Document of The World Bank Report No:ICR000045 IMPLEMENTATION COMPLETION AND RESULTS REPORT ( BELG-24242 IDA-32870 IDA-H0330 ) ON A CREDIT AND A GRANT IN THE TOTAL AMOUNT OF SDR 19.3 MILLION (US$ MILLION 26.2 EQUIVALENT) TO THE REPUBLIC OF BURUNDI FOR A SECOND SOCIAL ACTION PROJECT January 26, 2007 AFT: Human Development 3 AFC09 AFRICA REGION CURRENCY EQUIVALENTS ( Exchange Rate Effective 06/30/2006 ) Currency Unit = BIF BIF 1.00 = US$ 0.001019 US$ 1.00 = BIF 980 Fiscal Year January 1- December 31 ABBREVIATIONS AND ACRONYMS AfDB African Development Bank BURSAP I First Burundi Social Action Project BURSAP II Second Burundi Social Action Project CAS Country Assistance Strategy CDC Community Development Committee CURDES Centre Universitaire de Recherche pour le Développement Economique et Social CWIQ Core Welfare Indicators Questionnaire ECD Early Child Development EU European Union GOB Government of Burundi ICB International Competitive Bidding IDA International Development Association IEC Information, Education and Communication IFAD International Fund for Agricultural Development ISTEEBU Institut de Statistiques et d'Etudes Economiques du Burundi NCB National Competitive Bidding NGO Non-governmental Organization PAD Project Appraisal Document PMR Project Management Report PMU Project Management Unit QBCS Quality-Based and Cost Selection QUIBB Questionnaire Unifié pour les Indicateurs de Base du Bien-être QUID Questionnaire Unifié pour les Indicateurs de Développement SOE Statement of Expenditures UNDP United Nations Development Program Vice President: Gobind T. Nankani Country Director: Pedro Alba Sector Manager: Laura Frigenti Project Team Leader: Pamphile Kantabaze Burundi Second Social Action Project CONTENTS 1. Basic Information........................................................................................................ 1 2. Key Dates.................................................................................................................... 1 3. Ratings Summary........................................................................................................ 1 4. Sector and Theme Codes ............................................................................................ 2 5. Bank Staff ................................................................................................................... 2 6. Project Context, Development Objectives and Design............................................... 3 7. Key Factors Affecting Implementation and Outcomes .............................................. 9 8. Assessment of Outcomes.......................................................................................... 12 9. Assessment of Risk to Development Outcome......................................................... 15 10. Assessment of Bank and Borrower Performance ................................................... 16 11. Lessons Learned...................................................................................................... 19 12. Comments on Issues Raised by Borrower/Implementing Agencies/Partners......... 20 Annex 1. Results Framework Analysis......................................................................... 21 Annex 2. Restructuring (if any) .................................................................................... 26 Annex 3. Project Costs and Financing.......................................................................... 27 Annex 4. Outputs by Component.................................................................................. 29 Annex 5. Economic and Financial Analysis (including assumptions in the analysis).. 41 Annex 6. Bank Lending and Implementation Support/Supervision Processes............. 42 Annex 7. Detailed Ratings of Bank and Borrower Performance.................................. 44 Annex 8. Beneficiary Survey Results (if any) .............................................................. 45 Annex 9. Stakeholder Workshop Report and Results (if any)...................................... 46 Annex 10. Summary of Borrower's ICR and/or Comments on Draft ICR................... 47 Annex 11. Comments of Cofinanciers and Other Partners/Stakeholders..................... 60 Annex 12. List of Supporting Documents .................................................................... 61 MAP.............................................................................................................................. 62 1. Basic Information Country: Burundi Project Name: Second Social Action Project Project ID: P064510 L/C/TF Number(s): BELG-24242,IDA- 32870,IDA-H0330 ICR Date: 01/26/2007 ICR Type: Core ICR Lending Instrument: SIL Borrower: GOVT OF BURUNDI Original Total Commitment: XDR 19.3M Disbursed Amount: XDR 19.1M Environmental Category: C Implementing Agencies TWITEZIMBERE Cofinanciers and Other External Partners 2. Key Dates Process Date Process Original Date Revised / Actual Date(s) Concept Review: 01/26/1999 Effectiveness: 12/23/1999 12/23/1999 Appraisal: 05/17/1999 Restructuring(s): Approval: 10/05/1999 Mid-term Review: 10/30/2002 Closing: 12/31/2003 06/30/2006 3. Ratings Summary 3.1 Performance Rating by ICR Outcomes: Satisfactory Risk to Development Outcome: Substantial Bank Performance: Satisfactory Borrower Performance: Satisfactory 3.2 Quality at Entry and Implementation Performance Indicators Implementation Performance Indicators QAG Assessments (if any) Rating: Potential Problem Project at any time (Yes/No): No Quality at Entry (QEA): Satisfactory Problem Project at any time (Yes/No): No Quality of Supervision (QSA): None DO rating before Closing/Inactive status: Satisfactory 1 4. Sector and Theme Codes Original Actual Sector Code (as % of total Bank financing) Central government administration 8 8 General education sector 20 20 Other social services 34 34 Roads and highways 19 19 Water supply 19 19 Original Priority Actual Priority Theme Code (Primary/Secondary) Other social protection and risk management Primary Primary Participation and civic engagement Primary Primary Child health Secondary Secondary Other rural development Secondary Secondary Other environment and natural resources management Secondary Secondary 5. Bank Staff Positions At ICR At Approval Vice President: Gobind T. Callisto E. Nankani Madavo Country Director: Pedro Alba Emmanuel Mbi Sector Manager: Laura Frigenti Arvil Van Adams Project Team Leader: Pamphile Menahem M. Kantabaze Prywes ICR Team Leader: Pamphile Kantabaze ICR Primary Author: Paul Geli 2 6. Project Context, Development Objectives and Design (this section is descriptive, taken from other documents, e.g., PAD/ISR, not evaluative) 6.1 Context at Appraisal (brief summary of country macroeconomic and structural/sector background, rationale for Bank assistance) Since independence in 1962, Burundi has suffered from recurrent social and political conflict, which reached crisis proportions in 1993 with the assassination of the President. Following a military coup in 1996 an international embargo was imposed, and the Bank suspended its disbursements from October 1996 to May 1997. In 1998, the Government began peace discussions with political parties and armed groups. As the peace process progressed, the embargo was suspended. The Bank prepared and discussed with Government an interim strategy of assistance, whose broad objectives were, inter alia, to support the ongoing peace process and to provide urgent support for the recovery of poor communities and the rural economy. From the beginning of the crisis in 1993, poverty worsened rapidly and considerably. In terms of household expenditures, rural poverty was estimated to have increased by 80 percent and urban poverty to have doubled. Social indicators, related to health, education, and nutrition, also fell. BURSAP II consists entirely of a social fund that targets the poor and conflict-damaged regions of rural Burundi. Its approach is to significantly decentralize power to rural communities. The Project was well aligned with the objectives of the Interim Strategy. It would provide employment and wage income in poor rural communities, restore community infrastructure, and expand delivery of social services. The Project would further bolster recovery by increasing community capacity to participate in the selection, financing and operation of local investments. Moreover, it would help provide a social foundation for peace at the local level by bringing people together to select and to contribute to a project for their mutual benefit. The resulting improvement in local economic conditions and organizational capacity would then underpin and reinforce the peace process. 6.2 Original Project Development Objectives (PDO) and Key Indicators (as approved) Schedule 2 of the Development Credit Agreement (DCA) objectives were to assist the Borrower in : (i)strengthening the capacities of local community groups to work together in selecting, implementing, financing, monitoring and maintaining priority community services; and (ii) alleviating poverty through the creation of employment opportunities and the improvement of social services. The Section A of the PAD objectives were to ensure that : (i) Diverse groups work together to select, partly finance, monitor and evaluate, use, and maintain their priority community investments; and (ii) the poorest and most vulnerable groups benefit from 3 improved social and economic services and improved social protection -- through transitory wages and social services that limit the risks of poverty. The key performance indicators for the realization of the Project's objectives were: · For Objective 1: (i) Qualitative assessment of cooperation of diverse groups in selected projects; (ii) % of women on community development committees (CDC) reflects m/f ratio in the commune; (iii) % of projects that reflect community priorities; (iv) % contribution of the community to the project in material, labor and money; (v) % of CDC that are still functioning one year after completion of the infrastructure sub-project; and (vi) % of infrastructure adequately maintained one year after completion. · For Objective 2: (i) Days of employment created; (ii) number of beneficiaries of infrastructure by type: Education, Health, Water supply, Economic infrastructure (markets,roads, forestry); (iii) capacity utilization rate of the infrastructure one year after construction; (iv) prevalence of wasting, stunting, and underweight among pre-school aged children; and (v) qualitative assessment of adult-child interaction. 6.3 Revised PDO and Key Indicators (as approved by original approving authority), and reasons/justification The original project development objectives (PDO) were not revised. 