evaluation Agenda
The following evaluation objectives were set out in the PRM Statement of Work (PRM, 2007b):
· Assess the impact of PRM’s assistance to Burundi refugees and returnees in the environmental health sector as provided by PRM’s NGO partners.
· Assess the impact of PRM’s assistance to Burundi returnees provided (financed) through UNHCR.
· Provide recommendations for PRM’s core indicators on reintegration assistance.
· Highlight lessons learned for future reintegration operations.Based on PRM’s objectives, the evaluation agenda comprised five main evaluation questions, each one followed by a set of supporting questions and ‘evaluation items’:
I. Did Burundi refugees return?
1.1 Repatriation of Burundian refugees:
1.1.1 Main statistics and structural characteristics of return
1.1.2 Prospects for completion of return.
II. Did refugees return in safety and dignity?
2.1 Was return voluntary?
2.1.1 The ‘push’ and ‘pull’ factors
2.1.2 Mechanisms and practices to ensure voluntary return
2.2 Was return safe?
2.2.1 Security en route and at transit centers
2.2.2 Health protection and health care support
2.2.3 Support of the vulnerable
2.3 Was return protective of the returnees’ civil status?
2.3.1Support with rapid legalization
2.4 Was return supportive of the economic interests of the repatriates?
2.4.1 Cargo allowances for possessions and livestock
2.4.2 Streamlined registration for post-return benefits
2.4.3 Temporary lodging near claimed land.
III. Did returnees achieve reintegration?
3.1 Comparative welfare of the returnee and incumbent population:
3.1.1 Welfare indices per household member in both population groups
3.1.2 Welfare variance among returnees and incumbents
3.1.3 Comparison of returnees and incumbents by type of household assets
3.2 Returnees’ opinions:
3.2.1 Do they regret the decision to return?
3.2.2 Did post-return experience match returnees’ expectations?
3.2.3 Did returnees restore their welfare to benchmark levels from the past?
IV. Was reintegration assistance provided and if yes, did it help reintegration?
4.1 Supply and sufficiency of assistance by sector:
4.1.1 Food
4.1.2 Cash
4.1.3 Land
4.1.4 Shelter
4.1.5 Water and sanitation
4.1.6 Health and education
4.1.7 Safety.
V. Will return be durable?
5.1 Survivability of aid results
5.2 Prospects for nationwide peace and stability
5.3 Prospects for community-level peace and stability
5.4 Prospects for government leadership in the development process
5.5 Sustainability as a matter of transition from relief to development.
As the title of this evaluation suggests, PRM has grouped expected program results in two categories: repatriation and reintegration. The evaluation agenda follows this categorization: the first two questions refer to the achievement of repatriation, while the other three refer to the results and sustainability of reintegration. Ultimately, the main point of PRM’s interest in this impact evaluation is twofold: Did return happen? Will it last? -- The evaluation team designed and implemented this study with an unflagging focus on both interrelated questions.
2. Evaluation Program and Implementation
The evaluation was implemented in June – November, 2008. A review of pre-existing evidence was followed by field surveys, interviews of key informants, on-site observations, and post-field discussion and analyses to answer the main question: Did UNHCR and bilateral partners deliver results on PRM funding? The evaluation program included the following activities:
1. A background study was conducted to define evaluation scope and technical content (Chapter 4 of Complete Report). It included the following activities:
· Literature and document review to cover humanitarian assistance research, stakeholder policy and project documentation, media reports, and statistical data.
· Design of logical and conceptual frameworks for the evaluation. Central to this study, program impact was defined as program outcome, sustained over time. Program outcome, in turn, was defined as a high-level, systemic program result. Repatriation and reintegration outcomes as well as prospects for their sustainability were assessed in conjunction with program inputs, processes and outputs.
· Review of aid funding in Burundi by program category, assistance area, donor, and implementing partner. Four program categories were identified, based on donor funding objectives, and one of the four was targeted in this evaluation. This focus program category was titled ‘Assistance to returnees in UNHCR and/or PRM program domain’. The total program budget in this tier of relief and development aid expenditures in Burundi was estimated at $112.2 M or 22.9 percent of the total in 2003-7.
· Stakeholder review to establish ‘who does what and where’. Based on the results, a representative sample of stakeholder organizations was selected for field contact and close-up review.2. Evaluation tools were designed (see report’s Chapter 5 for the tool design and Volume 3 for the tools themselves). The study toolkit included the following:
· A reintegration measurement framework, based on the concept of household welfare. The key aggregate measure is welfare index. It included four components: (i) household current assets: consumption and savings; (ii) household physical assets: land, housing (structural and location characteristics), livestock, fruit-bearing trees, and consumer durables; (iii) household human assets: health and education; and (iv) household safety.
· Two cross-sectional statistical models, one of them based on the following conceptual premise: The degree of reintegration can be ascertained by comparing the current welfare levels of the returnee and incumbent families. If they are found indistinguishable, there is strong reason to believe that reintegration was achieved. Otherwise, it was not, and assistance outcomes should be assessed in terms of reintegration progress rather than reintegration achievement.
· A household survey was designed to provide data for statistical modeling and for analysis outside the model. The selected sampling plan ensured a sufficiently powered sample that captured the diversity of socio-economic conditions in the three high-return provinces of Burundi. A stratified random sample included both the repatriates (intentionally over-sampled) and the incumbents (i.e., not refugees) living in communes of high, medium, and low return.
· A water testing survey and a plan of visual assessment of water and sanitation facilities.
· Health and School Survey Tools to ascertain the supply and quality of basic social services in the areas of return.3. Field assessment included:
· 38 interviews and discussions in Bujumbura and the three high-return provinces designated for the study: Muyinga in the North-East, Ruyigi in the East, and Makamba in the South of Burundi.
· A household survey of 204 households, including 100 returnee and 104 incumbent households. The Survey was conducted in 19 collines (villages), randomly selected from a stratified sample of communes that represent areas with high, medium, and low presence of returnees.
· A qualified expert in water testing conducted a survey of 27 water sources, including 25 standalone tap stands (protected springs) and 2 tap stands connected to water lines.
· A seven-day field trip on a circular route from Bujumbura across three provinces. The evaluation team leader and country coordinator visited 17 aid sites. They visually examined seven assisted return facilities (including two border checkpoints, four transit centers, and a temporary lodging facility); four health centers; two primary schools; and two gravity-fed water systems.
· Post-field discussions to debrief counterparts, clarify field information, and request additional data, based on needs identified during field observations.4. Post-field data management, analysis and reporting included the following activities:
· Transcription and content analysis of field interviews and discussions.
· Household survey data verification and cataloguing; computation of secondary variables.
· Econometric modeling and other micro-studies on available household data, statistics, stakeholder opinions, and visual observations.
· Preparation of the 3-volume project report with analyses, findings, conclusions and recommendations (vol.1), summary of field materials (vol.2), and evaluation tools (vol.3).3. Technical Findings and Conclusions
3.1 Focus Areas of PRM-funded Assistance and Expected Program Outcomes
The two-pronged multilateral/bilateral strategy of refugee assistance[1] has been consistently applied in Burundi: PRM allocated resources both to UNHCR and through directly managed grants to partner NGOs in the proportion of 85:15. This evaluation covered both lines of PRM funding. The main features of PRM’s multilateral and bilateral engagement in the repatriation and reintegration assistance in Burundi are summarized below:
a. Multi-lateral Funding and Program Results:
By committing support to UNHCR, PRM ensured U.S. vesting in the three main areas of refugee assistance that accounted for 88 percent of the UNHCR/Burundi spending in 2003-7: coordination and support services (36 percent), protection, human rights, and rule of law (29 percent), and shelter and non-food support (23 percent). The end products that PRM intended to ‘underwrite’ through UNHCR included the following:
· Voluntary, safe and dignified return – one of the key outcomes of UNHCR efforts in the ‘Coordination and Support’, and ‘Protection, Human Rights, and Rule of Law’ sectors of assistance.
· Post-return protection of returnees –the main intended outcome of the UNHCR household-level monitoring effort in the ‘Protection, Human Rights, and Rule of Law’ sector of assistance.
· Returnees’ ownership of land – the main intended outcome of the UNHCR conflict mediation effort in the ‘Protection, Human Rights, and Rule of Law’ sector of assistance.
· Returnees’ ownership of housing that, along with access to land, boosts returnees assets at the post-return stage and primes their successful reintegration in the long term. This is the main intended outcome of the UNHCR effort in the ‘Shelter and Non-food’ sector of assistance.
· In the ‘Coordination and Support’ sector of assistance, UNHCR worked to ensure effective participation of other stakeholders (Government of Burundi, donor agencies, and implementing partners) in supporting UNHCR programs, scaling them up, filling gaps, and bridging them into the development phase.b. Bilateral Funding and Program Results:
PRM’s bilateral funding strategy in Burundi met the congressional requirement that bilateral programs be used to fill gaps in multilateral funding. As a result of this study, PRM bilateral grants were assessed as non-duplicative to UNHCR programs in one or several of the following ways:
· Complementary across sectors of assistance: PRM’s main area of activity was ‘Water and Sanitation’, where UNHCR had minimal presence. In 2003-7, the total budget of PRM-funded bilateral grants in water and sanitation was $3.27 M -- 5.7 times more than the UNHCR’s $0.57 M. The expected outcome of the water and sanitation programs is improved environmental health. High-quality construction and maintenance of water systems and latrines, enhanced with community strengthening and hygiene education are the prominent contributors to sustainable reintegration.
· Complementary within a sector of assistance: In the ‘Protection, Human Rights, and Rule of Law’ -- one of the pillars of UNHCR’s assistance -- PRM provided non-duplicative ‘niche’ financing for GBV prevention and healthcare support. A knowledgeable Burundian professional who consulted for this evaluation referred to the phenomenon of GBV as “a product of virility, personal vendetta, and general frustration”. Whether these factors can be effectively tackled in the life span of one or several successive grants is unclear. Peace building and access to economic opportunity will gradually lead to an environment in which GBV will be repelled by a combination of redefined social norms, restraining laws, and their tightened enforcement. While expecting these solutions to materialize in the long term, GBV programs continue to generate useful outputs but, unlikely, measurable outcomes or impact.
· Complementary in time: The construction/rehabilitation of a primary school and two health centers was directly funded by PRM in 2006. UNHCR withdrew from building social service facilities in 2005, encouraged by some of their donors to make room for development partners. In UNHCR’s opinion, it created a one-year transition gap (ERv2: 49) that two PRM-funded grants helped to alleviate. The sustainable outcome to be expected from the PRM-sponsored construction/renovation projects is a steady provision of basic health care and primary school services in selected areas of high return. The supply/demand gap in these sectors reached critical levels after the population had grown rapidly because of massive repatriation, and the government eliminated user fees in the maternal and children’s care and primary schools.
· Complementary in terms of “the more the better”: UNHCR alone was unable to scale up shelter assistance to a level at which supply could keep up with post-return demand. Several bilateral programs were put in place to address the gap. PRM shelter grants to World Relief and recently to CORD have reinforced this multinational effort. The expected outcome of shelter assistance programs is in putting the repatriated and local vulnerable families on track to economic self-reliance. The shelter programs will prove sustainable if the currently built houses last until sufficient family, public, and private business resources are accumulated in Burundi to provide for the next round of housing renovation and expansion.
· Complementary by type of benefits: PRM’s emphasis on water and sanitation, and secondary attention to health and education contributed to the ‘public goods’ component of post-return aid, thus supplementing the UNHCR focus on household-level benefits. Aid programs that support communities as a whole are important for dissipating the sense of inequity that has been fomenting among the incumbent population in the high-return areas of Burundi. Durable functionality of assisted public facilities and improved access of all residents to drinkable water, sanitation, and basic social services represent sustainable outcomes to be expected from these programs.· Complementary in terms of intra-program synergies: Small-scale husbandry was part of the PRM-funded grants to World Relief. While agriculture is not a primary assistance area either for UNHCR or PRM, the bundling of interventions across several assistance sectors, including agriculture, provides for a comprehensive response to intertwined community needs and generates synergistic effects within the multi-sectoral programs. Since late 2007 and particularly in 2008, the Government of Burundi and UNHCR promoted the idea of peace villages: an integrated approach to assistance that combines land, housing and livelihood support. PRM pre-empted this strategy with their bilateral grants in 2006 and 2007. The impact of the GoB, UNHCR and PRM efforts is yet to be observed, however in this case it can be assessed in the logical framework of ‘attribution through contribution’: the concept of integrated assistance makes so much sense that, if properly fine-tuned and rigorously implemented, it is bound to produce a positive impact. The impact can be presumed even before observed as long as implementation accurately follows the concept and the plan.
The outlined program results meet the definition of outcomes and impact, adopted for this evaluation, or serve as proxies, highly correlated with outcomes and impact. Most of these results were addressed in the course of the evaluation.
3.2 Achievement of Repatriation
‘Support for the return of the uprooted to their home communities’ is a priority objective of the Migration and Refugee Assistance (MRA) and the Emergency Refugee and Migration Assistance (EMRA), funded through PRM programs (U.S. Congress, 2007: 79). Starting in 2005, PRM’s main expectation from their investment in the UNHCR operation in Burundi was successful repatriation of Burundi refugees[2]. By funding the UNHCR-assisted repatriation, PRM led the UNHCR donor community in their standing request for: (i) the achievement of return, and (ii) the assurance of voluntary return in safety and dignity. This section summarizes the evaluation findings about the achievement of return, while the following Section 3.3 will focus on the ‘quality of return’.Responding to the first evaluation question: ‘Did Burundian refugees return?’, the evaluation produced the following findings and conclusions:
· Repatriation to Burundi started in January 2002. By the end of October 2008, 463.2 thousand refugees returned to their home country. This number puts the completion of assisted repatriation within a one-year reach.