6.4 Main Beneficiaries, original and revised (briefly describe the "primary target group" identified in the PAD and as captured in the PDO, as well as any other individuals and organizations expected to benefit from the project) The target population was the inhabitants of unusually poor communes that are areas of return for refugees and internally displaced persons (IDP), usually harboring children and other dependents with high rates of malnutrition. Using a poverty index that was calculated by ISTEEBU - the National Institute of Statistics and Economic Studies, Twitezimbere (meaning "let us develop ourselves" in the local language) selected 73 priority communes. The poverty index was based on results from the 1998 priority survey that was financed under BURSAP I and publicized in 2000. 6.5 Original Components (as approved) The Project had three components (referred to as "parts" in the DCA): Component I (or Part A in the DCA): Community investments / activities (US$10.75 million) The component consisted of the carrying out and monitoring of sub-projects for the construction and rehabilitation of social and economic infrastructure, for early childhood 4 development services, and for other community activities submitted by beneficiaries. Component I had three sub-components: Sub-component 1 -- Community participation. Mobilization and strengthening of community capacity (and in particular of representative development committees) to participate in the selection, financing, and continued operation of the infrastructure. The Executive Committee (in which the Government and civil society are represented) and the Director of Twitezimbere were to decide each year which communes would benefit from project-financed activities, using well-defined criteria (see Section 7.1 on design and quality at entry). To increase participation, the Project would work with the local populations to ensure that they actually choose the infrastructure sub-projects to be financed, subject to a threshold earmaked amount of US$75,000 per commune. Communes would be allowed to choose a mix of investments, within the financial envelop. Moreover, the Project would reinforce community institutions that maintained the infrastructure. Multi-disciplinary teams, composed of personnel from Twitezimbere and qualified NGOs (Agences d'Encadrement), were to work with the communities to determine their priorities. The teams were to apply a methodology using community meetings along with graphics, thematic maps, calendar of the seasons, and activity profiles. This procedure was to be used to identify the poorest populations, in consultation with key sources, such as the "Bashingantahe" (elders) and other social or religious leaders. To ensure that the development committees were representative, the teams were to add members after consulting with the beneficiaries and the local administration; the aim was to have a better representation of women, young people and other groups. Specifically, at least 30 percent of the members of each development committee were to be women. The teams were to work with each develpment committee to produce a participatory assessment report. The report was to include: (i) a brief description of the community and the social and economic conditions in the commune; (ii) a map of public facilities and of the concentration of the population along with the economic activity in the commune; (iii) a list of priority investments with a brief explanation and justification; and (iv) the responsibilities of the committee and the community during and after construction. The report was to receive the approval of the local Authority or Administrateur de Commune. Sub-component 2 -- Infrastructure rehabilitation and construction. Contracting for the rehabilitation and construction of small-scale community infrastructure, such as classrooms, health centers, rural water supply, short roads, public markets, swamp drainage, and planting of forests. The roles of Twitezimbere and the Agences d'Encadrement were to remain the same as in the BURSAP I: that is either one would be able to prepare the technical and financial proposals, procure construction of civil works following the Manual of Procedures (Manuel d'Opérations Techniques), and supervise construction of the works. 5 Sub-component 3 -- Investments in young children and mothers -- early childhood development (ECD). During the participatory assessment process, the sub-component was to start by delivering a basic package of information, education and communication (IEC) to promote early childhood development activities (ECD) for a sub-set of the communes with infrastructure sub-projects. All collines from these communes would be eligible to participate. Through the participatory process, each community was to produce a brief assessment of the needs of young children and mothers, ranked in order of priority. Options could include micro-nutrition, gardens to provide nutrition, monitoring of the growth of infants and children, and trauma counseling. The identified activities were to be prepared for each commune and implemented with support from contracted Agences d'Encadrement. Training of community mobilizers, health workers and educators was to back these activities. Twitezimbere was to provide monitoring and evaluation. Component I on community investments/activities also included the provision of training and technical advisory services for: (i) the design and implementation of a communications program to strengthen Twitezimbere's capacity to communicate information about the project to the public; and (ii) assistance to local small entrepreneurs, civil servants in the relevant Ministries and Non Governmental Organizations (NGO) to improve their community activity skills. Component II (or Part B in the DCA): Poverty Monitoring (US$1.20 million) Strengthening the capacity of ISTEEBU and other statistical survey organizations inBurundi to undertake poverty monitoring and analysis activities through the provision of technical advisory services, training, and the acquisition of equipment. BURSAP I financed a relatively large-scale survey to produce detailed baseline information on poverty. BURSAP II, on the other hand, was to finance less expensive surveys that could be repeated in order to provide up-to-date follow-up information on the evolution of poverty. Furthermore, those surveys were to be oriented towards action to reduce poverty rather than simply gathering analytical information about poverty. To that end, the surveys were to provide the information on the condition of roads, water supply, health and education services, and local institutions, necessary for planning reconstruction programs. The relevant indicators were to be selected through a survey of statistics users in the Government, donor, and NGO sectors. Component III (or Part C in the DCA): Administration / capacity building (US$1.25 million) The Project was to be executed by Twitezimbere, an NGO, under an agreement (Convention) with the Government. Twitezimbere means "let's develop ourselves" in the Kirundi language. Twitezimbere is governed by an Executive Committee that represents a cross-section of Burundian society: the Ministries of Finance, Planning and Agriculture, NGOs, and individual members including the Legal Representative of Twitezimbere who chairs the Committee. Component III included: 6 · Strengthening Twitezimbere's capacity to implement, coordinate, supervise, and monitor the execution of the Project, through limited recruitment, the provision of training of its personnel, studies, technical advisory services, and the acquisition of required logistical means. A specialist in participation was to be recruited to reinforce and advance participation activities. Twitezimbere was also to launch a training program for its staff and for the staff of the Ministries concerned, NGO partners, UN organizations and consultants. · Establishment of a monitoring and evaluation system (M & E) to monitor overall project implementation, the implementation of the sub-projects, and the design of a program to overcome any constraints encountered during the implementation of the project, through the provision of technical advisory services. 6.6 Revised Components The original components were not revised but, upon Government request during implementation, the Project has benefited from a Belgian parallel grant (US$ 0.88 million), and an IDA Supplemental Grant (US$14.2 million). Thus, financial allocations per component have increased substantially, as described in section 6.7 below. 6.7 Other significant changes (in design, scope and scale, implementation arrangements and schedule, and funding allocations) There were two significant additions in funding for the Project: (1) in 2001, the Government of Belgium made a grant for early childhood development (ECD); and (2) in 2003, the Bank provided supplemental financing (in the form of a grant) for the Project. The Belgian Grant On November 6, 2001, the Bank, as Administrator of grant funds provided by the Belgium Directorate General for International Cooperation (DHIC), extended a grant of US$882,000 to Burundi for parallel financing of the early childhood development (ECD) sub-component of the project. The purpose of the Belgian Grant was to carry out and monitor ECD services. The Grant was given for the implementation by Twitezimbere of the following activities: a) Mobilization and training of women's groups to maintain essential services for child health, nutrition, psycho-social development and learning for children ages 0 to 6. b) Carrying out IEC campaigns to community groups and mothers in order to promote ECD. c) Distribution of basic packages of goods for child health care. d) Strengthening communities' capacity in child care, including child nutrition through the provision of training and advisory technical assistance. 7 e) Providing technical assistance services for monitoring children's growth and evaluating project outcomes. f) Training of trainers and capacity building for women's groups to set up informal pre-school education circles and other psycho-social interventions through technical advisory services and the acquisition of small educational equipment. Supplemental financing from the Bank The Bank's Transitional Support Strategy for FY02 and FY03 (February 12, 2002) supported national reconciliation and reconstruction through a lending program that included a Social Action Project along with an Emergency Recovery Credit, Emergency Road Rehabilitation Credit, and a Demobilization Project. The peace accord of 2000 and the swearing in of a transitional Government in November 2001 triggered an unexpected increase in the cost of the BURSAP II Project because its objective had to be realized over a larger territorial expanse (regions added in the far south and eastern frontier) and for a larger population (newly returned internally displaced persons and refugees). The integration of the returning refugees and IDPs required the rehabilitation of additional community facilities, such as schools and health centers. At the time of the Mid-term Review (MTR) in October 2002, the Project had already disbursed or committed 92% of its resources, so that its implementation momentum was satisfactory. The indicators of project outputs and outcomes demonstrated the positive results of the Project. The Government demonstrated its commitment to the project by paying its counterpart contributions despite its tight fiscal budget. The executing agency, Twitezimbere, was fully competent, as demonstrated by the indicators of Project outcomes and outputs, as well as by the independent financial audits for 2000 and 2001 and by the technical and social audit of project activities for 2000/2001. The reviews carried out by the Bank procurement and financial management specialists confirmed this favorable assessment. All other requirements of OP13.20 which governs supplemental financing were being met. The urgency of meeting the Bank's commitments to support national reconciliation and economic reconstruction also justified supplemental financing since the preparation of a new project would be more time-consuming. On April 14, 2003, IDA made a grant of SDR10.3 million (US$14.2 million) to supplement the ongoing IDA Credit for BURSAP II, following the initial project design (to decentralize powers to local communities) and for the same original project components and sub-components. Seventy-three communes, which were the most vulnerable ones, were selected to benefit from the grant. The highest priority use of the grant was to rehabilitate war-damaged areas and ease the reintegration of returning refugees and displaced persons. More specifically, it was agreed with Government that the supplement would finance the following: a) Community mobilization (US$1.0 million). The supplement would support community mobilization and the production of new or updated community development plans in 65 communes. b) Infrastructure sub-projects (US$8.4 million). The majority of the supplement 8 would support delivery of small-scale community infrastructure sub-projects to 65 communes, with priority targeting of the provinces of Ruyigi and Makamba. The sub-project financing limit for each commune would increase from US$75,000 to US$100,000 so that communities could decide to build short-rural access roads and other economic infrastructure. c) Early childhood development (US$3.0 million). The supplement would allow the Project to continue to support delivery of social services to pre-school children in 34 communes. Moreover, it would support construction of 150 shelters for informal pre-school circles. It was agreed that, over the medium-term, management and monitoring would be shifted to the Ministry of Education, while delivery of services would remain a community responsibility. Therefore, the supplement would support training to prepare for the integration of policy-making and monitoring aspects of the component into the Ministry of Education. d) Poverty monitoring (US$0.8 million). The supplement would support a rapid statistical survey of social and economic conditions (a Core Welfare Indicators Questionnaire -CWIQ) for use in monitoringBurundi 's progress in recovering from severe poverty. Moreover, the component would support training of candidates for the entry exams for the African statistical institutes. e) e) Administration (US$1.0 million). The cost of administration was about 7% of project costs. This covered Twitezimbere's costs for executing the Project as well as audits and special studies. 7. Key Factors Affecting Implementation and Outcomes 7.1 Project Preparation, Design and Quality at Entry (including whether lessons of earlier operations were taken into account, risks and their mitigations identified, and adequacy of participatory processes, as applicable) Overall, the quality at entry was satisfactory. The project risks were well identified, with appropriate risk minimization measures proposed in the project appraisal document (PAD). The lessons learned through the implementation of BURSAP I were used in the preparation of BURSAP II. The most important lesson from BURSAP I was that, in a country in a state of crisis and where insecurity is rampant, this type of demand-driven project to rehabilitate infrastructure at the village level could still be implemented. The process helps build up community solidarity, contributes to national reconciliation, and helps the country graduate to a "post-conflict" status. Another lesson was the importance of true community participation, including timely contribution by beneficiaries to sub-project costs, so that they would own the sub- projects and be very much involved in their maintenance and continued operation. Whereas under BURSAP I the priorities of the population may have been influenced by commune administrators or NGOs, the design of BURSAP II included a strengthening of community participation, particularly in the selection, financing, and management of infrastructure sub-projects and social services delivery. Therefore, in addition to an 9 improvement in the targeting of poor communities, the main change in BURSAP II, relative to BURSAP I, was to empower residents of communities to choose and maintain their own project investments. The specific changes were to: a) Use participatory assessments to establish local needs and investment priorities in each commune; b) If necessary, modify the community development committees through consultation with the public to assure that they are representative; c) Allocate a financial envelope of US$75,000 to the residents of each commune to finance their priority infrastructure sub-projects (but not including ECD activities); and d) Increase the choice available to communities by allowing them to choose a mix of infrastructure investments, within the financial envelope. The project design was kept simple: inBurundi 's emergency and civil conflict context, this was very appropriate for delivering urgently needed support quickly, taking into account the limited local capacity for project execution. The weak point in the design of BURSAP II was Component II on poverty monitoring which, in fact, was never really fully developed. The PAD did not include much information on its intended contents, but the project execution plan scheduled one statistical survey per year which turned out to be too ambitious (in any event, the need for such frequent surveys is questionable). Overall, this component would not have proven relevant within the design of this CDD-type operation but a macroeconomic management support operation would have been a more appropriate context. However, the Government was concerned about strengthening capacities of ISTEEBU for monitoring poverty and human development indicators during the very early period of post-conflict socio-economic recovery. 7.2 Implementation (including any project changes/restructuring, mid-term review, Project at Risk status, and actions taken, as applicable) The major factors affecting implementation were the security situation and the instability of the administration. With the evolution of the political situation (regime change, peace agreement, cease fire, elections), there were many changes in the administration, at the central level as well as at the level of the provinces and communes. Over the life of the project, the Minister of National Reconstruction and Planning (MPDRN) changed six times, while many provincial governors and commune administrators also changed repeatedly. Despite all these fluctuations, and maybe because it was apolitical, Twitezimbere managed to keep the support of the administration at both the central and provincial and commune levels. Because of the lack of security and the extreme poverty that prevailed in certain areas, the execution of sub-projects was at times very difficult. Work sites had to be abandoned, and then started again, to be abandoned once more, depending on the security situation. Many other government entities gave up and left, but not Twitezimbere. No 10 Twitezimbere staff lost their life in this endeavor, but it was certainly a daunting experience. Needless to say, these conditions affected the supervision of activities. On the positive side, Twitezimbere was able to continue to operate because it had gained the confidence of the communities. Also, the population was already used to mobilizing itself to participate and contribute to the execution of sub-projects in cooperation with NGOs. Finally, the availability of a detailed and precise implementation manual helped. On the negative side, there were some technical issues which made the execution of sub- projects difficult, such as difficulties in finding seasoned wood for construction and a shortage of skilled construction contractors. 7.3 Monitoring and Evaluation (M&E) Design, Implementation and Utilization According to the PAD, Twitezimbere would recruit a local consulting firm or contract with the ISTEEBU to monitor project performance. Actually, there was no need for such sub-contracting; the M&E function was carried out satisfactorily by a Twitezimbere's dedicated project management team, led by a deputy director for M&E. 7.4 Safeguard and Fiduciary Compliance (focusing on issues and their resolution, as applicable) In April 1999, an environmental analysis was conducted for the Project. The report identified both positive and negative environmental and social impacts of future sub- projects, and presented mitigation measures for the negative impacts for the various types of sub-projects. Based on the report, environmental guidelines were included in Twitezimbere's manual of procedures. In 2003, in connection with the supplemental financing, Twitezimbere adopted a Social and Environmental Risk Management Framework to prevent or mitigate any damage to the environment resulting from Project works. Twitezimbere followed the policies and procedures set forth in the Framework. The Framework also defined the legal status of land used by the project, which is nearly all public, and assured fair compensation in case of displacement of people or expropriation of property. It should be noted that the project never displaced anyone from their homes or expropriated any land. Various audits and procurement post-reviews documented a continual improvement in fiduciary management. No major fiduciary problems have been associated with the implementation of BURSAP II. Indeed, the Project had the highest performance rating among 10 projects in the review of the fiduciary management of the IDA portfolio in February 2006. 