· Repatriation advanced at an annual pace consistently slower than UNHCR’s annual projections. Had the projected annual numbers been met, the number of repatriates reported to date would have been reached in 3.5 years instead of 6.5 years. This would have changed the demographic growth pattern in Burundi from the internally driven to repatriation-driven. The annual population growth rate in the three provinces of high return would have increased from the difficult-to-withstand 5.2 percent to an even more challenging 7.8 percent. This might have led to a breakdown in the socio-economic fabric of durable return.
· The repatriates are not at all a drag on their home country’s resources, judging by the comparative human capital that they represent: their average level of education (years of schooling) and self-assessed health, estimated in the household survey, are on a par with the incumbent families. However, the lack of a matching increase in Burundi’s physical capital renders this additional human capital largely redundant. UNHCR made a prominent effort to boost physical assets in the high-return areas through their program of assisted housing construction and by facilitating access to land. However, the trickle-down impact of these investments on the economy’s aggregate factor productivity would transpire over time. In the meantime, each new repatriate adds to a trend that may be best characterized as accelerating negative economic returns. This posits a realistic conclusion that the country could hardly afford any more repatriates any sooner.
· Repatriation assistance was managed in two / three stages: (i) transportation of returnees from refugee camps in Tanzania and Rwanda to a ‘primary’ transit center in a province of entry; (ii) from a province of entry to the final destination area; (iii) ‘secondary’ transit centers were established en route from the ‘primary’ transit centers to the communes of return to serve peak loads of returnees in selected high-return areas. A specially designed extended-stay transit center, called a temporary lodging facility, came to play an important role in the facilitation of old caseload return.
· The geographic pattern of repatriation distinguishes three hub-and-spoke clusters of provinces in Burundi. They are formed by each of the three main entry provinces, bordering with Tanzania (a hub) and several provinces of final destination (spokes) for which the respective entry province serves as the principle ‘supplier’ of returnees.
· The involvement of the main entry areas in the repatriation assistance happened in succession. Muyinga took a head start in 2002-3 and carried the main load of facilitated repatriation to seven provinces, including three (Ruyigi, Gitega and Cancuzo) that later were transferred under the operational authority of the UNHCR office in Ruyigi. In 2004, Ruyigi became the main province of entry for assisted repatriates. In 2005-7, Makamba ‘processed’ the largest numbers of assisted returnees.
· UNHCR had predicted repatriation. However it took 2.5-3 years for the repatriation assistance program to get up to speed and establish itself in all the main entry provinces. Compromised security due to setbacks in political transition determined the delayed and geographically uneven response to spontaneous repatriation in 2002-4.
In summary, repatriation is a fact of life now (short of the concluding one-year effort), and UNHCR has been there to make it happen. PRM’s leading contribution to the UNHCR/Burundi budget in the years of high return (2003-7), has delivered on the core expectation that, if nothing else, UNHCR will always provide the service of returning people to their home country. The quality of this basic service that PRM ‘procured’ from UNHCR in the past five years is reviewed in the next section.
3.3 Quality of Repatriation
The material in this section provides an answer to the second evaluation question, namely: ‘Did refugees return in safety and dignity?’ Three specific questions were formulated and answered during the evaluation to prepare and explicate the answer to the general question: (i) Was return voluntary? (ii) Was it safe? (iii) Was it protective of the returnees’ civil status and economic interests?3.3.1 Achievement of Voluntary Return
Upholding the voluntary nature of return is the cornerstone of the protection of human rights in the context of repatriation. UNHCR’s core policy principles and operating procedures revolve around the assurance of voluntary return. This qualitative aspect of return is an important area of inquiry for a PRM-sponsored impact evaluation for three reasons: (i) Voluntary return signifies an interruption of the vicious circle of forced displacement. Conversely, forced return would represent a new permutation of it. The assurance of voluntary return can thus be seen as a valuable benefit of the repatriation assistance, and an important humanitarian payoff on PRM’s contribution to UNHCR budget. (ii) There are reasons to believe that an internally motivated return has stronger chances to mature into durable return at the reintegration stage. (iii) By awarding a grant to the Voice of America’s project of strengthening radio broadcasts as an information medium for targeted promotion of return to the Burundian youths, PRM has cast its support for the principle of informed return and, thus, put the voluntary repatriation agenda on the Bureau’s priority list. The equivalence of voluntary return and informed return is elucidated in the complete version of this report.Responding to the sub-question ‘Was return voluntary?’, the evaluation produced the following findings and conclusions:
· The assessed interplay of the support and risk factors of voluntary repatriation, suggests that the voluntary return is a complex and dynamic construct both in concept and in practice. In the context of repatriation, few choices could be made outside the Hegelian perspective on freedom as the recognition of necessity. Burundian refugees had to chart their decisions in response to a contradictory mixture of push and pull factors, in which positive and negative incentives (reward and fear) were intertwined.
· Clearly, UNHCR could not be expected to put a tap on all the sources of pressure and uncertainty that surrounded the repatriation and impeded the free will to return or not to return. The pressure from the host country and the ‘descent to the valley’ acceleration of the return process when the end is already in sight create a ‘streamlined’ approach to repatriation with less focus on the sovereign choice of the repatriates themselves. Yet, these phenomena are part of the reality, dictated by political factors outside UNHCR control and organizational pressures that are inherent in UNHCR’s institutional sense of timing, effectiveness and efficiency.
· Much to UNHCR’s credit, the tripartite ‘bridge’ between Tanzania and Burundi was instrumental in moderating the initial attempt to fast track repatriation beyond the capacity to manage or absorb it. By helping set a more realistic pace of return, UNHCR gained a significant amount of time to implement a carefully scripted and well organized information campaign in support of voluntary return.
· The multi-media information support has been customized by broadly defined refugee categories and narrowly profiled risk groups. A wide range of media included modern and traditional communication channels, enhanced further by an informal ‘information grapevine’. Taking advantage of the relatively short distances between refugee camps and return areas in Burundi, UNHCR was able to arrange a ‘sneak preview’ of life in post-conflict Burundi by arranging refugee travel to the areas of return and facilitating radio/phone communications across the Tanzanian border. According to household survey data, 16 percent of the repatriate families had direct visual exposure to their area of return prior to repatriation. By enabling this important and laborious component of their assisted repatriation program, UNHCR strengthened informational and, ultimately, motivational resource behind the voluntary decision to return.
· UNHCR’s policy-level commitment to voluntary repatriation was implemented through strong procedural safeguards that included voluntary registration and the assignment of personal identification and clearance functions to the Voluntary Repatriation Form (VRF). By linking the readmission process to the information in VRF, UNHCR hardwired repatriation to the principle of voluntary return.
· The protection of voluntary return is an on-going effort. UNHCR has to walk a fine line of promoting return without forcing it. This may present a particular challenge at the end stage of repatriation when the principle of voluntary repatriation can be sidetracked on both sides of the border. The heightened attention of PRM and other donors to the problem of naturalization of the residual caseload in Tanzania may help UNHCR cope with the outlined risk and avoid ‘overdoing’ the return[3].
In summary, the voluntary nature of return was assured both at the process and outcome levels. PRM’s support of the assisted return program in Burundi helped protect the basic human rights to free movement and truthful information in the tenuous environment of massive repatriation.
3.3.2 Assurance of Safe Return
Safe return is the second commitment and concern of UNHCR in the triad of voluntary, safe and dignified return. Safe return is assured through security en route and at transit centers, by identifying cases of social vulnerability and health risks prior to return, and taking special care of them throughout the return process. The safety strategy at the return stage is linked to the post-return protection effort as a quintessential part of the UNHCR’s program of reintegration support. The return to Burundi from the standpoint of returnee’s personal safety was examined in discussions with UNHCR Protection Section and in a series of field observations. The summary of findings is presented below:· Personal safety protection has been integral to the assisted return program of UNHCR/Burundi. UNHCR’s core competences have been successfully engaged in protecting refugees in all stages of their trip home: from refugee camps and old settlements to clearance at the border; from there to the entry transit centers; and on to the ‘secondary’ transit centers and/or areas of return. The overarching strategy of personal protection was to expedite all stages of return: to make the ‘door to door’ travel time as short as possible and thus shorten the exposure to the risk of adverse events. The favorable geographic setting of Burundi was a contributor to quick repatriation, and so were effective convoy management, transit center operations, and management of the domestic leg of return.
· Field observations and opinions collected in the course of this evaluation supported the conclusion that the assisted return program has come a long way since its start-up in 2003-4 when the transportation standard of 50 people per truck was not sustained, a transit center operation was characterized as “cramped, making the process of registering the returnees a complicated, lengthy and stressful affair”, and returnees, reportedly, had to walk up to 30 km to reach their colline because UNHCR fleet did not have smaller vehicles to transport returnees below commune level (Refugees International, 2004).
· In addition to minimizing all safety risks by managing return at a brisk pace, UNHCR worked with competent implementing partners to control specific safety risks. Healthcare support en route and at transit centers was targeted to at-risk returnees. The definition of at-risk groups, used in the assisted return program, was corollary to social vulnerability, medical history, and current health status. Up-to-date patient information produced in a pre-return health exam and recorded in the personal bill of health, enabled a customized approach to case-level management of health risks during return. Approximately 15 percent of the returnees in a sample convoy were monitored for health problems and social vulnerability. This ratio suggests that health and vulnerability protection was a generously, yet realistically sized component of the assisted return program.
· The level of trans-border coordination between UNHCR and partners was found to be sufficiently high to enable on-schedule departure and arrival times, a smooth passage of convoys through the border, and a seamless transfer of information, responsibility, and operations control from the protection teams of UNHCR and partners in Tanzania to those in Burundi.
· The lack of a clearly defined mandate and authority, as well as the absence of financial incentives continued to plague the Reception Committees – a nationwide community-level initiative inspired by the Government of Burundi jointly with UNHCR to ensure that the return is successfully concluded and the established safety standards are upheld through the very end of the process.
· The safety-related incidents including a few cases of physical assault were captured in the household survey in the three provinces of high return. The incidence rates for three out of four types of adverse events were higher among the unassisted repatriates than among the assisted ones. This limited statistical evidence adds to the conclusion that the UNHCR-led protection effort effectively minimized the potentially injurious effects of return on the repatriates’ personal safety. With complete confidence, the evaluators have concluded that the personal safety component of the assisted return is a much better alternative to a laissez-faire approach when it comes to moving people through terrain with an uncertain-to-compromised level of security.
· The reviewed system of support and protection was designed for the fast-paced return from refugee camps that takes several days to complete. In 2008, this system has been challenged by the experience of repatriating the 1972 caseload. A separate transit center and a temporary lodging facility ( a ‘tertiary’, long-stay transit center of sorts) had to be built, the latter to accommodate the final stages of repatriation of landless families with poor chances to reclaim their land. It takes 3-4 months to move these people to a peace village or alternatively defined place of residence. ‘Frozen in return’, these old caseload returnees are supported by the same UNHCR partners who provide health and protection services to convoys and at short-stay transit centers. However, engaging with a community of several hundred people for several months requires a longer attention span to a more diverse range of problems. The ‘fuzzy logic’ of supporting prolonged return does not fit in a simpler algorithm of supporting returnees for 3-4 days. There were indications at Gitara that AHA found itself somewhat out of its depth. Illustratively, the organizational and technical skills of MSF or IMC could be more appropriate in a situation where return does not end until reintegration is partially accomplished.
Overall, the question whether the return was safe is answered in the affirmative. The UNHCR donors’ money was well spent in this segment of the assisted return program. As UNHCR’s largest sponsor in Burundi in the previous several years, PRM can take credit for the accomplishment of safe return by UNCHR and partners. In this particular case, the logic of assigning credit is somewhat indirect. ECHO actually used its earmarked allocations to UNHCR to fund a major part of the return logistics. The funding earmarks were supplemented by strong ‘brand promotion’: EU stickers on the gates and walls of the return facilities (transit centers, temporary lodging facility in Gitara) leave no doubts as to the origins of donor support. At the same time, ECHO/Burundi rightly pride themselves for their agile funding strategy, intended to fill current gaps and structural imbalances in the rest of the UNHCR program portfolio. It will be logical to conclude therefore that UNHCR’s use of PRM’s unearmarked funds in other segments of the assisted return program saved ECHO resources for concentrated investment in the logistics of safe return.
3.3.3 Protection of Returnees’ Legal and Economic Interests
With reference to the threefold policy of voluntary, safe, and dignified return, the protection of returnees’ civil status and assets contributes to their personal and economic safety and dignity on a par with physical protection. Furthermore, these components of the assisted return program play a unique role of priming repatriates for reintegration. The evaluation examined both aspects of the UNHCR assistance return program: (i) helping returnees’ legalization in Burundi; and (ii) protecting their economic interests during return. The main findings and conclusions follow:a. Assistance with Legalization:
· UNHCR’s empathetic approach to returnees problems’, their drive for practical solutions, and their established partnership with the Interior Ministry and province governor offices resulted in a simpler and quicker legalization process. Essential parts of this process have been woven into the ‘technology’ of assisted return and are now completed at the transit centers before repatriates make their first contact with Burundian authorities. UNHCR became a ‘family integrator’ for many repatriated families: when people register for return as one family, are listed on the Passenger Manifest as such, and travel together, they will be recognized as a family at the registration upon arrival and, most probably, from then on. The streamlined naturalization process is an effective component of the assisted return program.
· The trade-off on UNHCR success was partial bypass of the public administration systems in the areas of high return – not a step toward system sustainability in the post-UNHCR future.