7.5 Post-completion Operation/Next Phase (including transition arrangement to post-completion operation of investments financed by present operation, Operation & Maintenance arrangements, sustaining reforms and institutional capacity, and next phase/follow-up operation, if applicable) 11 Because of the participatory process, communities developed a strong sense of ownership of their sub-projects. As the target rate of 95% may have been met with ease in most communes of the Project, the maintenance and utilization record of completed sub- projects is expected to be satisfactory. The Government is also working on its decentralization policy and is committed to take into consideration local development institutions such as the CDC set up under BURSAP II prior to the issuance of 2005 Communal Law. A new IDA-financed project, namely Community and Social Development Support Project, is already under preparation and is aimed at - inter alia - ensuring that communities and population have access to basic services and are supportive of the peace process and economic reform agenda... The new Project will draw on Bank's previous experience in implementing community driven projects in the country such as BURSAP I and II. The Bank's involvement will foster cooperation and coordination among bilateral and multilateral donors and will help government leverage further funding from them for the implementation of the decentralization strategy and will deepen the support to communes. With the implementation of the new communal law, strengthening the relatively strong NGOS that have become active on other ongoing projects, Twitezimbere and some of its peers seem likely to get definitely involved in a more visible leadership role in promoting and implementing participatory approaches of community development. Experience with Twitezimbere has provided a good example of a model for institutional strengthening at the community and civil society levels, which the new Government has somehow adopted in the context of preparation of a decentralization policy. BURSAP has done a lot to feed into the preparation of the new project. 8. Assessment of Outcomes 8.1 Relevance of Objectives, Design and Implementation (to current country and global priorities, and Bank assistance strategy) The project's original development objectives are still consistent with current development priorities and both country and sectoral strategies. Such continuous consistency has been evidenced through (a) the supplemental financing that was approved in 2003; (b) the April 11, 2005 Interim Strategy Note for FY 2006-2009; and (c) the full Poverty Reduction Strategy Paper (PRSP) that had been approved by the Government and the Parliament and sent to the Bank and IMF at the time of this Implementation Completion Report (ICR). The Project's design and implementation arrangements are even more relevant now than they were at the time of project approval. 8.2 Achievement of Project Development Objectives (including brief discussion of causal linkages between outputs and outcomes, with details on outputs in Annex 4) The project development objectives have been reached. By using the results of the 1998 priority survey, the Project was able to target the poorest communes; it also made 12 extraordinary progress in empowering communities. This is an extraordinary achievement considering the on-going civil conflict and the absence of any elected local government. Annex 1 presents the outcome, output and process indicators, and Annex 4 discusses in great detail the outputs by component. The following are some of the Project achievements: a) BURSAP II arranged for the election of community development committees (CDC), which decided how funds will be used. This is undoubtedly the most important achievement of the Project. b) The beneficiary participation has been real: all the sub-projects financed by Twitezimbere have been selected based on the population priorities, and on average beneficiaries contributed about 8.5 percent of the costs of the sub- projects. c) During the whole project period, Twitezimbere completed 1,219 sub-projects: 657 for water supply, 278 for education, 94 for health, 178 for pre-school circles, 10 for rural roads and bridges, and 2 for social centers (for details by type and by year, see the table in Annex 4). d) More than 1.4 million person/days of temporary employment were created under the project, about four times the target of 350,000. e) Technical and social audits were generally positive regarding the quality of community mobilization and participation, as well as the quality of the infrastructure. f) The participatory process brought the major ethnic groups together to work on investments that are in their common interest and, in this way, supported national reconciliation. g) The ECD sub-component, which had been suggested by the Bank as a response to extreme conditions (poor nutrition, poor health care and poor preparedness of young children for primary school) was a success. It led to improved nutrition and health among children at low cost. Compared to the target rate of 15%, the Project achieved a severe malnutrition rate of children 0-6 of 13.3%. Also, according to school directors and teachers, the best students in the first years of primary schools are often those that participated in the program. The 2006 social audit indicates that the population is satisfied with the Project and are asking for more facilities. It was commented by some of the people that: "Only Twitezimbere has not abandoned us. Twitezimbere stayed with us and helped us very much." For all infrastructure sub-projects, the main benefit for the population was improved access, such as the reduction in the time and effort required to go to a school or a health center or to fetch drinking water. Infrastructure-specific benefits included an increase in school enrollment as a result of the construction of additional classrooms and an improvement in the health status of the population due to the health centers and water supply sub-projects. 13 The component on poverty monitoring has been much less successful, mostly because of I|STEEBU's weak management and lack of qualified statisticians. There has been basically no progress in building the capacity of the ISTEEBU. The results of the 1998 priority survey have been used by Twitezimbere for the selection of the poorest communes, but the usefulness of the two rapid surveys that have been carried out was jeopardized by the delays in their completion. 8.3 Efficiency (Net Present Value/Economic Rate of Return, cost effectiveness, e.g., unit rate norms, least cost, and comparisons; and Financial Rate of Return) As discussed in somewhat more detail in Annex 5 on Economic and Financial Analysis, Twitezimbere has been very efficient. With a low ratio of disbursements for operating expenses to total disbursements of less than 10% for the whole project period, Twitezimbere's performance has been highly satisfactory. Also, the costs of the facilities built by Twitezimbere are less than those of similar investments carried out by other organizations (Annex 5). 8.4 Justification of Overall Outcome Rating (combining relevance, achievement of PDOs, and efficiency) Rating: Satisfactory The project design and implementation arrangements are, if anything, even more relevant now than at the time of project approval. The project development objectives have been achieved and, with the exception of the poverty monitoring component, the outputs have been substantial. The population is satisfied with the Project, which Twitezimbere implemented very efficiently. In recognition of its experience in community participation, Twitezimbere has been given a leadership role in the design and implementation of the process of PRSP community consultations. 8.5 Overarching Themes, Other Outcomes and Impacts (if any, where not previously covered or to amplify discussion above) (a) Poverty Impacts, Gender Aspects, and Social Development The Project encouraged the full participation of women. However, the elections of the members of the CDCs were free without any quotas for women. The target of 50% of women in CDCs has not been reached. On average, the percentage of women in CDCs is 41%, but there are variations between provinces and communes. In a very few communes, there are no women at all, whereas in others the percentage of women is about 70%. Although the Project targeted the poorest communes, at the level of each commune the participation of the poorest of the poor in the whole process remained low. 14 (b) Institutional Change/Strengthening (particularly with reference to impacts on longer-term capacity and institutional development) The Project has been very useful for NGOs working as Agences d'Encadrement. Small and medium enterprises (SME), individual consultants, and consulting firms have been established, thanks to the business opportunities provided by the infrastructure sub- projects. On the other hand, the Project's impact on ministries and government technical services has been limited, and on ISTEEBU it has been virtually nil. (c) Other Unintended Outcomes and Impacts (positive or negative, if any) In the early years of the Project, Burundi did not have elections or local governments; "administrateurs de communes" were local representatives of the central government. Therefore, the election by the people of the CDCs was a very positive initiative; it created an electoral basis for representative local government. Thus, BURSAP outcomes have paved the way for the government policy for decentralization. The quality of the participatory rural appraisals (PRA) carried out under the Project has been good. The social audit noted that the populations liked to be consulted, and that the local administration was also very satisfied that the populations were consulted since their involvement facilitated the implementation of sub-projects. 8.6 Summary of Findings of Beneficiary Survey and/or Stakeholder Workshops (optional for Core ICR, required for ILI, details in annexes) Since this is a Core ICR, no beneficiary survey or stakeholder workshop was held at the end of the Project. However, in early 2006 an individual consultant carried out a social audit of BURSAP II. The report, issued in April 2006, indicates that the populations are satisfied with the Project. Annex 8 provides a summary of the report. 