· The quick legalization process has been strained lately by the complexities of the return of refugees since 1972. Critical to the issuance of the National ID Card is the returnee’s knowledge of his/her commune and colline of return. In hundreds of 1972 cases this information would remain unknown either to UNHCR or returnees for several months after their return to Burundi.b. Transfer of Assets:
· The assisted repatriation program protected returnees’ economic interests by: (i) moving their possessions and livestock; (ii) quickly registering them for, and dispensing the initial set of cash and in-kind benefits; and (iii) providing special support for repatriated refugees since 1972.
To conclude this set of findings, UNHCR effectively helps returnees to start anew in Burundi not only by helping them arrive physically, but also by making sure that their legally defined identity and possessions catch up.
3.4 Achievement of Reintegration
Following a comprehensive review of the assisted return results in Section 3.3, the next two sections present the second set of the evaluation findings, namely those referring to the outcomes of assisted reintegration.3.4.1 Integrated Assessment of Reintegration
a. Econometric Model 1: Comparative Welfare of Repatriates and IncumbentsTo answer Evaluation Question 3 ‘Did returnees achieve reintegration?’, an econometric model was developed for statistical estimation and supplied with data from a household survey, conducted as part of this study. The model tested the hypothesis that household welfare in Burundi depends on whether the households are incumbent or repatriate. Put simply, there is a difference in welfare between incumbents as a group and repatriates as a group. If this is found to be true, reintegration has not been fully accomplished. If, based on statistical evidence, the hypothesis is rejected, that is the welfare levels are found not to differ between incumbents and repatriates, then reintegration has progressed to essential levels. Importantly for the evaluation agenda, the welfare index was constructed to capture main lines of reintegration assistance. Therefore, if the repatriates welfare is found indistinguishable from that of the incumbents, it is reasonable to assume, that the aid program played a role in achieving this parity. Multiple statistical analyses, performed on this model have led to the following conclusions:
· On a representative population sample selected in three high-return provinces, repatriates and incumbents were found indistinguishable by the level of household welfare.
· The post-return experience of the sampled refugee population is 4.2 years. This implies that repatriated population matched the local welfare standard within 4.2 years after return.
· The model does not specify how long it took the repatriates to achieve welfare levels similar to those of the incumbents. It will take a much larger household sample to study the evolution of comparative welfare by post-return month/year in order to understand when exactly the repatriates reach the incumbents within this post-return 4.2-year period.
· The welfare of both the repatriates and the incumbents depends predominantly on their household size: welfare levels drop with the increase in the number of household members. The drop is precipitous in the household size sub-range from very small to medium.Based on the above, it will be fair to conclude that by now repatriates indeed have reintegrated with the incumbents in the communities of return.
b. Econometric Model 2: Welfare Variance among the Repatriate Households
The second econometric model was designed to take a closer look just at the repatriate families: particularly, to find out what features of their pre-return and post-return past might have significantly helped reintegration. The model related welfare to the number of years in displacement, whether time in refuge was spent in camps or settlements, whether return was assisted or not, how much time elapsed after return, whether head of household has vulnerability status, and the household size. The results of testing the model are summarized below:
· Welfare among the repatriate households is strongly influenced by household size. The relationship is inverse: larger families imply lower welfare. This variable alone accounts for 28% of the variance in household welfare in the repatriate population. Household size is by far the best predictor of welfare levels across the repatriate households.
· The variable ‘Years in displacement’ was identified as a small but statistically significant factor of welfare among the repatriate households. Welfare index values tend to grow with the increase in the number of years spent by refugees in displacement. Given that the 1972 refugees have the highest value of this characteristic (they were displaced for the largest number of years, compared to other returnees), it will be reasonable to conclude that their relatively advanced livelihood skills and observed entrepreneurial spirit give them an edge over the rest of the repatriate community in providing for themselves and improving their well-being.
· The fact that the other variables were found insignificant does not necessarily discard their importance. It is plausible that on the fifth year after return[4] the experience of asylum (camps or settlements) and return (assisted or unassisted) faded and current factors such as family size drive the welfare variance among the households. The important exception, according to the findings from this model, are the intrinsic resilience and coping skills that the 1972 ex-refugees acquired in their pre-return years and that continue to serve them and their descendants in Burundi.3.4.2 Assessment of Reintegration by Welfare Component
Following the integrated assessment of welfare by means of econometric modeling, the evaluators decomposed the welfare index and used household survey data to compare the repatriates and the incumbents by their current assets (consumption and savings), physical assets (housing), human assets (health and education), and social assets (safety). The objective of this analysis was to validate or challenge the previously made conclusion that the welfare levels in both populations of interest are indistinguishable and reintegration, therefore, has happened. The results of this analysis are outlined below:a. Household Consumption and Savings:
· The levels of consumption plus savings are very close in both populations of interest: repatriate households stand at 96 percent of the incumbents in per-household terms and at 97.5 percent per household member. These indicators certainly reinforce the finding from the econometric test about the convergence of living standards across both populations.
· In per-household-member terms, the repatriates are slightly behind the incumbents in food consumption (98.3 percent). They are considerably lower on savings (88.3 percent of the incumbents). Their non-food consumption is 8.1 percent higher than in the incumbent households.
· The average monetary value of per capita food consumption is USD 0.93/day in the repatriate families and USD 0.95/day in the incumbent families. The share of food in the current consumption (without savings) is 81.5 percent and 80.9 percent in the repatriate and incumbent families, respectively. The savings are worth 9 days of consumption in the repatriate households and 11 days in the incumbent households. If spent exclusively on food, savings will keep the repatriate families going for 10 days, and the incumbent families for 12 days. The very low dollar value of daily per capita food consumption, very high share of food in the total household budget, and the short-lived savings uniformly point at the parity between the repatriates and the incumbents that can be characterized as equalization in extreme poverty.
b. Housing:
· Based on household survey data, both the repatriate and incumbent families, typically, live in houses made of adobe bricks, with clay/earth floor. They use kerosene/oil for lighting and firewood for cooking. A marked difference is the predominance of metal roofs on houses of the repatriates (93 percent) thanks to the massive shelter assistance provided by UNHCR and bilateral donors over the years of high return and inclusion of 22 corrugated iron sheets in the UHCR-sponsored construction material kit. The 49 percent of the incumbents’ houses still have thatch/straw roofs. An important progressive feature of the repatriates’ houses is the high prevalence of metal-roofed outhouses with pit latrines – also, an integral part of international shelter assistance.
· There are no significant differences between the repatriates and the incumbents in terms of the time it takes them to get from home to a local source of drinking water, food market, health center and primary school.
· The physical and location characteristics of housing were integrated into the Housing Quality Scores, designed as part of this evaluation. The scores have shown a close-to-normal distribution, almost identical across both populations of interest. The high similarity of housing quality scores for both types of households is consistent with the similarity of housing characteristics, discovered in the study.
In summary, the repatriate households are on a par with the incumbent households as far as the quality of their housing is concerned. The supply of housing is determined by two factors: availability and ownership status. All households in the sample-based survey had shelter. The ownership versus tenancy was not measured this time. A predecessor survey conducted at the end of 2006, showed the ownership rate of 89 for the repatriate families and 95 percent for the incumbent families (JAM, 2007: 39).
A comprehensive assessment of housing supply and quality in the high-return provinces of Burundi attests to the fact of reintegration between the repatriates and the incumbents.
c. Human Capital -- Education and Health:
The following summary indicators of education and health were estimated on the household survey data produced in this study:
· Average number of years of schooling per household member: 3.60 in the repatriate households and 3.24 in the incumbent households;
· Self-assessed health (on a one-to-five scale from ‘very bad’ to ‘very good’): 3.40 in the repatriates households and 3.46 in the incumbent households.Given PRM’s priority attention to environmental health, environmental health status was assessed on the repatriate and incumbent children younger than five years. The epidemiological focus was the incidence of diseases that are highly sensitive to micro-environmental hazards, such as lack of safe water; poor sanitation and hygiene, indoor air pollution, and overcrowding. The incidence of such diseases must be indicative of the outcome of programs in water, sanitation, and habitat/community strengthening that PRM funded as part of the reintegration effort in Burundi. The main findings follow:
· 91.6 percent of children in the incumbent families and 78.6 percent of children in the repatriate families were ill with at least a fever over a two-week period, preceding the survey.
· The incidence rate for diarrheic diseases was 39 percent for the incumbents and 31 percent for the repatriates. The symptoms of severe diarrhea, likely to involve bacillary forms, was markedly lower in the repatriate families: 4.6 percent compared to 10.5 among the incumbents. The coordinated effort to increase supply of, and access to potable water and sanitation, on the one hand, and aggressive promotion of modern hygiene, on the other, must have contributed to this difference. In the communes of high return all the families usually benefit from assistance programs in water and sanitation. However, since the survey also included communes with low presence of repatriates, the intake of watsan assistance in those communities could be much lower. Hence, the observed sample-wide difference in the incidence of conditions directly linked to water and sanitation. Another contributing factor could be the lasting positive effect of health education and hygiene promotion in refugee camps.
· With similarity to diarrheic disease, the upper respiratory disease rate is lower in the repatriate children. However the symptoms of acute lower respiratory diseases (ALRIs) – one of the main causes of death in small children – are more incident in the repatriates. It is unclear whether this is a response bias, or there are adverse circumstances that increase their risk exposure. Children’s susceptibility to ALRIs is usually higher in families on the move, without permanent housing, and/or with parents who are negligent or do not have established access to health care. Some of these circumstances are corollary to the early post-return period.
Further inquiry was focused on the access to, and supply of water as important determinants of environmental health:
· Survey data suggest that the repatriate households are distanced from their sources of water by a somewhat longer commute time than the incumbent families: 58 percent of the repatriates have their water source within a 30-minute roundtrip walk compared to 68 percent of incumbents.
· At the same time, the water sources used by returnees are of higher quality than those used by the residents: 93 percent of the repatriate households use water from protected springs / taps, compared to only 61 percent of incumbents.
· An integral measure of water supply is the water shortage rate. Here again, the repatriates are better-off compared to the incumbents. The ‘no shortage’ rate is 53 percent and 44 percent, respectively. Seven percent of the repatriate households experience systematic shortage of water, vis-à-vis 13 percent among the incumbent households.
These findings should be put in a properly defined perspective of water use in Burundi. Most water sources and systems are a public good: built for shared use. If drinking water comes to a colline, it would be available to a cluster of rugos or to a large part of the colline, if not the whole colline. The significant international effort of protecting water sources and building gravity-fed water systems was geared to the communes of high return and helped both the repatriates and the incumbents in those communes. The observed significant difference in water supply between the repatriates and the incumbents must be due, mostly, to scant availability of drinking water in communes and villages with low presence of returnees.
Adherence to four modern hygiene practices was estimated in both populations of interest: (i) water management and storage practice, (ii) household refusal handling, (iii) children’s defecation, and (iv) feces removal. Consistent with the findings in other parts of the comparative welfare study, the repatriate and the incumbent families are similar, judging by their adherence rates in all the four examined areas. The noteworthy finding is that people are significantly more knowledgeable and conscientious about water than about sanitation. The water component of public health education must have benefited from the flow of aid in the water source protection and water system building, as well as in community education on the techniques of water collection, storage, and use. Given that the observed pediatric diarrhea rates are still high and worrisome, the weak chain in public health may be elsewhere (not in water). Most probably, it is in sanitation. And building pit latrines may not be enough. Study results suggest an important need for teaching people disposal and excreta management practices as aggressively as it has been done for water management practices.
d. Safety:
The repatriate and incumbent families agree about the level of, and threats to their safety:
· Most respondents in both groups think of their current life as reasonably safe.
· The predominant opinion that safety has considerably or somewhat improved over the past two years or since return, is shared by 84-92 percent of respondents in both groups.
· According to another shared opinion, two threats to safety still exist at sizeable levels and have aggravated over the past 3-4 years: (i) theft, looting, and damage to property; (ii) armed gangs and/or militias.In summary, welfare levels in each of the components (consumption, savings, housing, health, education and safety) were found similar in both populations of interest. Returnees have an edge over incumbents in the areas of children’s health that are highly correlated with the quality of micro-environment: water, sanitation, and indoor air pollution. On the whole, component-specific welfare analysis corroborates the main finding from econometric modeling that reintegration has been achieved.
3.4.3 Assessment of Reintegration as Perceived by Returnees
Following the technical analysis of welfare at the integrated and component-specific levels, the third round of assessment was focused on ‘perceived reintegration’ – as seen by the returnees themselves. The objective of this third round of reintegration assessment was to reinforce or challenge conclusions from the previous two rounds.In this part of the study, a more inquisitive approach is taken: returnees are not related to their neighbors but instead are asked to assess whether their life in Burundi has improved compared to their past and whether the post-return reality delivered on pre-return expectations. The following are not citations of returnee opinions but their attitudes synthesized by cross-matching their answers, followed by evaluators’ comments:
· ‘We, repatriates may well be on a par with the rest of the local population [as shown in the previously presented welfare estimations], however if we compare our current life to our own levels from the past, the picture will be mixed at best’. -- If compared with the pre-displacement past, today’s well-being is much or somewhat better for 30 percent and somewhat or much worse for 64 percent of the returnees. If compared with their pre-return past, the current well-being is much or somewhat better for 42 percent and somewhat or much worse for 51 percent of the returnees. Repatriates, it appears, have already achieved the current welfare level of their neighbors, but are yet to reach the levels that they had enjoyed in the past, particularly before they were displaced from their homes in the 1970s and/or 1990s. This implies that Burundi as a country has not yet restored the living standards of its citizens to their pre-crisis levels.
· ‘Many of us, repatriates of the recent three years, expected to do better soon after the return than we actually did’. -- If first-month post-return experience of the fairly recent returnees (from 2006-8) is compared to their pre-return expectations, it turns out that reality did not quite deliver for 77 percent of them on the expectation of international aid (77 percent rated their experience as somewhat or well below expectations), for 67 percent on the expectation of support from the Government of Burundi and local authorities, and for 28 percent on the expectations in terms of political and economic situation in Burundi. On the positive side, the receptivity in the host community was assessed well above or somewhat above expectations by 53 percent of the respondents, and only 18 percent expected more.
· Countering their own attitude of under-achievement summarized in the previous two conclusions (‘We are still worse-off than in the past, and the reality did not measure up to our expectations’), the returnees never felt disappointed about their decision to return: 88 percent of them felt very or somewhat satisfied that they came back to Burundi at the end of their first post-return month and 98 percent, 4.2 years after the return.
The overall sentiment of the repatriates appears to be this: ‘What we expected and what we achieved is secondary to our main accomplishment: ‘We are back in our homeland.’ Humbling this study’s prevalent focus on the material aspects of reintegration, the repatriates’ have asserted the preponderance of reintegration with their inner self as Burundians. – A truly dignified way of thinking about, and living through the return experience!
3.5 Attribution of Reintegration to Assistance
A large part of Burundian repatriates possess strong resources of self-motivation that contribute to their successful reintegration but do not substitute for reintegration aid. Assistance plays a dual role of solving the returnee immediate problems and strengthening their self-reliance in the long term.The value of reintegration aid can be inferred from returnees’ opinions: 58 percent believe they considerably benefited from being a returnee in Burundi. This share varies by return period: ‘considerable beneficiaries’ account for 39 percent of those who returned prior to 2002, 59 percent of those who returned in 2002-4, and 63 percent of those who returned in 2005-8. It is reasonable to conclude that the positive side of being a returnee is chronologically correlated with assisted return. Returnee status is valued the highest in 2005-8: most of that period was covered by a full-fledged reintegration benefit package under the assisted return policy of UNHCR.
Assistance, provided by UNHCR and supplemented by bilaterally funded projects, contributed to all the components of household welfare: current consumption, physical assets, human capital, and safety. If this assistance is found to be of high volume and quality, the finding will provide the fourth layer of evidence of successful reintegration. (The previous three were developed from the study of integrated welfare, component-specific welfare, and returnees’ perceptions.) The following summary presents key snapshots of evidence about the quantity and effectiveness of reintegration assistance by sector.
3.5.1 Food Assistance
The centerpiece of food assistance is the food package. Other benefits in the food security assistance sector that had tangencies with the return assistance program included WFP-supplied food for school-based meals (cantines scolaires), and FAO-supplied seeds.The food package was introduced at the start of the assisted return program in 2002 and progressively extended from the original 3-month to a 6-month entitlement (since August 2007). Spontaneous returnees “received food rations for either one month or three months” in 2002 (UNHCR, 2002: 116) and continued to receive them with somewhat less commitment from UNHCR than did assisted returnees.
Observed reintegration effects of the food aid are outlined below:
· Returnees creatively adapted the food package to their coping strategies. Food from the food package was consumed, sold for cash, traded for more (cheaper) food, traded for more diverse food, and shared with host families. The outlined practices represent a rational approach to post-return coping. UNHCR reasonably controlled these practices by changing from a lump-sum distribution of food assistance to distribution in six monthly installments; and by introducing a cash grant in 2007 so that returnees could have spending money without selling their food package.
· The food package played a visible role in returnee nutrition: it accounted for an average 26 percent of the total food consumption in a sample of households who came back in 2008 (up to nine months after return). Within the first six months, its share is estimated at 1/3, and much higher in the first weeks after return.
The food package, clearly, plays an important role in tiding the repatriate families over to their first post-return harvest. The delivery of this benefit is effective, as was demonstrated by a high visibility of food from the UNHCR assistance program in the household data on food consumption.
3.5.2 Cash Support
The cash allowance was established in August 2007. Key informants in this evaluation praised it as a high-impact benefit in the UNHCR-sponsored program of reintegration support. With the average family size of close to 6 persons in Burundi, an infusion of cash at the amount of BFr 50,000 per family member (currently $42) provides families with spending resources commensurate with the price of a small to medium plot of land.The main reintegration effects of this benefit are as follows:
· An infusion of discretionary income -- a recognized way of jumpstarting the post-return household economy.
· Cash available to the repatriates, eases strain in their relations with host families immediately after return. The allowance thus plays a role in improving social cohesion in host communities.3.5.3 Land Conflict Management
UNHCR/Burundi showed a strong capacity for strategic leadership in land-related areas of policy development and implementation assistance. The Government of Burundi recognized the political and professional authority of UNHCR by putting them to the fore of the stakeholder consultation process, currently in place under the auspices of the Peace Building Commission (Burundi Configuration). Additionally, UNCHR has become the GoB’s trusted partner in implementing the jointly planned peace village project – a bold attempt at cutting the Gordian knot of landlessness among the old caseload returnees.The following is a summary of findings about the conflict-mediation help that returnees receive as part of the UNHCR program of reintegration assistance:
· Conflict mediation is provided by the Bashingantahe (traditional wise men justice), the Commission on Land and Other Assets (known by its French acronym as CNTB), and UNHCR partners. They all work to maximize the volume of conflict resolution through out-of-court settlements.
· UNHCR implementing partners use the following basic methods: (i) provide guidance and advice to contending parties and leave them to their own devices. (ii) help mediate toward an amicable arrangement; and (iii) refer plaintiff to an alternative mediator or for hearing before court. These services are provided free of charge. All the returnees with self-reported need are eligible.
· Since 2008, the UNHCR protection office has increased emphasis on prevention: detailed interviews are conducted with repatriates from the 1972 caseload during their return, and the findings are utilized to steer returnees towards return areas and land access options that would be less fraught with conflict.
· The demand for land conflict resolution support was tentatively assessed by matching population estimates to household size and dispute incidence rates. It was roughly estimated that approximately 8,000 disputes are under the purview of either the traditional justice or CNTB or UNHCR implementing partners. This accounts for about 40 percent of the estimated number of disputes. The remaining 60 percent may be in a latent stage or ‘raging in the open’ but without recourse for resolution.
· Only a modest share of registered complaints gets resolved at the mediation level: UNHCR partners settled 22 percent of their registered cases in 2006 and 23 percent in 2007 (UNHCR, 2007; …2008b: 171). CNTB resolved anywhere between 3 and 6 percent of registered complaints and had another 19 percent pending at the end of the Commission’s second year. The estimated unresolved ratio of 85 to 90 percent includes registered but unresolved cases and cases that were not filed with any conflict mediation circuit.
· The effectiveness of mediation depends greatly on the contending parties’ willingness to follow their own amicable arrangement or abide by a mediator’s ruling. So far, the will for peace through asset sharing has prevailed in the Burundian communities. Social cohesion stood its ground at two levels: many returnees were accepted back without dispute; and enough good will was displayed in resolving disputes amicably.UNHCR has done its best to control the level of land-related conflict in high-return areas. Ultimately, international relief agencies will be unable to go against the tide of public sentiment if it turns bitter over the land tenure in Burundi. Clearly understanding the pressing need for systemic solutions, UNHCR/Burundi experts have advocated for an expeditious revision of the land code and consolidation of the currently unused and misappropriated land resources into a publicly controlled pool of land for distribution to landless returnees. The Government’s action was slow at best. The land access agenda is developed further in the ‘Sustainability’ part of the evaluation.
3.5.4 Housing Construction
‘Shelter and Non-food Assistance’ is the third largest area of the UNHCR/Burundi program of reintegration assistance, after ‘Coordination and Support Services’ and ‘Protection, Human Rights, and Rule of Law’. The funding for this assistance area (primarily, of its housing component) reached $26.1M in 2003-7.The landscape of Burundi was visibly transformed by the 49.1 thousand houses built or renovated under UNHCR programs in 2002-7, and 32 percent more (based on 2006-7 data), built with bilateral funding. PRM contributed to both lines of housing assistance -- as the largest funding source for the UNHCR program of reintegration assistance for a number of years, and as a grantor for World Relief with its housing support effort in the province of Makamba. Later, another shelter assistance grant was awarded to CORD.
The key findings about the entitlements and reintegration effect of the housing program are presented below:
· The program’s standard benefit consists of a standard set of construction materials and a limited advisory and problem-solving support with building a detached family house and an outhouse with pit latrine. Eligible groups include the repatriate families and a quota-based number of qualified incumbent families – those with confirmed vulnerability status and residing in communities of high return.
· With few exceptions, families build their own house, using construction materials and advice from the housing assistance program. The do-it-yourself approach contributes to beneficiary’s vesting: improves construction quality, encourages better maintenance of the house, and, ultimately, maximizes program contribution to successful reintegration.
· 65 percent of the repatriated population of 2002-7 (both assisted and unassisted returnees) were provided with housing assistance from UNHCR.
· In 2006-7, bilaterally funded programs added 11,650 assisted houses to the UNHCR-sponsored 24,786 houses, assisted in 2006-7 and 49,276 houses assisted in 2002-7. Bilateral programs did not target returnee population as consistently as UNHCR. Some faith-based donors prioritized IDPs. Others did focus on the reintegration agenda, however, they established higher than UNHCR quotas for the local families, and these quotas varied by organization and area. Adding non-UNHCR houses to the estimation of supply for returnees requires caution. Tentatively, bilateral grants and contracts may have increased the supply rate from 65 percent (achieved by UNHCR) to 75 percent of the total returnee need.
· At the end of 2007, the largest numbers of returnees waiting to be helped with the construction of their houses were in the following provinces (thousand persons): Muyinga, 39.4; Ruyigi, 35.4; and Makamba, 14.3. These three provinces of high return accounted for 67 percent of the nationwide ‘waiting list’. These numbers, however will be lower if houses, assisted under bilateral programs, are counted. In 2006-7, 59 percent of the total number of assisted houses in Makamba were provided under non-UNHCR programs; in Muyinga, 34 percent, and in Ruyigi 11 percent.
· Assistance supply rate (ratio of supply to need) varies widely by province: from 96 percent in Bururi to zero percent in the City of Bujumbura, Bujumbura Rural, Muramviya and Mwaro. This variation highlights the following policy dilemma: which areas should be targeted first, those with the highest number of prospective beneficiaries yet to be helped, or those where the supply gap is the highest in percent of the need? – Solutions to this dilemma are outlined in report’s ‘Recommendations’ section.The assisted housing construction program has impacted on reintegration in the following ways:
· Housing assistance has put returnees on a par with the incumbent families in one of the four components of the welfare index: housing was weighted at 30 percent of the total welfare. Thus, the housing assistance program contributed significantly to the observed welfare parity between both populations of interest.
· Beneficiaries developed a vested interest in housing assistance by fighting to get it, up to the point of gaming the need verification system. While some gaming tactics are reprehensible, others may not necessarily merit harsh judgment: they indicate rational adaptive behaviors and, ultimately, mean that the repatriate households have effectively integrated assistance into their economic decision-making and put it to work for their reintegration.
· Programs of assisted housing construction boosted asset accumulation in the household sector, and produced a ‘multiplier effect’ on the economy by supporting private entrepreneurship: significant demand was created for professional and community construction services (brick-making, masonry, carpentry).
· UNHCR and bilaterally-funded projects contributed to equity in the Burundian communities by allocating a 10-percent or larger quota of assisted housing construction to the vulnerable among the incumbent families.
· Assisted housing construction became a community-strengthening experience in the high-return areas of Burundi. The shelter committee movement strengthened inter-family ties and promoted the culture of self- and mutual help by sharing resources of time and skills. The housing assistance program raised the standard of public accountability of the colline chiefs. The momentum of community strengthening, developed in the program, is likely to have a lasting effect.The assisted housing construction program had mixed secondary effects on reintegration, some of them short-term, others at a more systemic level:
· The supply/demand gap, identified in the program, results in a tightened approach to rationing. Rationing implies denial of benefits to some of the prospective beneficiaries, based on a comparative need rating. In an environment of privation, those who may be in relatively less need, are nevertheless very poor. Rationing is always perceived as discrimination by those excluded from an entitlement; much more so in the subsistence setting of Burundi. Leaving part of the repatriate and incumbent community out of the program, either because of restrictive eligibility rules, or changed funding targets, or slow implementation, creates a potential fault line under the edifice of sustainable reintegration.
· Disparate benefit coverage rates across neighboring communes put a strain on the resources of social cohesion.
· The UNHCR-sponsored construction boom in Burundi has contributed to environmental distress, particularly, deforestation as a corollary of demand for lumber.The housing assistance program is a sensitive node of organizational mandates, social and economic policies, and cultural preferences. UNHCR, as the key player in the assisted housing construction in Burundi, had to navigate between considerations of cost efficiency and operational simplicity, on the one hand, and the pressures created by non-aligned expectations from the returnees, host communities, implementing partners, and authorities of Burundi, on the other. On the whole, the housing program proved to be an effective consensus-based intervention, even if consensus was interpreted somewhat differently by its participants.
3.5.5 Water and Sanitation
Water and Sanitation assistance (referred to as ‘Watsan’ in the international assistance parlance) is a priority area of humanitarian response for PRM. As was mentioned previously, PRM’s bilateral spending on water and sanitation was 5.7 times larger than UNHCR spending in 2003-7. Given the ‘market making’ role of PRM watsan projects, watsan program results were evaluated with the emphasis on PRM-funded grants. These grants, given their volume, are representative of the entire humanitarian response in the water and sanitation sector in Burundi.PRM’s bilateral grants cover three key determinants of environmental health: water, sanitation, and personal hygiene, out of nine that WHO included in the environmental health agenda in humanitarian emergencies and disasters.