9. Assessment of Risk to Development Outcome Rating: Substantial The evaluation of the risk that the development outcome will not be maintained has to be done at several levels: at the level of Twitezimbere and the project approach in terms of community participation, and at the level of the completed sub-projects (to the extent that the achievements of the component for poverty monitoring are limited, the question of its sustainability does not arise). Regarding completed sub-projects, the evaluation is favorable. Despite the conflict and insecurity in the country, none of the facilities built or rehabilitated by Twitezimbere has been destroyed. The populations -- proud of their role in creating these facilities -- appear to be doing their best to protect them, a clear benefit of community participation. The facilities are operated and maintained, albeit perhaps minimally at present. The notion of maintenance is not yet fully understood or accepted by beneficiaries, but it is expected that they will provide routine maintenance at a minimum. The sustainability of the ECD sub-component was supposed to arise from its low cost, 15 due in part to the volunteer work by coordinators and local mothers. Despite the strong demands of communities, the sustainability of all the aspects of this sub-component (health, nutrition, psycho-social development) is unlikely. There are many other priorities that require huge amounts of money, and public resources are limited. Regarding Twitezimbere and the community participation approach, the situation is somewhat mixed. Twitezimbere has become a key partner of the Government on community development and a key contributor to government decentralization policy. In the past, Twitezimbere has demonstrated a capacity to survive and to raise money from donors and, capitalizing on its community participation expertise and experience, Twitezimbere should be able to be awarded contracts and work as a NGO. However, in the absence of a BURSAP III, the scope of activities is bound to be much reduced, and many staff have already left Twitezimbere looking for other employment opportunities. The community participation approach promoted by Twitezimbere has been followed by other projects and organizations, but it is uncertain whether government ministries will be as eager as an apolitical organization as Twitezimbere to adopt it and use it in future projects. Actually, Government's policy is unlikely to contract Twitezimbere's services for the purpose of implementation of new projects. Instead, new projects will be run by Government Ministries that may sub-contract some services to private entities like Twitezimbere. In this context, despite the proven track record of Twitezimbere and the unfinished agenda (in terms of all the education and health facilities, etc. that need to be constructed or rehabilitated), the Government and the Bank will not be continuing to provide direct support to Twitezimbere. On balance, the risk to development outcome is rated as substantial. 10. Assessment of Bank and Borrower Performance (relating to design, implementation and outcome issues) 10.1 Bank (a) Bank Performance in Ensuring Quality at Entry (i.e., performance through lending phase) Rating: Satisfactory The performance of the Bank through the lending phase has been satisfactory, ensuring quality at entry. The design of BURSAP II took into account the lessons from BURSAP I. It included a strengthening of community participation, but was still kept simple. There is an area where the Bank's position is somewhat difficult to justify. This relates to the reliance on ISTEEBU for the poverty monitoring component, considering that ISTEEBU had already a poor track record under BURSAP I. Why was it included in BURSAP II in the first place? The position most often quoted cited in response that the ISTEEBU is the only agency permitted to conduct official government surveys in Burundi and, therefore, the Bank had no choice other than to work with them is not a very strong argument. 16 (b) Quality of Supervision (including of fiduciary and safeguards policies) Rating: Satisfactory The quality of supervision has also been satisfactory. The implementation support provided by the Bank was affected by the security situation. During its life, the Project had two different task team leaders, first a Headquarters-based staff and then a country office-based staff. The Bank country office provided a continuous presence in the field and maintained good working relations with Government, Twitezimbere and other partners involved in the Project. Following the ceasefire (signed in November 2004) between the Government and the main rebel groups, security improved. Fighting ceased in 16 out of 17 provinces of Burundi . For the first time since 1996, field missions were cleared by the UN Security Field Office, allowing the use of ground transportation instead of humanitarian aircrafts. Therefore, supervision became easier. To facilitate supervision and also to ensure that Twitezimbere benefit from good professional advice, the Bank saw to it that technical and social audits as well as procurement and financial reviews were carried out at reasonable intervals, in addition to the usual financial audits. The Bank's role has been very positive for mobilizing additional resources for the project (such as the Belgian Grant) and for processing quickly a very substantial supplemental financing in the form of a grant. Another positive aspect of the Bank involvement is the flexibility that it showed to adapt to Burundi 's changing circumstances. Examples of this flexibility are the increases in the financing ceiling (from US$75,000 to US$100,000, and then to US$125,000 per commune) and the reduction (from 12% to 7%) in the required community contribution to the costs of sub-projects. Following the 2003-2004 great famine, particularly in the northern regions, the community contribution had become impossible; the Bank agreed to finance 100% of sub-project costs in the provinces of Kirundo and Muyinga which have been officially declared "disaster areas." On the other hand, with respect to the poverty monitoring component, it seems that the Bank was too flexible. In view of the continuing poor performance of the ISTEEBU throughout the Project, it would have been more appropriate for the Bank to cancel the component on poverty monitoring. Instead, the Bank team agreed to support Government request to conduct 2002 and 2005 CWIQ surveys which could not be financed otherwise. (c) Justification of Rating for Overall Bank Performance Rating: Satisfactory The Bank ensured good project design. It played a positive role by securing additional resources for project implementation and it displayed flexibility to adapt to Burundi 's changing circumstances. The overall Bank Performance is satisfactory. 17 10.2 Borrower (a) Government Performance Rating: Moderately Satisfactory Government was committed to the objectives of the Project and supportive of Twitezimbere. Central Ministries were closely involved either by sitting on Twitezimbere Executive Committee (Finance and Planning) or by providing technical support at the sectoral and decentralized levels (Health, Education, Communal Infrastructure, Agriculture & Livestock). Government paid its counterpart contribution in a regular and timely manner. Local Governments, namely Provincial and Communal Authorities, were more involved, by providing administrative and logistical support in various aspects of the project implementation at community levels. On the other hand, ISTEEBU's performance was unsatisfactory due to weak institutional capacity. The project execution plan scheduled one statistical survey per year, but after much delay the ISTEEBU delivered only two surveys in all. One reason for the initial delay was that ISTEEBU could not start a new survey when it had not yet published the results of the 1998 priority survey that had been financed under BURSAP I. The results of the priority survey were finally published in February 2001 (under BURSAPII funding) and this delay severely limited its usefulness. More generally, there were delays in the production of routine statistics and complaints from several donors. Furthermore, the project execution plan included the training of Burundian statisticians. Unfortunately, this training was not provided despite more than two years of effort by Twitezimbere and the Bank to work with the ISTEEBU and then the Faculty of Economics and Administration at theUniversity of Burundi. Finally, the Government allocated an appropriate mix of resources from BURSAP II and mostly from a different Bank- financed project - Economic Management Support Project - to start the training program at the University of Burundi. Contribution from BURSAP II consisted in delivering training on data collection and processing (using 2005 CWIQ/QUIBB Consultant services and providing computers and software). In balance, the Government's performance rating is moderately satisfactory. (b) Implementing Agency or Agencies Performance Rating: Highly Satisfactory Implementing Agency Performance TWITEZIMBERE Twitezimbere, a not-for-profit association, has been responsible for the execution of the project. Despite the civil war and the insecurity, its staff has continued to execute the Project, often traveling through contested areas. The staff displayed personal courage and commitment to the rural population, and was in many respects heroic. They carried out missions in dangerous regions with on-going guerilla warfare. The leadership of Twitezimbere was particularly good at 18 diplomacy, both with Government and local populations. It illustrates the importance of recruiting the right staff who can establish good relations with the Government and the beneficiary population during a civil conflict. The staff has been stable. Overall, project execution by Twitezimbere has been highly satisfactory, with low administrative costs. A series of audits showed a constant improvement in performance, with respect to both the technical expertise for community development and the project's fiduciary management. The Project had the highest performance rating among 10 projects in the review of the fiduciary management of the IDA portfolio in February 2006. The only weakness was in statistical work. Twitezimbere which sub- contracted the execution of the second component (poverty monitoring) to the national statistical institute, the ISTEEBU, was not qualified to take over responsibilities from the ISTEEBU. Twitezimbere cannot be held responsible for the failure of the ISTEEBU. (c) Justification of Rating for Overall Borrower Performance Rating: Satisfactory With a "moderately satisfactory" Government performance, and a "highly satisfactory" Twitezimbere performance, the overall Borrower's performance is rated "satisfactory." 