In the logical framework of this evaluation, durable return is achieved through sustainable reintegration. The latter depends on the quality of aid results and their support and enhancement at the development stage. The objective of inspecting watsan facilities, built or rehabilitated with PRM funding, was to ascertain their sustainability to date, that is the level of their functionality and compliance with basic standards of quality and safety. By comparing functionality levels today with those reported at the end of PRM-funded projects, conclusions can be drawn on how resilient to deterioration aid results are or, conversely, how precipitously they erode. Based on the findings, a time can be estimated between the end of an aid project and post-delivery deterioration to minimally acceptable values, e.g., those established by SPHERE minimal standards. This knowledge sets the time frame for transition from relief to development. If deterioration occurs too soon, construction quality and maintenance issues are likely to be at hand. It may be surmised that the relief agencies should strengthen their quality control systems, or more attention should be paid to community mobilization to ensure proper maintenance of aid results. Based on this logic, functionality of watsan facilities and the agenda of quality and durability of aid were closely interrelated in the field assessment summarized further in this subsection.
The evaluation included the following four sample-based assessments:
· Survey of protected freestanding water points;
· Assessment of gravity-fed water systems;
· Assessment of community participation;
· Assessment of sanitation facilities in schools and health centers.
a. Survey of Water Points, Gravity-fed Water Systems, and Community-based Maintenance:
The results of the water tests and field observations provide reason to conclude that the PRM-funded effort in the water sector has resulted in the improved quality and supply of drinking water.
The quality of water proved to be a lasting benefit for at least 1-2 post-project years, as was identified in a somewhat undemanding test, performed in the dry season. As to the supply of water, it has diminished visibly both in the standalone water sources and in the water lines.
Errors at the pre-construction stage are unlikely: A highly qualified foreign-educated water engineer in the employ of IRC conducted feasibility assessments, developed technical specifications, and managed a competitive bidding process to procure construction services. An IRC regional water engineer provided senior-level oversight and signed off on system design. It can be surmised that construction process has been controlled rigorously: there were instances of a dismissed construction crew and even a demolished water object, poorly built. In summary, this is the case when a properly staffed and managed process speaks for an imminently positive outcome.
The observed problems with water infrastructure occur at the post-project stage due to inadequate maintenance, bordering on neglect. Neither local authorities nor communities developed sufficient vesting in the water systems. Water as a public good was not internalized by those who stood to benefit from it. Part of the problem, it seems, is that communities of high return are in flux. Intensive repatriation unsettled, in a way, even the incumbents. The flood of new people was used as an excuse for a broken water tap, for a stolen padlock, for rows at a tap stand. The hands-off attitude in a member of the local Water Board could be attributed in part to her sense of alienation from the commune that she felt, was out of control, ‘invaded’ by the repatriates. A massive presence of international aid projects must have added to the unsettling sentiment in the community members and their laissez-faire view of the problems on the premise that ‘we need to figure out where it all goes, before we try to get back in control of local life here’.
The overall conclusion is that water projects have been producing ‘diminishing motivational returns’ from the beneficiary population. Less appreciation means less care and less sustainability of aid results. Possible solutions are outlined in the ‘Recommendations’ subsection of this report.
b. Assessment of Sanitation, Based in Schools and Health Centers:
Upgrading sanitation from the hardly existent to the compliant with SPHERE minimal standards was an effort, integral to all the pertinent projects of reintegration assistance. Sanitation facilities have been observed in transit centers, near assisted houses, schools, health centers, and a local market. As part of their school construction and rehabilitation effort, UNHCR jointly with UNICEF and IRC built 356 latrines for school students and 96 for school staff (UNHCR Database). Pit latrines, built under PRM-funded grants, were assessed in Gacu primary school (Rukungere Colline / Giteranyi Commune / Muyinga Province), Nkero primary school (Nkero / Gisuru / Ruyigi), and at two health centers at the end-stage of construction as of the evaluation visit date: Munagano health center (Munagano / Munagano / Muyinga) and Rimbo health center (Rimbo / Nyanza Lac / Makamba). Pit latrines constructed by UNHCR partners were observed at Masama primary school (Bweru /Kirambi / Ruyigi) and Kirambi Health Center (Kirambi / Bweru / Ruyigi).
Working latrines contribute significantly to the development of modern hygiene habits in school children and controlling diarrheic diseases. Contrariwise, dysfunctional latrines exacerbate environmental health risks. The line between the functional and the dysfunctional in school-based sanitation is drawn at the end of the first capacity utilization cycle: once latrines were filled up in Gacu they became useless and dangerous. In Masama and Nkero, the problem has not yet materialized. Thanks to more favorable student/latrine ratios, the time to first emptying has been extended. Neither of the two schools has determined how they will deal with the problem of emptying their latrines. All the thinking is directed at finding an NGO that would come to help. The evaluators’ encouragement to seek support at the local commune administration did not seem to have much traction.
Health center-based sanitation was examined in two soon-to-be-open health centers, started by construction on PRM funding in 2006: Munagano health center in Muyinga province, almost built by World Vision and Rimbo health center in Makamba province almost built by World Relief. Adequate design, sufficient supply (user/latrine ratio) and reasonable quality of construction support the conclusion that sanitation in those facilities was not left out. Sanitation was also observed in two functioning health centers, in Nyarunazi health center in Muyinga province and Kirambi health center in Ruyigi province. Both were built under UNHCR programs. Latrines are of standard design; and adequately maintained.
As a general conclusion, health center-based sanitation works much better than sanitation in primary schools. In primary schools, the term ‘heavy duty’ seems too inexpressive to render the pressure on relatively few pit latrines, expected to serve hundreds of children during a 15-minute break.
3.5.6 Education
Gacu primary school in Muyinga province and Masama primary school in Ruyigi province were assessed from the following two angles: (i) their resources and performance as educational institutions in the context of intensive repatriation; and (ii) the quality of their construction / rehabilitation under PRM and UNHCR funding.UNHCR and bilateral donors, including PRM, invested significantly in strengthening the primary education system of Burundi. By the time UNHCR withdrew from active assistance to the social service sectors in 2006, they have built 44 primary schools with 245 classrooms and rehabilitated 49 classrooms; constructed 30 school cafeterias; and 126 two-room houses for school teachers. School materials were distributed to returnees as part of the UNHCR- sponsored ‘school kits’. Intensive language and cultural training was offered to the repatriate children who had difficulty with Kirundi and French. UNHCR funded visits of expert missions from Burundi Ministry of Education to the camps in Tanzania (UNHCR, 2008b: 171). The purpose was to facilitate the repatriate schoolchildren’s transition to Burundian schools and maximize continuity in educational standards and experience.
The UNHCR-led effort to expand capacity of the primary school network in Burundi was extremely useful. However, the need for additional class space in primary schools was met at 77 percent, according to the evaluators’ estimation. The acute shortage of classroom space was accompanied by the government’s slow response with scaling up staffing and operating funding for the primary school network. The overall conclusion is that primary schools were unable to absorb intensive repatriation. Reintegration in the area of primary school education happened in a rather dissatisfactory way: the school network and resources sagged under the demographic pressure of repatriation; and both the repatriate and the incumbent children shared the consequence in the high-return areas of Burundi.
One of 44 UNHCR-built schools and the school, partially rehabilitated and expanded on PRM funding, were assessed for aid survivability. Durable results depend largely on the quality of construction, at least in the first few post-construction years. Both schools were built in 2006: Masama school was constructed by CADI, a Burundian partner of UNHCR; Gacu school was constructed by World vision on PRM funding. Site observations revealed a highly uneven quality of construction. Quality-related issues are explained by the following factors:
· Construction technology: The observed problems of cracking walls, crumbling pavement, and degrading window seals and door blocks are rooted in substandard concrete. The technology of on-site concrete production appears to be compromised due to violation of the cement-sand ratio; builders’ failure to use sieves to improve the quality of sand; and non-compliance with the water-cement ratio in the absence of power concrete mixers.
· Coordination and planning: Lack of coordination and/or attention to comprehensive design and planning resulted in the dysfunctional school cafeteria in Masama school: built but never matched to drinking water.
· External Quality Control: Customer feedback as an important quality control circuit was disabled in the observed case of Gacu school.
· Management objectives: There is an opinion at least in some NGOs that the relief aid, provided under cooperative agreements, stipulates an intent rather than an obligation to achieve a result. This self-excusing attitude diffuses management focus on goal achievement and quality control.The UNHCR- and PRM-sponsored construction and rehabilitation of primary schools and the accompanying measures in support of primary education have eased but not resolved the acute shortage of capacity and resources that continues to beleaguer the education sector of Burundi. When assistance cannot be scaled up to demand it is particularly important that limited aid resources are not used on building facilities that start crumbling within three months after the ribbon is cut. There is a relief outcome even in a poorly built school when the alternative option is no school at all. However, the effectiveness of such relief aid is short-lived and development spin that it produces is none. On the contrary, quality construction leads to durable results. In that case, relief becomes an insipient point of development.
3.5.7 Health
Two health centers (HCs), rehabilitated on UNHCR funding[5] and two health centers constructed with PRM funding were visited during this evaluation. UNHCR-sponsored health centers were surveyed using a health center survey tool, designed for this study. PRM-sponsored health centers have not been opened yet. Their configuration, construction experience, and development context were examined.Both functioning HCs are well supplied with pharmaceuticals; provide children’s immunizations at high coverage rates despite a rudimentary cold chain; almost never do lab tests; and provide most of the women’s and children’s care expected from a primary HC. Patient care is focused on office-based services. Outreach is limited to community sensitization and entrusted to a team of community volunteers, referred to as health agents. Population in the catchment area exceeds HC capacity by 1.8-2 times. Health outcomes are not monitored in the local population because health professionals are not involved in death registration and do not provide any outreach follow-up on the episodes of care initiated in the health center. Office-based health care is managed in a ‘revolving door’ mode, seen in many parts of the developing world: patients come to see a health worker; get prescription, pay for it, unless exempt; receive medication, and are sent home.
The government did not respond to repatriation with increased financing of primary health care. As a result, the returnees’ entitlement to three months of free services has widened the gap of unreimbursed care in HCs’ budgets. Here is the sentiment of a provider of care: “Nearby, we have a peace village with a lot of poor people who cannot pay. We spend medications on them and nobody repays the cost. Our stock is depleted by this flood of indigenous patients” (ERv2: 197).
Even if returnees are prepared to pay, they are likely to be disappointed by the quality of care: with the patient population gone significantly up, HCs are now more short-handed than before and are experiencing an across-the-board shortage of resources. Similar to the education sector, reintegration in health occurred by putting repatriates on equal footing with residents only to see both populations worse off.
Given that two HCs have been delayed by construction for two years and one of them (in Munagano) because of a non-compliant design, there seems to be a need for a more result-oriented approach to grant implementation management. Pertinent ideas are outlined in the report’s ‘Recommendations’ section.
The second HC that is not open yet is located in Rimbo, Makamba province. It is configured to standard requirements of the Burundi health authorities; well designed and built, and has access to a water line unlike Munagano HC. In an attempt to answer the question why health centers are of uneven quality, even when financed by international organizations with shared objectives and ideology of refugee assistance, the evaluators looked into the process-related aspects of the PRM/World Relief grant. The following findings are relevant for understanding the success of a particular construction project but, more importantly, they provide an insight into the underpinnings of durable aid results in general. To summarize the findings, World Relief used a PRM grant to bundle HC construction with environmental health, shelter, and livelihood strengthening assistance. By integrating interventions in such a holistic way, World Relief were able to capture interrelated human needs at the individual, household and community levels. The role of health as an important contributor to reintegration was properly understood, and so were the roles of improved house design and small husbandry as contributors to health. By engaging the community in a truly participatory process, WR developed accurate knowledge of community needs and rallied grassroots support for program interventions. By establishing dialogue with the national and province governments, WR shifted ownership to, and developed vesting in those who control resources. By building well and getting people involved in the process, WR mobilized social support for Rimbo HC that will contribute to its sustainable operations. Given the high percent share of returnees in Nyanza Lac commune, the WR experience is strictly within the boundaries of humanitarian assistance to repatriation. At the same time, the WR’s strategic and comprehensive approach to livelihood strengthening in Rimbo, places their achievement firmly in the range of advanced development. Thus, relief and development are neither opposing nor sequential concepts: they can be achieved by, and within one program. By proving that this is possible, WR has qualified for a best-practice project, in the evaluators’ opinion.
3.5.8 Protection of Safety and Human Rights
The dual purpose of post-return monitoring in Burundi is (i) to provide an on-going “assessment of the degree of returnees’ reintegration in the Burundian society” (UNHCR/Burundi, 2008: 2), and (ii) “to assure adaptive protection for the repatriates, compliant with international standards” (UNHCR/Burundi, 2007: 1). Monitoring at the reintegration stage is a logical continuation of the protection agenda that precedes and accompanies the return.The monitoring information is used for protection planning and case-based follow-up. The follow-up is focused, primarily on feeding case information to appropriate implementing partners so that they could address post-return problems pertinent to their areas of reintegration support. Monitoring data informs and guides the following assistance areas: vulnerability identification, land conflict mediation and legal support, facilitation of national ID cards, and post-return assistance to unaccompanied minors and separated children. The protection program, thus, does not represent a separate line of assistance to beneficiaries. Instead, it creates an additional information loop that ties other areas of reintegration assistance into a more coordinated and effective whole.
The post-return monitoring program of UNHCR/Burundi relies on two valuable assets, created over several years of perseverant effort: a nationwide interviewer network and a well established practice of data analysis. With these assets in hand, the post-return monitoring program can qualify for an international best practice in a matter of months, (i) if transition is made to a more rigorously designed population sample; (ii) a data management program is upgraded to serve a more diverse set of analyses; and (iii) a funding partner is identified to extend sample-based monitoring to the incumbent population. What at the relief stage is protection monitoring, at the development stage should evolve to a Current Population Survey (borrowing from the U.S. Bureau of the Census terminology). These ideas are elaborated in the report’s ‘Recommendations’ section.
3.6. Prospects for Durable Return
This section presents analyses to answer the final evaluation question: “Will return be durable?” – The previous analyses have shown that reintegration was achieved as an outcome of the return and post-return assistance. However, it is yet to be consolidated in a non-reversible achievement to become an impact. “The reintegration is far from over. We must monitor very carefully. In the worst case scenario, if we do not accommodate people’s needs, the repatriates may leave again. Durable return is the main goal, yet the hardest to achieve. It certainly has not been achieved yet” (CA, 2008: 114-5349).The following sustainability factors will define the prospects for irreversible reintegration and durable return:
· Survivability of aid results. The results of reintegration assistance must be durable enough to survive until handover to development. The painstaking analysis in the previous sections of this report has shown that assistance results tend to deteriorate over time at a pace that varies by assistance area. Processes that drive deterioration were reviewed. There is enough reason to conclude that relief aid is not immune to decay. The pace of decay is not catastrophic but does speak in favour of a quick transition to development. The survivability question is thus transformed into the relief-to-development question: ‘Is development effort up to speed or at least in sight to sustain the reintegration momentum developed by UNHCR and bilateral donors at the relief stage?’
· Peace and stability. Of overriding importance is political future of Burundi. Relapse to instability is likely to disrupt further reintegration. It also may trigger a new wave of forced displacement, thus, undoing even the return. Development programs are designed for several years and, therefore, are sensitive to medium-term political risks. The development stage will be hindered by the lack of clarity about the country prospects for the next 3 to 5 years.
· Government and community strengthening. Sustainable reintegration is crucially linked to the willingness and capacity of the state to reassume responsibility for the rights and well-being of its citizens (UNHCR, 2008: 4). Lack of public leadership and control, and community participation will render development largely ineffective.
The levels of these risks to sustainability were reviewed in four stakeholder opinion scans that were systematized and mapped on an 5-point scale from ‘very pessimistic (1) to ‘very optimistic’ (5).
a. Nationwide Prospects for Peace and Stability:
Stakeholder opinions about nationwide prospects for peace and stability are mostly in the range of pessimistic to neutral. The prevalent view is that the presidential elections of 2010 will be the watershed event that will either trigger a relapse to instability or will allow Burundi to continue on a path to normalcy. There is a widespread mood of troubling uncertainty as to what will happen after the elections. One optimistic opinion emphasizes the nascent consumer confidence as an indication of future progress towards a more stable environment in Burundi.
b. Community-based Prospects for Peace and Stability: Land
At the community level the key risks to social cohesion are land-related conflict and certain divisive effects that reintegration aid programs have produced in the regions of high return.
Land is, by far, the most contentious issue in the daily life of the communities of high return. Land-related conflict is endemic. Acuity of the land issue varies by region: informants consistently highlight Makamba as the most contentious province. The current practice of land conflict resolution are ineffectual: mediation programs do not cover all disputes; close a relatively small number of cases; and render non-binding decisions. As a result, stakeholders in the seemingly resolved disputes remain dissatisfied and in a potentially vindictive mood. Government and quasi-government organizations tend to be more optimistic in their views of the situation, compared to the international community of UN agencies, donors and implementing partners, who rendered their opinions in an alarmist key. All stakeholders considered, opinions gravitate towards the sub-range of ‘very pessimistic’ to ‘neutral’.
To validate opinions with a more objective insight, the recourse for conflict resolution was reviewed on a sample of 265 cases, settled amicably or resolved by CNTB decision over the second quarter of 2008 Ten settlement methods were identified and grouped by notional level of discontent (‘grudge level’) that those methods are likely to produce in any or both of the disputants. Sample-based estimations have led to the following findings:
· In 48 percent of the settled disputes (‘grudge level’ =3), land was taken away from both stakeholders or remained /was transferred in the hands of one of them. Current public discourse, suggesting that land-conflict resolution is all about sharing land, strayed from the point: almost half of the reviewed settlements did not involve sharing at all.
· The high- and medium-grudge cases account for 70 percent of all the reviewed settlements. These are ‘zero-sum’ settlements: one disputant gets only what the other one looses. Discontent is inherent in such cases. Notably, the game-theory term ‘zero-sum game’ has the synonym ‘conflict game’.
· Presence or absence of compensation summarizes the difference between the low and high levels of discontent. This difference may not necessarily be dichotomous. If the amount of compensation is considered inadequate by the recipient, the settlement is likely to be perceived dissatisfactory as well. In the reviewed case sample, the amount of compensations gravitated toward the range of BFr 15,000 – 50,000, which seems low. Plausibly, a measurable part of the ‘low-grudge’ settlements would be less benign in the perception of the disputants then in the proposed classification. This implies that ‘high-grudge’ cases account for a majority of the settled disputes.
· A sample-wide ‘grudge index’ has been estimated at 2.18. This means that the amount of discontent is between medium and high, closer to medium. Importantly, the evaluators did not have a case sample from Makamba. In all likelihood, the inclusion of that highly-contentious province would increase the ‘grudge index’ value.
UNHCR has made a significant intellectual and monetary investment in helping Burundi to break through the confines of the zero-sum approach. Two lines of UNHCR’s effort are important in this regard: (i) Cash grant; (ii) Peace villages. The cash grant program brings money ‘to the table’ while peace villages bring land. Both strategies thus inject resources and enable Pareto-efficient solutions to the problem: making returnees better-off without making incumbents worse-off.
To enable faster progress, UNHCR invested in the government’s plan to develop peace villages as a conduit to new land and sustainable habitat. Up to 15 peace villages are envisaged. A separate ‘subject node’ was developed in the content analysis to scan stakeholder opinions about peace villages.
The concept remains elusive. All the interviewed tend to put a utopian spin on the peace village as a Burundian version of the ‘City of the Sun’ (the latter was also placed in a sub-equatorial space): a place where the vulnerable are well protected, ‘private property is acquired and improved’, and ‘agriculture is much followed’. The returnees will live together with the former IDPs and the incumbents; and people of various trades will share their skills for the common benefit. A popular expectation of the peace village is the creation of an ambiance opposite to that of a refugee camp: people of different walks of life will break the psycho-emotional confines of refugee’s former existence with the diversity of their experience and the shared interest in the free and peaceful life renewed. Expectations to the contrary have been heard as well: a peace village, with its fairly regimented space and organization, will provide a comfortable déjà-vu for the ex-refugees, whose perception of social hierarchy and success are linked to the eco-social space of the refugee camp[6].
The evaluators have visited two peace villages. Nyarunazi peace village in Buhinyuza commune, Muyinga province was the first one built under the ‘villagization’ initiatives. In the local account of the experience of this very first peace village, the conceptual centerpiece of villagization (to endow people with land and a self-sufficient habitat) was lost in implementation. Land was not promptly allocated to the 300 families of Nyarunazi village. Nine months later, in September 2008, the people of Nyarunazi were in a standoff with the local authorities: de facto landless yet reluctant to accept infertile land offered to them in the vicinity of the village. Government officials, on their part, are reluctant to let people farm on land that is more to their liking but is located further out. The government’s concern is that once people get land away from the village, they will abandon the village to live close to their land.
The emerging signs of bureaucratic pressure threaten to turn a peace village into its opposite. It is unclear whether ‘villagization’ has prospects to succeed in Burundi or will join the fate of similar pieces of social engineering tried in various parts of the world in recent decades – strong in concept and flawed in practice.
To be fair, peace villages as any complex innovation are entitled to a period of trial and error. Much will depend on the government’s ability for experiential learning. Nyarunazi can be seen as a productive pilot.
Gitara peace village in Makamba province is under construction. It helps to visit a nearby temporary lodging facility to understand the pragmatic side of villagization: most of the 500+ old caseload returnees have no place to set foot in Burundi. They are ‘locked up’ in a secondary transit camp for 3-4 months until their peace village is built. All but one in a random group of 15 temporary lodgers in Gitara said they looked forward to living in the peace village. One old man insisted that he knew where his land used to be and he wanted to go back to his land instead. This man will probably join “grudge level 3”. The rest may or may not be happy in their peace village but will use it as a good starting point for reintegrating with the country that they have not been to for almost 40 years or never at all.
Rather than thinking of peace villages in the somewhat esoteric terms of a sustainable habitat and an antipode to refugee camps, the evaluators are inclined to strip the concept to its robust core: peace village is an ex-refugee camp. However, this time, the gates are open and the freedom to move in, out and around is granted and encouraged. A peace village buys important time for the landless returnees to get started in their home country, to look around, to prepare for, and respond to an opportunity once it comes to those gates. Unless discredited by bureaucratic incompetence, villagization has the potential to inhibit an accumulation of discontent around the contentious issue of land in Burundi. Peace villages are a peace-building tool but the devil is, and will continue to be in implementation.
In the meantime, the first timid steps have been taken by the international community to come to terms with the fundamental need for a publicly funded compensation of landless people in Burundi. In May 2008, the Peace Building Commission made the following statement: “For Burundi, the modalities for compensation remain to be elaborated by the various stakeholders, including the time-frame for receiving claims. Compensation can include monetary and non-monetary settlements, such as training” (PBC/Burundi, 2008). Unassuming as this statement is, it may be heralding the next and more far-reaching stage in building the economic foundation of peace and stability in Burundi.
4. Summary Conclusions
a. Repatriation:The benefits provided under the assisted return program in Burundi have ensured comprehensive physical, legal, and economic protection of Burundian refugees during their repatriation to the home country. These protections have been well designed, sensitively adjusted to the evolving need, and rigorously executed. The beginning of the assisted return in 2003-4 was afflicted with organizational problems, due, primarily, to the delayed deployment of UNHCR resources in the areas of high return. In 2005-8, the assisted return program attained its maturity and consistently delivered on expectations of the repatriates, the Government of Burundi, and UNHCR’s donors alike. If there is a succinct way to define the UNHCR-led program of assisted return to Burundi, it would be ‘evidence-based pragmatism’ – a set of conceptual and operational features that puts the program in the range of best-practice responses to refugee crises and humanitarian crises in general.
b. Reintegration:
Reintegration was evaluated on a set of technical evidence (levels 1 to 3) and based on perception of its primary beneficiaries – the returnees (level 4).
The technical side of reintegration has been examined at three levels: (i) econometric estimation of the difference among the repatriate and incumbent families by an integrated measure of socio-economic achievement – the welfare index; (ii) estimation of variance and similarity between the two population groups in each welfare component; (iii) estimation of the likelihood of reintegration based on the amount and reintegration effects of the post-return assistance. The econometric model produced an unambiguous finding that the repatriates and the incumbents are indistinguishable by welfare level at the present time. The ‘present time’ in the context of this analysis is defined by the average post-return experience of 4.2 years. If there was a welfare gap immediately after return, it disappeared by now -- 4.2 years after return. The second round of analysis examined the comparative standings of the repatriates and the incumbents by their current assets (consumption and savings), their durable physical assets (housing), their human assets (education and health), and their safety. The finding from this study corroborates the econometric output: the repatriates and the incumbents possess assets at very similar levels. There are factors, for example family size, that determine living standards of Burundian households at a statistically significant level. Ex-refugee status is not among such factors. The intuitively defined near-term purpose of refugee assistance – to put returnees on a par with the resident population of Burundi – has been achieved for an average repatriate family as of September 2008. The analysis of reintegration benefits in seven sectors of assistance – food, cash, land, shelter, water and sanitation, education and health – leads to the conclusion that assistance played an important role in the achievement of reintegration. Reintegration assistance was designed, scaled, timed, and delivered in response to key material and social needs and effectively enough to make the repatriates themselves to acknowledge that, ‘Yes, we benefited from our returnee status in Burundi’.
The importance of the material aspects of reintegration has been put in perspective by the repatriates themselves. They expressed strong satisfaction with their decision to return despite the fact that their current well-being is below pre-return expectations, let alone their pre-return and pre-displacement living standards. The reviewed evidence from the household survey suggests that the return and reintegration have been a profoundly self-motivated process in Burundi.
c. Sustainability:
Given the returnees’ deeply-felt desire to be back to their home country despite all the post-return hardship, can UNHCR and PRM conclude that their efforts are bound to succeed in Burundi, if minimally effective? – For the short term, the likely answer is ‘Yes’. Long-term, there is not enough evidence to conclude with confidence that the near-complete repatriation marks an irreversible end to forced displacement. There are factors that support sustainability yet have not been put to work at full strength. There are also negative factors that have not been entirely curbed. The main areas of concern include a latent discontent over uncompensated redistribution of land; a divisive role of aid when it is excessively targeted to repatriates beyond the first two-three years; a partial loss of momentum in the relief-to-development transition; and lack of confident projections of post-election policy trends.
The land issue is central to the near- and long-term prospects for sustainable reintegration. The overcrowding on the currently farmed land, compounded by the legacy of mismanaged land ownership rights and lack of progress toward overdue reforms in the agrarian sector makes it difficult to endow repatriates with land – a source of complications at the current stage of the assisted return program and, potentially, a stumbling block on the road to sustainable reintegration.