11. Lessons Learned (both project-specific and of wide general application) The Project was a key to providing community access to basic services in the context of a conflict situation, and was an excellent preparation for the upcoming Community-Driven Development (CDD) Project. The risk to the outcomes were significant given the unpredictability of the situation at the time of effectiveness but the successful participatory approach has become an effective contribution to community reconciliation and thus far the Project has been able to implement the community development component that was the cement for Project activities. The following lessons of wide and general application can be drawn from the successful implementation of BURSAP II: a) Community participation can ensure ownership by the communities of the whole approach and process, including community mobilization and the selection of sub-projects, the monitoring of construction and the subsequent operation and maintenance of infrastructure. b) In an emergency and civil conflict context, there is a need for simplicity of project design for delivering urgently needed support quickly, taking into account 19 the limited local capacity for project execution. c) During a civil conflict, it is essential for the implementing agency to maintain good relations with all stakeholders in order to implement sub-projects. d) In a post conflict phase, there is the potential to work with local governments, if these are established. e) A social fund can mobilize communities to work with the formal health and education structures to supply social services, rather than limit itself to bricks and mortar. f) The following are some project-specific lessons: g) The use of qualified NGOs and other entities to work as Agences d'Encadrement has enabled Twitezimbere to carry out the community mobilization in almost the whole country in a relatively short period of time. However, Twitezimbere still needed to have sufficient technical staff to closely monitor the execution of sub- projects in order to ensure that their quality is satisfactory. h) ECD activities can be undertaken in rural areas at a low cost provided that implementation arrangements are based on community participation and emphasize proximity aspects. i) The design of a monitoring component for a social fund project should ensure that only competent organizations are recruited for implementation. j) Subcontracting was a challenge due to limited NGO capacity and administrative requirements. k) The Bank should have taken early action to design proper indicator to monitor poverty. l) 11) This type of project could be a key contributor to improved development outcomes in a difficult situation. 12. Comments on Issues Raised by Borrower/Implementing Agencies/Partners (a) Borrower/implementing agencies The ICR team has had ad hoc exchanges with Twitezimbere especially on the data provided. The team has formally requested the Government's comments which were expected until the last minute. However, a summary of the Borrower's completion report is provided in Annex 10 of this ICR. (b) Cofinanciers Comments have been requested from the Office for Belgian Technical Cooperation but we have not received any reaction. (c) Other partners and stakeholders (e.g. NGOs/private sector/civil society) 20 Annex 1. Results Framework Analysis Project Development Objectives (from Project Appraisal Document) Schedule 2 of the Development Credit Agreement (DCA) objectives were to assist the Borrower in : (i)strengthening the capacities of local community groups to work together in selecting, implementing, financing, monitoring and maintaining priority community services; and (ii) alleviating poverty through the creation of employment opportunities and the improvement of social services. The Section A of the PAD objectives were to ensure that : (i) Diverse groups work together to select, partly finance, monitor and evaluate, use, and maintain their priority community investments; and (ii) the poorest and most vulnerable groups benefit from improved social and economic services and improved social protection -- through transitory wages and social services that limit the risks of poverty. The key performance indicators for the realization of the Project's objectives were: · For Objective 1: (i) Qualitative assessment of cooperation of diverse groups in selected projects; (ii) % of women on community development committees (CDC) reflects m/f ratio in the commune; (iii) % of projects that reflect community priorities; (iv) % contribution of the community to the project in material, labor and money; (v) % of CDC that are still functioning one year after completion of the infrastructure sub-project; and (vi) % of infrastructure adequately maintained one year after completion. · For Objective 2: (i) Days of employment created; (ii) number of beneficiaries of infrastructure by type: Education, Health, Water supply, Economic infrastructure (markets,roads, forestry); (iii) capacity utilization rate of the infrastructure one year after construction; (iv) prevalence of wasting, stunting, and underweight among pre-school aged children; and (v) qualitative assessment of adult-child interaction. Revised Project Development Objectives (as approved by original approving authority) The original project development objectives (PDO) were not revised. 21 (a) PDO Indicator(s) Actual Value Achieved Indicator Baseline Original Target Formally Value Values (from approval Revised Target at Completion or documents) Values Target Years Indicator 1 : Evaluation of the cooperation of diverse groups in the choice of infrastructure Value (quantitative Not or Available positive Not Applicable positive Qualitative) Date achieved 07/07/1999 12/31/2005 12/31/2005 06/30/2006 Comments No baseline info was available at appraisal. Based on the previous project's ICR (incl. % (positive), this qualitative indicator was monitored through supervision missions, achievement) MTR and in the end of project (EOP) social audit. Indicator 2 : Percent of women in CDCs Value (quantitative Not or Available 30% Not Applicable 41% Qualitative) Date achieved 07/07/1999 12/31/2005 12/31/2005 06/30/2006 Comments No baseline info was available at appraisal. The project performance at MTR (incl. % (40%) and at completion (41%) is documented in the final social audit. Arusha achievement) Peace Agreement and strong political commitment have resulted in 30% m/f ratio after 2005 elections. Indicator 3 : Subprojects that reflect community priorities Value (quantitative Not or Available 100% Not Applicable 100% Qualitative) Date achieved 07/07/1999 12/31/2005 12/31/2005 06/30/2006 Comments Documented through Implementation Status Reports (ISRs), MTR, final social (incl. % and technical audits. All Sub-projects were identified through 'participatory rural achievement) appraisal' methodology. Indicator 4 : Percentage of community contribution to projects (in kind or in cash) Value (quantitative or 15% 12% 7% 8.5% Qualitative) Date achieved 12/31/2000 12/31/2003 12/31/2005 06/30/2006 Comments Initial target was 12% vs 15% achieved at completion of BURSAP I. The Bank (incl. % agreed to a reduction of the target to 7% because of extreme poverty, famine and achievement) drought. The project performance was 11.6% at MTR vs 8.5% at completion. Source:Technical audit. 22 Indicator 5 : Percentage of Community Development Committees that are operational one year after completion of their infrastructure sub-project Value (quantitative 0 80% Not Applicable 80% - see comment or below Qualitative) Date achieved 07/07/1999 12/31/2005 12/31/2005 06/30/2006 Comments Based on sites visits and ISRs, both Technical and Social Audits have concluded (incl. % on significant achievements. Though no specific analytical work was done, the achievement) estimated achievement is as high as 80%. Indicator 6 : Percentage of infrastructure sub-projects adequately maintained one year after completion Value (quantitative Not 90% - see comment or Applicable 95% Not Applicable below Qualitative) Date achieved 07/07/1999 12/31/2005 12/31/2005 06/30/2006 Comments (incl. % Technical and Social audits were much positive but no specific analytical work achievement) was done. Estimated achievement is 90%. Indicator 7 : Number of days of (temporary) employment created Value (quantitative Not or Applicable 350,000 Not Applicable 1,431,250 Qualitative) Date achieved 07/07/1999 12/31/2005 12/31/2005 06/30/2006 Comments (incl. % Source: Reports prepared by Twitezimbere achievement) Indicator 8 : Infrastructure Utilization Rate one year after completion Value (quantitative Not 95% - see comment or Applicable 95% Not Applicable below Qualitative) Date achieved 07/07/1999 12/31/2005 12/31/2005 06/30/2006 Comments Estimated achievement at MTR was 95%. Nationwide, occupation of classrooms (incl. % is at least 200%. Health facilities that are 100% functioning receive 20% of achievement) expected patients but this is a health sector issue. Pre-school circles have been overcrowded. Indicator 9 : Malnutrition Rate (weight/age) of children under 6 Value (quantitative Not 18.35% (not measured at or Available Not available 15% completion) Qualitative) Date achieved 07/07/1999 12/31/2003 12/31/2005 09/30/2002 Comments The rate of 18.35% resulted from a national MICS survey that was made in 2002. (incl. % The target of 15% was therefore proposed for the supplemental financing during achievement) 2003-2005. No survey was conducted at completion. 23 (b) Intermediate Outcome Indicator(s) Original Target Actual Value Indicator Baseline Values (from Formally Revised Achieved at Value approval Target Values Completion or documents) Target Years Indicator 1 : Number of participatory assessments to prepare community development sub- projects and plans (number of communes) Value (quantitative or 0 73 Not Applicable 93 Qualitative) Date achieved 07/07/1999 12/31/2003 12/31/2005 06/30/2006 Comments Twitezimbere had targeted the 73 poorest communes based on a poverty index (incl. % calculated by ISTEEBU at the time of appraisal. At completion 100% or 116 rural achievement) communes had been covered by diverse sources, 93 of which were covered by Twitezimbere. Indicator 2 : Community infrastructure sub-projects completed (average number of sub- projects per year) Value (quantitative 100 (was the proposed or 0 Not Determined target for 2003-2005 203 Qualitative) supplemental financing) Date achieved 10/31/2002 12/31/2005 12/31/2005 06/30/2006 Comments At MTR, the Project average performance amounted to 123 infrastructure sub- (incl. % projects per year. At completion, a total of 1,219 sub-projects were completed. achievement) See table in Annex 4 for details. Indicator 3 : Number of Agreements/Contracts signed with intermediary NGOs, community based organizations (CBOs), and SMEs Value (quantitative or 0 Not Determined Not Applicable 1519 Qualitative) Date achieved 07/07/1999 12/31/2005 12/31/2005 03/31/2006 Comments (incl. % At mid-term Twitezimbere had signed 193 contracts. achievement) Indicator 4 : Number of Members of CDCs trained in Child Health and Nutrition Value (quantitative or 0 Not Determined Not Applicable 740 Qualitative) Date achieved 10/31/2002 12/31/2005 12/31/2005 06/30/2006 Comments Twitezimbere had trained 340 participants at MTR. (incl. % achievement) 24 Indicator 5 : People receiving IEC on ECD Value (quantitative or 0 Not determined Not Applicable 112,240 Qualitative) Date achieved 12/31/1999 12/31/2005 12/31/2005 06/30/2006 Comments (incl. % Achievement at MTR (2002) was 74,085 participants achievement) Indicator 6 : Quality of infrastructure Value (quantitative Not or measured favorable Not Applicable favorable Qualitative) Date achieved 07/07/1999 12/31/2005 12/31/2005 06/30/2006 