5. Recommendations
The following recommendations are grouped by assistance sector and areas of opportunity identified in this study. Most recommendations are addressed to one or several ‘action agencies’, or a broadly defined category of stakeholder organizations. The authors of this report evaluated program results and prospects for those results to be sustained. The knowledge developed in this study substantiates recommendations as to what can be done to maximize program outcomes, slow down their erosion and prevent their reversal at the post-program stage. Who should act on these recommendations is a separate matter that was not the subject of this evaluation. Therefore, while the ‘what’ recommendations are defensible with the evidence developed in this study, the ‘who’ recommendations are speculative. A competent selection of addressees will require a thorough assessment of the stakeholder mandates and capacities, as well as the norms of international law and practice -- a useful follow-up task but not a part of the reported evaluation.
a. Repatriation:
As long as the international community is consistent in their adherence to the principle of voluntary return, they cannot control the beginning or pace of repatriation: refugees are entitled to move back to their home country when they wish. In doing so, they have the right to accept or bypass repatriation assistance. This, basically, puts UNHCR in a reactive mode: they initiate repatriation assistance after they see that repatriation has gotten underway. The catch-up period can be minimized with accurate prediction of, and preparation for the start of the repatriation. However, if refugees choose to return to a relatively unsafe environment, UNHCR’s response is bound to be delayed by the lack of security clearance.
To the extent the pace of repatriation can be influenced, it should be influenced with consideration for the host country’s absorption capacity and asylum country’s policy to end the refugee presence on their territory. The ‘absorption capacity’ factor must be decomposed into such variables as (i) the host country’s ability to provide returnees with fixed assets (land and/or producer’s durable equipment); and (ii) the degree to which social cohesion can withstand increased competition for limited resources. The design and empirical testing of a push-pull equilibrium model is a technically viable but impractical task: immediate political agendas dictate the pace of repatriation more than theoretical optima.
Of practical importance is a continuous policy dialogue or, more to the point, an on-going tri-partite bargaining to correct both for an excessively rushed and an unreasonably slow pace of repatriation. With no knowledge to what extent this recommendation is already part of the practice or politically appropriate at all, the evaluators nevertheless recommend that major donors help UNHCR negotiate a realistic pace of repatriation with the asylum and the home country governments; and, importantly, insist on close coordination between UNHCR offices in both countries. In the case of Burundi, the recommendation would be addressed to all UNHCR donors and, primarily, to PRM and ECHO as the leading donor agencies. The evaluation has evidenced that ECHO actively participated in the tri-partite coordination and closely monitored the situation in Tanzania. This was achieved by two organizational means: (i) by placing the Burundi and Tanzania operations of ECHO under the responsibility of a single representative (‘Technical Assistant’), and (ii) by linking country planning and operations to a continuous feed of evidence from first-hand observations in Tanzania and the high-return areas of Burundi. It is recommended that PRM study the ECHO model of refugee assistance to see if it might be of use to enable a more active engagement with the return planning process on both sides of the border.
b. Reintegration:
b1. Housing Construction:
1. While the UNHCR operation is designed with economies of scale in mind and therefore focused on the provinces with the largest numbers of repatriates, niche programs could be implemented to address small pockets of need in provinces where the number of returnee families is not that large, yet does not deserve to be neglected. It is recommended that PRM consider this and next recommendation when planning their housing grants in Burundi.
2. Two targeting strategies are recommended to combine the ‘As Many As Possible per Province’ policy with the ‘Equal Chance’ policy of geographic targeting. The first targeting policy takes into account the province-specific numbers of people not yet provided with housing assistance: the larger pool of yet unassisted returnees is likely to draw more attention and assistance. The second ‘targeting formula’ adjusts the first one by multiplying the number of people in need by the number of months of construction it will take to reach those people with assistance at the present pace of construction. The proposed ‘time-weighted need’ strategy emphasizes provinces that, despite relatively few returnees, will have to wait longer, because the amount of assistance allocated to them is negligibly low. Provinces with a larger target population and, at the same time, highly preferential treatment in the previous years, will be somewhat de-emphasized. Based on this strategy, geographic priorities can be flexibly adjusted on a yearly basis, particularly, when the time comes to prioritize persistently neglected small targets after big ones were sufficiently addressed. Table 3.39 of the Complete Report provides ‘priority indices’ that quantify this approach to targeting.
3. UNHCR and bilateral donors are advised to exercise extreme caution when changing their funding targets in ways that void program eligibility for previously waitlisted Burundian families.
4. The targeting policy of program administrators should be focused, the evaluators believe, on how to provide one house per eligible family on program-defined terms and in an expeditious manner, and less on how this house would be adapted for family needs and economic strategies. If the benefits of assistance end up used in ways unpredicted, the architects of assistance should not necessarily feel disappointed: the economic rationality specific to the host country, just performed a reality check on program design and adapted it accordingly.
5. UNHCR’s mandate and limited resources restrict overall coverage and specific targeting policy to the defined population of concern – Burundian repatriates. Development programs must get up to speed in Burundi to address the resulting micro-geographic variation in assistance levels.
6. Bilateral programs may want to start putting resources into a customizable house design and construction assistance. To do this in the true spirit of development, they should start engaging user charges by encouraging people to pay for upgrades. This will introduce some operational complexity but also will help to turn housing assistance into a more rational and sustainable endeavor. In practical terms, PRM and Habitat for Humanity have funded housing assistance through World Relief on opposite policy principles: PRM to provide free houses to the repatriate families; Habitat for Humanity to do the same on the basis of cost recovery. Perhaps, both donors could coordinate their funding to develop a middle-ground approach, whereby their resources would be pooled to enable assisted housing construction with partial cost recovery for the additional cost of customization.
7. It is hoped that reforestation, precipitated by assisted housing construction, will continue to be addressed in order to remedy the growing pressure on the environment. The environmental factor of reintegration is not actively considered as part of the relief agenda. At the least, relief donors should assume the role of conscientious environmentalists when helping development agencies design their strategy and programs for Burundi. PRM may want to discuss the environmental priorities of rural Burundi with USAID.
b2. Water and Sanitation:
8. Further assistance in water and sanitation should be linked to geographic need, as the primary targeting criterion. So far, the need was secondary to the number and share of the returnee population in a given area. The neediest communes were chosen from a short list of those with the highest concentration of repatriates. It is now time to start selecting communes based on comparative need and regardless of who lives there. This shift in the targeting policy would be easier for development donor agencies to accomplish, since they are not bound by relief mandates that limit their priorities to specific populations of concern. If PRM cannot follow up on this recommendation because of their refugee-centered approach under MRA, it is recommended that they establish a coordinative process with USAID. The recommendation to USAID would be to work in low-return communes of high-return provinces to offset the contrasts in the levels of assistance between contiguous micro-geographic areas.
9. The evaluation identified an important need for teaching people disposal and excreta management practices as aggressively as it has been done for water management practices. It is recommended that PRM (re)target their grants in sanitation and community strengthening to emphasize the importance of public health education in these respective areas.
10. Water source protection can be strengthened by ameliorating the area adjacent to the source. Wherever appropriate, the following additions to a standard technology are worth consideration: (i) digging a trench across the slope to stop erosion and intercept the downstreaming surface water in order to reduce the risk of re-contamination of the water source; (ii) planting long-root grass or perennial plants to stabilize topsoil around the water source; (iii) where appropriate, building a fence to prevent random unhygienic activity in proximity to the water source; (iv) stricter adhearance to the rule of providing cemented area for laundry and bathing. PRM’s Cooperative Agreements should contain more specific requirements as to the technology and results of water source protection and construction/rehabilitation of tap stands connected to water lines.
11. The strategic and technical void is apparent in matching water assistance to local lifestyles: water supply has been geared to people’s houses, while most families spend their days in the field. Quoting from a Burundian water expert, “The same field is used as the rest room even if the worker is not very far from his home. He will drink raw water from a nearby stream or river even if he has drinking water at home”. The problem has been confirmed in SPHERE guidelines, yet no solution or discussion agenda proposed either in Burundi nor, it seems, internationally. PRM may want to call upon their trusted partners in the environmental health assistance to initiate a discussion of, and practical response to the issue of matching watsan to the workplace environment of rural Burundi.
12. The baseline and end-of-project measurements of microbiological content are essential for assessing the achievement of basic project results in the watsan sector of assistance. It is recommended that the requirement and basic specifications of such tests be included in Cooperative Agreements as a standard clause. So far, they were performed at partners’ discretion.
13. To ensure objectivity, it is strongly recommended that pre/post tests be performed in the same month and, importantly, during the rainy season, preferably when environmental health risks are known to be at their highest. In Burundi, the recommended months for water testing are October-November. This is another recommended addition to the PRM’s CA template.
14. To enable ‘legacy’ evaluations of past projects (e.g., to assess survivability of aid results), it is recommended that standard requirements for aid site identification be included in PRM’s Cooperative Agreements. Each watsan site should be specified as follows: (i) description of the exact geographic location on the administrative map of Burundi (in the case of Burundi: Sous-colline, Colline, Zone, Commune, Province); (ii) the closest identifiable reference point; (iii) two jpeg image files with photographic pictures of the site: one up close, another one in a topographic setting; with month, date, time, and orientation specified in the ‘picture summary’ section of the pic file properties window; (iv) contact information of a local official and organization (in the case of Burundi, a member of the local commune administration) who was a focal point for the project at the end of activities in a given aid site.
15. Undoubtedly, the biggest problem both in water and in sanitation is maintenance. Further watsan investment is becoming counterproductive. An effort to produce more positive results by implementing new projects is increasingly offset by the crumbling of the results of the previous ones. The following three recommendations respond to the identified problem:
16. A palliative solution for PRM would be to deploy a ‘maintenance’ grant that would encourage an implementing organization to revisit their past project sites or engage with new sites, where assistance was provided by other NGOs. The objective is to address a plethora of maintenance problems and restore functionality of the previously built water and sanitation systems to an optimal level. An important objective would be to investigate causes of failure and provide empirical material for a survivability analysis to inform watsan interventions in the future. Additional rationale for a maintenance grant is outlined in Recommendation 46.
17. A more far-reaching solution that is strongly recommended to PRM and/or USAID is to promote professionalization and privatization of watsan maintenance. Rural Burundi after years of intensive watsan assistance looks like one big maintenance job for a skillful and motivated plumber. Operating a pump truck presents a ‘sure-shot’ business opportunity also. Professionalization may precede privatization or, more realistically, become an incipient point of privatization. The immediate objective is to substitute an independent contractor for a team of ineffectual water technicians in the employ of commune water administrations (three per commune in the case of Gisuru/Ruyigi). Public and private contracts will serve as a vehicle for stronger incentives, operational autonomy, and responsibility for quality and end results. Expected performance gains are outlined in the report’s full version. Start-up assistance may include access to capital and training, and a push with work orders.
18. Relief aid should be scaled to the prospective population rather than the current population in the communes of high return. The observed decline in water supply in Gisuru occurred in great measure because of an influx of repatriates at the post-project stage and resulting over-utilization of the tap stands. In the visited primary schools, the inadequate student/latrine ratios have resulted from negligence or errors at the planning stage, compounded by the rapidly expanded school enrollment at the height of repatriation. UNHCR data have sufficient predictive capacity that grant applicants can use to project the number of returnees by destination area. It is recommended that PRM request such projections from RFA responders as part of their pre-application need assessment. A forward-looking approach to planning watsan interventions will make systems and facilities more resilient to demographic pressure.
b3. Construction of Schools:
19. Dependent on how aggressive grantors want to be in strengthening the quality of construction under their projects, a range of measures can be taken. Some of them are outlined below:
20. UNDP and UNHCR may consider joining efforts with the Government of Burundi and bilateral donors to create a national register of pre-approved construction organizations and self-employed teams. A credentialing process would include as follows: (i) a questionnaire would be sent to NGOs with prior construction experience in Burundi, asking them to list their construction crews from past and current projects and rate the quality of their work; (ii) quality standards would be defined in Good Construction Practices to be developed in coordination with the policies and guidelines existing in Burundi; (iii) additional training would be provided to pre-selected building teams; (iv) Their written commitment to compliance with professional standards would be elicited; (v) terms and conditions of periodic renewal of accreditation would be determined.
21. UNDP and UNHCR may consider facilitating business start-ups in the construction industry and, as a matter of priority, a start-up of a few province-level organizations that would rent or lease construction equipment to NGOs or subcontracting builders for the duration of their projects. Fuel-powered concrete mixers, and pick-up trucks illustrate the range of technology that could raise labor productivity and improve quality of construction works.
22. PRM would include preferred choice of a pre-approved construction team as a standard requirement in the cooperative agreement template. Even if the proposed national accreditation system is not set up, PRM can request that grant applicants detail their search and selection strategy in their grant applications. This would increase their attention to the selection of qualified builders.
23. Customer participation in the handoff process at the end of the construction project and other types of feedback should be encouraged as a good practice of grant implementation. The presence and content of such feedback may be examined during Interim Program Evaluation and judged as a criterion of successful grant performance.
24. A zero-tolerance policy may be considered under PRM grants and communicated to the grantees with respect to quality flaws that jeopardize safety of users in general and children in particular. Cracked school walls, holes in the walls, and missing doors illustrate poor quality that directly impacts on children’s safety and health.
25. As a standard requirement of PRM grant application process and project reporting, all construction sites (as well as any other assistance sites) should be referenced to specific geographic locations.
b4. Construction of Health Centers:
26. To the extent cooperative agreements want to target results, rather than progress towards stated objectives, the application and review process can be strengthened in several aspects. An illustrative set of PRM requirements for grant applications may include as follows:
27. Range of health care to be provided by the intended health center with reference to: (i) valid health sector laws and regulations of the host country; (ii) international care standards for comparable health sector resource settings. More specifically, instead of a generic ‘health center’, World Vision should have been required to propose a facility that meets the MOH of Burundi definition of the primary HC and provides a range of care to a defined patient population that complies or diverges from the WHO guidelines for the first-level HC in such areas as immunization; maternal, newborn and reproductive health (including GBV care); management of common symptoms in children (cough and fever, diarrhea), management of TB, malaria, and provision of general care (e.g., stitching wounds, changing wound dressing, incision of abscess), and so on, as applicable.
28. Health Center location and resource configuration: exact location, catchment area population, physical plant (buildings and structures matched to clinical and support functions), staffing schedule, equipment list; with reference to standard regulatory requirements of the host country.
29. A memorandum of understanding co-signed by Province Health Administration, or a letter of support indicating what, where, and in what configuration will be built.
30. Competitive selection process to ensure skilled workforce for construction.
31. Basic specifications of similar facilities built by the applicant in the host country or other countries.32. This outline could be continued, however, additional requirements may escalate the cost of preparing application, create excessive dependence of a prospective grantee on the local bureaucracies, and discourage the application.
b5. Monitoring for Safety and Protection:
33. The following recommendations may be worth consideration to increase the value added from the UNHCR post-return monitoring program, both its people and its information:
34. Share the network of field interviewers with other evaluation programs and teams (provided their time is reimbursed at cost, plus UNHCR management fee, plus an opportunity cost of their distraction from their main duties in the monitoring program).
35. Initiate a panel study of post-return experience. Contrary to the current logic of avoiding repetitive interviews, a representative cohort of returnees and incumbents would be identified and tracked for months and years, to understand their post-return adaptive strategies and socio-economic evolution. By moving their interviewers to professional employment, UNHCR is likely to have lowered the turnover rate in their monitoring team. A stable team of interviewers is an important asset for a cohort study as it minimizes non-sampling errors that happen, among other reasons, due to lack of uniformity in asking questions and interpreting answers. Cohort assessment is rarely possible due to a short attention span in most organizations. The current set-up in the UNHCR/Burundi monitoring program makes it possible, thus enabling a longitudinal insight into the process of reintegration (tracking reintegration as it happens for a group of families in real time).
36. The new population census, conducted in Burundi in the summer of 2008, has provided the first post-crisis population sampling frame, thus, creating an excellent opportunity for upgrading monitoring to a more high-powered design. If UNHCR uses this new opportunity, they will be able to reduce the volume of their monitoring activity while increasing the explanatory power of the resulting data.
37. Unless Ligue Iteka’s teams can be in all the communes all the time, UNHCR will have to identify certain stratification criteria in order not to leave out areas that are distinctly different according to a specified one or several variables of interest. In the household survey designed for this study, the evaluators decided to sample from the communes with high, medium and low presence of returnees. As a result, they were able to identify certain phenomena that attest to a socio-economic depression among the incumbents, such as their reduced access to good sources of water. This, apparently, happens in the communes that are bypassed by reintegration assistance because they have relatively few repatriates. The share of returnees in the local population is a variable of interest that was used differently in the logical framework of this evaluation compared to its use in the UNHCR monitoring program. UNHCR sends more interviewers to where there are more repatriates. The evaluators’ logic was to sample commensurately from all the three pools of communes in order to understand the part of Burundian life that is not in the spotlight of current international assistance yet is important for capturing the overall progress of reintegration. If returnees are in alignment with the local population, when part of the local population falls out of this alignment, the achievement of reintegration can be questioned. – UNHCR may have other considerations for stratifying its population sample in order to control its size yet make it responsive to their data needs.
38. More analytic output can be produced from the existing monitoring database if cross-tabulation is used more actively. Cross-tabulation implies that a dataset queried from one database field is subset to another data field. In the current evaluation, several examples of cross-tabulation were presented as part of the household survey data analysis. For example, the distribution of returnees by level of satisfaction with being a returnee in Burundi (as a proxy of their satisfaction with post-return assistance) was calculated initially for all the returnees in the sample and then for smaller groups of returnees, formed by sub-period of return. With reference to the UNHCR/Burundi monitoring dataset, it may be interesting, for example, to match the self-reported levels of food security to the number of years elapsed after repatriation, to see how resilience to food shortage changes over time.
39. To fulfil the first of the two previously mentioned objectives of the monitoring program (gauging a ‘degree of returnees’ reintegration in the Burundian society’) it is important that monitoring is extended to the incumbent (resident) population. In 2007, only 39 incumbent families were interviewed among 23,692 families. As long as reintegration is a comparative concept, measuring the socio-economic pulse of the incumbents remains an important part of an ongoing reintegration assessment. Developing this second prong in the UNHCR-led monitoring program could be an honourable undertaking for any bilateral donor or UNDP.
c. Support of Sustainable Reintegration and Durable Return:
40. The land issue is the most important challenge to sustainable reintegration. A strategic response to this challenge requires a system of compensation, accessible to any of the two contending parties in land-related disputes. Compensation should be high enough to incentivize an opt-out solution to a dispute. Either claimant or defendant should be encouraged to relocate and/or take on a path to non-agricultural livelihood. Compensations, therefore, should be set at levels that would provide the recipient family with cash to buy land elsewhere, or with start-up capital for self-employment or business activity in the current or alternatively chosen place of residence. Public land distribution is a strategy that supplements or partially substitutes for cash compensations. It is strongly recommended that PRM communicates the urgency of these issues to USAID and the Peace Building Commission (Burundi Configuration).
41. The following steps will help create a much-needed system of land-related compensations:
42. Tapping into a major line of loans to Burundi provided by IMF under the Poverty Reduction and Growth (PRGF) facility. “PRGF-supported programs are designed to cover only areas within the primary responsibility of the IMF, unless a particular measure is judged to have a direct, critical macroeconomic impact” (IMF, 2007). Areas typically covered by the IMF include advising on prudent macroeconomic and financial policies and related structural reforms such as exchange rate and tax policy, fiscal management, budget execution, fiscal transparency, and tax and customs administration. It is important, therefore, that PRM discuss with USAID the preparation of a position paper on behalf of the international relief and development donor community, explaining to IMF the macroeconomic drag that landless returnees or dispossessed incumbents will create in terms of GDP foregone, wage levels across the economy driven down, budget revenues curtailed, and budget outlays escalated.
43. To establish compensations at economically viable levels, a system of land valuation is important. Its design should be integrated into the current work on the new land registry. Comparative valuation of land, either by tracking prices of previous transactions or directly appraising land productivity by combination of fertility and location will be important for future land banking, in addition to a system of fair compensations. All sources of influence on the Government of Burundi should be utilized to include the land valuation issue in the land reform agenda. PRM should advocate prompt action in their discussions with UNHCR/Burundi, USAID and other UN and bilateral donor agencies.
44. Physical availability of land is needed for the system of compensations to function effectively. Further consolidation of land in a public inventory and its distribution on the principles of accountability and fairness should be supported by the current and future development aid programs.
45. Mainstreaming repatriation assistance is a way to reduce the divisive effect of international aid. Given that PRM’s funding vehicles are earmarked for refugee assistance, PRM cannot spend on the general population except by producing public goods in the water, sanitation, and construction/ rehabilitation of social service facilities in the communities of high return. Even under the existing system of earmarks, geographic targeting can be refined. Initial simulations performed on data from the UNHCR/Burundi Statistical Database suggest a multivariate targeting formula with the important new variable: ‘Average Post-return Experience in Months’ by commune. This variable must be used in conjunction with several others: ‘Number of unassisted returnees’, ‘Aid Delivery Lag’ (in months of program effort); ‘Presence of Other Donors’. The previously used variables of Number of Returnees’ and ‘Percent of Returnees in the Local Population’ should remain in the formula but be statistically de-emphasized.
46. While expecting the development stage to take over in the next few years, PRM may want to consider a ‘maintenance grant’ to slow down or stop the normal process of deterioration of previous aid results. This grant can be used as a pilot for developing professional workforce and private business start-up in Burundi in the critically important areas of maintenance, construction, and community strengthening.
d. Additional Lessons Learned and Recommendations for PRM:47. This evaluation has produced a set of methodologies and data collection tools that respond to PRM’s request ‘to provide recommendations for PRM’s core indicators on reintegration assistance’. The multi-tier set of reintegration indicators was designed and successfully tested in this evaluation, and is recommended for future use:
48. A composite Welfare Index serves as the statistical basis for a comparative integrated assessment of welfare among the repatriate and the incumbent populations in the areas of return. Welfare parity is an indicator of successful reintegration under the prevalent definition of reintegration. Welfare parity does not imply that every family is at the same level of welfare. It implies that welfare variance is not influenced by the presence or absence of refugee experience in a given sample population. The econometric model, used in this study for testing the predictors of welfare variance is applicable in any country. The Complete Report’s Annex B (pp.250-258) provides a detailed description of the concept and statistical content of the Welfare Index and, also, explains the design of the econometric model.
49. The second tier of indicators enables an ‘itemized’ reintegration assessment in several welfare components, namely: household consumption and savings; household fixed assets (land and housing); household human capital (education and health), and household safety. These components and their statistical descriptors are applicable to any country where PRM supports reintegration assistance.
50. The household survey tool was designed and successfully applied to collect data for the itemized and integrated welfare assessment (see Volume 3 of the Evaluation Report). The household survey’s design was found valid at 90 percent. The remaining 10 percent of questions were misspecified (off target): did not produce meaningful information. This is because the questionnaire included elements of repatriation assistance that are part of the UNHCR assistance program at a generic level but not in Burundi. For example, way stations, apparently, make sense in countries with much lengthier return routes and times than in Burundi. Similarly, part of the return assistance that was correctly captured for Burundi may not be valid in other countries. On the whole, an up to 20-percent customization rate should be planned for in future use of this household survey.
51. The reviewed two tiers of reintegration indicators -- integrated and component-specific welfare -- measure reintegration on technical evidence. The third tier of reintegration indicators measures ‘perceived reintegration’. This measurement is based on the returnees’ self-assessed satisfaction with return. The household survey questionnaire enables an insightful inquiry in this area. Subsection 3.4.3 of the Complete Report illustrates analysis that PRM should be able to replicate in other countries.
52. In addition to providing data for the welfare analysis and ‘perceived reintegration’, the household survey contains a viable template for a concise but effective assessment of the environmental health status (Questionnaire’s Section 13). The incidence rates of diarrheic diseases and acute lower respiratory infections in children are the proven indicators of the accessibility and quality of water and sanitation; as well as the levels of overcrowding and indoor air pollution. The Complete Report’s Annex B explains the environmental health assessment algorithm and how household survey data should be interpreted (more specifically, how reported symptoms should be converted into cases of marker conditions). Given PRM’s priority attention to environmental health, the assessment of relevant disease trends is strongly recommended and can be conducted using the methodology designed for this study.
53. The evaluation has produced the Health Center Questionnaire, the School Questionnaire, and a focus group guide for the assessment of community-based organizations. These tools have been successfully used in the current study and can be recommended for PRM program assessment in other countries.
54 Visual observations played a very important role in this study. They enabled findings that could not be captured from either technical or opinion-based evidence. Visual evidence must be representative and rendered in informative photographic pictures. This is a recommended line of investigation for future PRM assessments of reintegration. However, it is time-consuming, requires extensive site visiting, including travel to remote areas of endemic crime; and therefore cannot be expected as a standard activity. To expect it at all, PRM may want to strengthen its support for the evaluators’ field work, e.g., comply with danger payment requirements per U.S. Department of State regulations.
55. As the next step, PRM would benefit from an operations support consultancy that would help integrate the recommendations of this study into PRM’s grant management templates, such as Cooperative Agreement, Interim Program Evaluation, grant reporting framework; and outcome/impact evaluation framework.
56. It is recommended that PRM joins other bilateral donors such as the Government of Norway and Belgium in supporting periodic external evaluations of their refugee assistance programs. An experience-based standard in the international assistance is to spend 5 percent of a program portfolio budget on evaluations.
- End of Summary Evaluation Report -
To request a copy of the full Evaluation Report, please email the U.S. State Department’s Bureau of Population, Refugees, and Migration.
UNHCR has the UN mandate to provide for the protection and long-term solution of the plight of refugees, and has special expertise in performing these duties. The U.S. government sees UNHCR’s role as unique in this regard, and relies on UNHCR to ensure these duties are carried out. … Because of UNHCR’s multilateral approach and worldwide presence, it plays a critical role in providing and coordinating refugee assistance … although the U.S. government relies on other organizations to carry out refugee assistance as well ’. (OMB, 2004: 4)
2 In previous years, the emphasis was on supporting Burundian refugees in Tanzania and foreign refugees in Burundi.
3 UNHCR/Burundi is cognizant of the adverse implications that the end stage of repatriation may have for the principle and practice of voluntary return. They are poised to take these challenges on. In their own words, “informed decision, voluntary registration, not forcing anyone back, and than dealing with the residual caseload … Those who are not coming back will require a prolongation of protection. This is to be provided either by the current country of asylum or by a third country of asylum – the country where refugees will be resettled under a resettlement program. So, yes, we work to preempt the push-pull imbalance that otherwise might have developed by this time.” (CA, 2008: 2-168).
4 As was mentioned earlier, the average post-return experience in the repatriates sample is 4.2 years.
5 Out of 11, reported in 2006, at the end of UNHCR’s active involvement in the rehabilitation of social service facilities in Burundi.
6 See (Turner, 1999) for an insightful study of how the social environment of the refugee camp impacts on the behavioral patterns and social values of its inhabitants.
The Office of Electronic Information, Bureau of Public Affairs, manages this site as a portal for information from the U.S. State Department. External links to other Internet sites should not be construed as an endorsement of the views or privacy policies contained therein.