Comments (incl. % Qualitative value, as observed during ICR mission and reported by social and achievement) technical audits. Indicator 7 : Cost of administration as a percentage of total disbursements Value (quantitative Not or Available 10% Not Applicable 9.77% Qualitative) Date achieved 07/07/1999 12/31/2005 12/31/2005 06/30/2006 Comments (incl. % Twiteximbere claimed spending less than 2% overhead as a percentage of total achievement) infrastructure investment. Indicator 8 : Mean length of the sub-project cycle, from identification to completion (in months) Value (quantitative Not or Applicable 3.0 Not Applicable 4.5 months Qualitative) Date achieved 07/07/1999 12/31/2005 12/31/2005 06/30/2006 Comments (incl. % Based on 11 months average length at MTR, The target of 8 months was therefore achievement) proposed for 2003-2005 supplemental financing, which has been widely achieved. Indicator 9 : Statistical surveys undertaken Value (quantitative or 0 4 2 2 Qualitative) Date achieved 07/07/1999 12/31/2003 12/31/2005 06/30/2006 Comments (incl. % The initial target of one survey per year has proven unachievable, even not achievement) relevant. By MTR the target was reduced from 4 to 2 surveys in all. 25 Annex 2. Restructuring (if any) Not Applicable 26 Annex 3. Project Costs and Financing (a) Project Cost by Component (in USD Million equivalent) Components Appraisal Actual/Latest Percentage of Estimate (USD M) Estimate (USD M) Appraisal COMMUNITY INVESTMENTS -- PARTICIPATION 1.80 1.30 72.22 COMMUNITY INVESTMENTS: INFRASTRUCTURE 14.40 18.80 130.56 COMMUNITY INVESTMENTS: EARLY CHILD DEV. 4.50 6.60 146.67 POVERTY MONITORING 1.70 1.00 58.82 ADMINISTRATION 2.00 3.00 150.00 Total Baseline Cost 24.40 30.70 Physical Contingencies 2.60 Price Contingencies 1.10 Total Project Costs 28.10 Front-end fee PPF 0.00 0.00 0.00 Front-end fee IBRD 0.00 0.00 0.00 Total Financing Required 28.10 30.70 (b) Financing Appraisal Source of Funds Type of Actual/Latest Percentage of Cofinancing Estimate (USD M) Estimate (USD M) Appraisal Borrower 1.10 0.90 81.82 INTERNATIONAL DEVELOPMENT 26.20 27.20 103.82 ASSOCIATION 27 (c) Disbursement Profile 28 Annex 4. Outputs by Component Part A or Component 1. Community activities and investments Sub-component 1 -- Mobilization and strengthening of community capacity Rating: Highly Satisfactory The process 1.1. 1 The process starts with the choice of communes to be served, that will be included in each annual work program approved by the Executive Committee of Twitezimbere. The selection criteria are the poorest communes or the communes with the least economic or social infrastructure based on the poverty index developed by the ISTEEBU, and the communes with the highest proportion of returnees in the population. Twitezimbere selects and recruits qualified NGOs or other entities that will work as Agences d'Encadrement for project implementation. 1.1. 2 The process of community mobilization can be summarized as follows. Twitezimbere and/or the Agence d'Encadrement organize public meetings to inform the population about the project and to seek their views on development of their commune. On the basis of these consultations, the population elects a community development committee (CDC). Twitezimbere and/or the Agence d'Encadrement then work with the committee to produce a community development plan that lists their priority investments. The community development plan is then validated through a public meeting, and approved by the local Authority or Administrateur de Commune. The CDC mobilizes the population's contribution in labor and materials, monitors construction and maintains their investments. The participatory process brings the major ethnic and social groups together to work on investments that are in their common interest and, in this way, supports national reconciliation. Activities 1.1. 3 At the beginning of the project, an issue was that Twitezimbere staff seemed more interested in building infrastructure, and in particular water supply sub-projects, rather than in getting the community mobilization started. The Bank approved only a small amount of such infrastructure to respond to government requests for provision of water supply to some health centers built under BURSAP I. There was clearly a trade-off, over the medium term, between delivery of community infrastructure in the field and deepening of community participation. In 2000, Twitezimbere started to build capacity for community mobilization and decision making through intensive training of mobilizers (Agences d'Encadrement). The aim was to increase capacity of community mobilization in 2000 so that the project could accelerate delivery of infrastructure in later years. 1.1. 4 In 2000, Twitezimbere successfully launched the community mobilization process in rural areas. First, Twitezimbere formed 5 teams of mobilizers and trained them through a learning-by-doing exercise in the field. Then, the teams mobilized the population in 8 29 communes, by organizing community meetings where they encouraged community members to articulate their problems and identify alternative solutions. Fortunately, Twitezimbere was able to persuade people from all ethnic groups to participate in the meetings. The process concluded with agreement on community development plans that list sub-projects by order of priority and on community development committees to guide execution and maintenance. Twitezimbere took community participation further than originally planned by direct election of the development committees, a very desirable step. Twitezimbere formalized this innovation within the manual of procedures and sought the approval of Government for this change. Twitezimbere was able to integrate local administration, and in particular the Administrateurs de Commune, into the participatory process. Moreover, early on the Ministry of Plan decided that all community projects should be identified through a participatory process. Twitezimbere was also able to persuade the UNDP and the UNHCR to accept the identification of sub-projects selected through this participative process. Coordination efforts with other donors have led to agreement with the UNDP to finance several sub-projects that appear on the plans but could not be financed through the IDA credit, including a half million dollar potable water project in Kirundo province. In 2000, the participatory process was completed in 16 communes and begun in a further 3 communes, nearing the objective of 20 communes for the calendar year. 1.1. 5 In 2001, Twitezimbere contracted individuals for undertaking rapid rural appraisals, establishing 15 fully operational teams. The project supported 41 rural appraisals (for 41 communes) from its launch in January 2000 to end-February 2001. This means that the project had already begun to serve about a third of the communes inBurundi and was running well ahead of schedule. By June 2001, Twitezimbere had achieved a striking success by mobilizing 68 communes, more than half the communes in the country, during the first 18 months of project execution. The magnitude of this achievement should not be underestimated, considering that the country was engaged in a complex civil and ethnic conflict. The project contributed to peace building as well as development by engaging all groups in activities that serve their common interest. Projects supported by the IFAD, Action Aid and others have also mobilized a sizeable number of communes and established community development committees. Therefore, by the end of 2001, most of the communes in the country had been mobilized, with only the most unsecured communes remaining un-mobilized. The completion of so many plans was a major success because Twitezimbere could now finance the priority investments chosen by the population. 1.1. 6 Impressively, by the end of 2002, Twitezimbere had completed the community mobilization process. Out of 116 communes, 93 CDCs had submitted their Community Development Plans. IFAD and other donors, using the same methods, had mobilized nearly all other communes. Only a few (4) particularly conflict-stricken communes have not been mobilized. The completed plans have formed a basis for financing of small scale- community investments by Twitezimbere and by other donors, such as IFAD, the UNDP and the European Community. 1.1. 7 In connection with the supplemental financing signed in April 2003, the project had initially targeted the 65 communes containing the poorest and most vulnerable groups. 30 The number of supported communes was later increased to 73 including the remaining 4 war-affected communes that were finally mobilized in 2005. 1.1. 8 Twitezimbere assisted CDCs to update development action plans. The project developed or updated community development plans in 93 communes to reintegrate the relatively important numbers of war-affected persons, including returnees and internally displaced people. In the communes, CDCs were restructured to integrate returnees, internally displaced persons, as well as demobilized ex-combatants. A few comments 1.1. 9 The updated action plans were validated by multiple partners such as: beneficiary representatives, CBOs, local and international NGOs as well as churches, members of Parliament and Senate, local administrations, etc. 1.1. 10 The Bank showed flexibility in adapting to changed circumstances. In December 2001, the Bank agreed to a request from Twitezimbere to reduce the minimum contribution from the beneficiary population from 12% to 7% of sub-project cost because of the drought, the conflict situation and the population movements; the reduction was to be for one year, up to the end of 2002. With the supplemental financing, the financing ceiling was increased from US$75,000 to US$100,000 per commune to take into account increases in costs and to allow for the financing of small rural roads. At the end of 2005, the ceiling was further increased to US$125,000 for the poorest communes that could not mobilize the additional financing necessary to carry out their priority sub-projects. 1.1. 11 Participatory exercises were organized in close collaboration with other partners involved in the community development area, including the IDA-financed Public Works and Employment Creation Project and the IFAD-financed Rural Development Recovery Project. 1.1. 12 In connection with the community mobilization and participation, Twitezimbere carried out many different activities, including: (i) information meetings with beneficiaries on community sub-projects in the priority communes; (ii) sensitization meetings with partners at the level of the provinces; (iii) follow up workshops with Agences d'Encadrement; (iv) training of members of the CDCs; (v) workshops for self-evaluation by CDCs; and (vi) inter-province visits for exchange of experience between CDCs. In this context, number of provinces have set up provincial action plans based on the outcome of Twitezimbere-led participatory processes in respective communes. 1.1. 13 Technical and social audits were generally positive regarding the quality of community mobilization and participation, but they did point out a few weaknesses. For example, in 2002 the technical and engineering audit reported that Commune Administrators were not as engaged in the sub-projects as intended in project design, and that this should become an area for improvement over the next year. The technical and social audit found that the share of women in the highest (commune level) CDCs was smaller than in the lower (colline and sous-colline level) committees. 31 1.1. 14 With regard to community participation, the Mid-term Review Report prepared by Government (October 2002), which favorably reviews project performance and reserves particular praise for its participatory approach, recommended that Twitezimbere conduct participative evaluations -through public meetings- at the completion of each infrastructure sub-project. The aim would be to gather quick feedback on the quality of the sub-projects. Achievements 1.1. 15 The election by the populations of the communes of community development committees was quite an achievement, since at the time Burundi did not have elections or local government but rather a highly centralized administration (administrateurs de communes were local representatives of the central government). 1.1. 16 Twitezimbere prepared methodology and terms of reference, and was the leader, for community consultation for the preparation of the PRSP. 1.1. 17 Twitezimbere contributed to exchange of experience on community participation in connection with other projects financed by IDA or other donors in agriculture, rural development, education, roads and public works. Conclusions 1.1. 18 Most related project development objective (PDO)'s indicators have been achieved. In particular, the project has succeeded in implementing Community Development Committees and Community Development Plans in all the communes in Burundi. Twitezimbere has been involved and has taken a leadership role in the design and implementation of the process of PRSP community consultations. BURSAP outcomes have paved the way for the Government policy for decentralization. Lessons learned have proven useful for the preparation of a new IDA follow-up operation (social and community development project under preparation). Moreover, the recently approved legal framework (Loi Communale) includes Community Development Committees (Comité Communal de Développement Communautaire) which will be put in place shortly, integrating former CDCs elected under BURSAPII. Sub-component 2 -- Contracting for infrastructure Rating: Satisfactory 1.2. 1 Rapid progress in community mobilization paved the way for future progress in realization of infrastructure sub-projects, since the mobilized communes have selected sub- projects for project financing. Twitezimbere finances the priority investments that appear in the community development plans, up to a maximum cost of US$75,000 per commune (the investments in each commune are referred to as a single sub-project). 1.2. 2 In 2000, Twitezimbere started work in 3 communes on 15 sub-projects that were identified as priorities in their community development plans, but the process accelerated 32 quickly. As shown in the table below, the project has completed a large number of relatively small water sub-projects but only a few larger sub-projects such as schools and health centers. During the whole BURSAP II, Twitezimbere financed and completed 1,219 sub-projects, as follows: Table: number of completed sub-projects by type and by year Type of sub- 2000 2001 2002 2003 2004 2005 2006 Total Cost / project sub- project US$ (1) Water spring 10 121 416 72 0 0 619 485 Water supply 3 4 5 4 9 8 5 38 18,984 Community college 1 9 32 16 25 27 4 114 37,164 Primary school 0 9 43 21 22 34 12 141 47,149 Housing for 0 2 8 4 2 3 4 23 26,841 teachers Health center 0 6 13 37 11 10 17 94 35,920 Shelter for pre- school circle 0 0 16 21 17 38 86 178 14,934 Rural roads & small bridges 0 1 0 0 0 9 0 10 890 Social center 1 0 0 0 0 1 0 2 6,633 Total 15 152 533 175 86 130 128 1,219 15,308 (1) Average cost per sub-project in US$ equivalent 1.2. 3 The main constraint on execution has been the difficulty of raising community contribution of 12 percent of sub-project costs in labor, local materials or cash. The constraint arose from the project mandate to serve the poorest communes first, especially since these were suffering from a drought and subsequent shortage of food. There was clearly a tension between the requirement of a 12 percent community contribution as a means of assuring that the community truly wants and will maintain the sub-project, and the objective of targeting the poorest communes. This tension has been worsened by drought, severe malnutrition, and even famine in several of the poorest communes, especially in the northern provinces. To ease this tension, Twitezimbere and the Bank agreed to reduce the community contribution to 7 percent for regions where the World Food Program had formally declared a famine. Furthermore, the Bank agreed that Twitezimbere could pay for transport of materials (bricks, gravel, sand, wood) contributed by the community. Actually, communities contributed on average about 8 percent of the costs of sub-projects, mostly in labor and raw materials such as bricks, gravel and sand, compared to the objective of 7 percent. 1.2. 4 Another constraint which manifested itself in recent years was the sizeable increase in construction costs, with many communities unable to fill the gaps between the sub- project costs and the allocation of US$75,000 per commune. To address this issue, the Bank agreed to increase the financing ceiling per commune to US$100,000, and later on to US$125,000. 33 1.2. 5 Technical and social audit reports commented that the engineering and physical quality of sub-project works was generally good. The weaknesses in construction resulted from the absence of sufficient supplies of dry wood for construction (mainly because of the disappearance of mature forests) and from problems with tile roofs. The technical specifications were modified to include metal frameworks and sheet-metal roofs. The audits recommended a doubling of the number of Twitezimbere engineers so that Twitezimbere could better supervise construction without having to rely on the assistance of a consulting firm. The audits also found several occasions where there was disagreement between the Community Development Committees (CDCs) and construction contractors over payment of workers and the speed of construction; they suggested that these disagreements could be resolved by increasing the number of site visits by Twitezimbere engineers. It should be noted that delays in beneficiary contributions were often the reason why contractors did not respect the implementation schedules. The social component of the audit also suggested that Twitezimbere will need to work with communities to organize a greater effort to maintain the works. 1.2. 6 The Mid-Term Review report prepared by the Ministry of Planning with help from a consultant favorably reviewed project performance and suggested some technical improvements in execution. Turning to the physical quality of the infrastructure, the report recommended that Twitezimbere train artisans in construction skills, and especially in tile and brick production. Finally the report recommended that Twitezimbere complete its preparation of maintenance manuals of infrastructure (manuals in Kirundi language were actually prepared) and train maintenance volunteers named by the community development committees. 1.2. 7 Beneficiary communities are generally quite satisfied with the infrastructure. They would have liked to be associated with bid evaluation and contract awards but, because the work was done in Bujumbura, they were not always involved. Also, in some cases the financing ceiling was not high enough to include in the sub-projects all the items that the beneficiaries would have liked to have, such as water, or beds and mattresses in health centers, or housing for teachers. The lack of water in many health centers and pre-school circles is certainly a major problem. 1.2. 8 Senior officials of the Ministry of Education claimed that they were not consulted by Twitezimbere on the selection of primary schools and community colleges to be rehabilitated or reconstructed, and they criticized the quality of the infrastructure works as being below standards (they mentioned the poor quality of the materials used, and contractors cheating on quantities). Those comments are rather surprising. The Ministry of Education, like other ministries, has a focal point for BURSAP that is kept fully informed by Twitezimbere of its activities in the education sector. Also, the Provincial Education Directorates and the Bureau de la Carte Scolaire have always approved the selection of education sub-projects. The Directorates have been associated in the supervision of the execution of sub-projects; in particular, the Provincial Directors or their representatives have always signed the minutes of the "checking and acceptance of work upon completion", an indication that they did not have major problems with the quality of the works. 34 Sub-component 3 -- Early Childhood Development Rating: Satisfactory The process 1.3. 1 The sub-component consists of investments in early child development (ECD) by supporting services to pre-school children in the areas of nutrition, prevention of infectious diseases such as malaria, and informal pre-school education. As part of its community participation effort, Twitezimbere holds community meetings to identify the main problems of children under 6 and to identify solutions. Then, the communities choose committees of mothers to act on several of the proposed solutions with support from the project. Twitezimbere mobilizes and trains groups of mothers that provide basic social services to pre-school children with the assistance of volunteer coordinators (encadreurs). The Belgian Grant 1.3. 2 One important characteristic of the sub-component is that it attracted additional financing. In November 2001, the government of Belgium made available US$882,000 for parallel financing of the early child development (ECD). The Bank was administrator of the grant funds provided by the Belgium Directorate General for International Cooperation (DHIC). Activities 1.3. 3 Twitezimbere first started pilot ECD programs in 4 communes, and by the end of the project had extended these activities to 30 additional communes. In total, 34 communes located in 9 out of the 17 provinces of the country, or about one-third of the number of communes in Burundi, have been served. The sub-component reached about 450,000 children under age 6. 1.3. 4 The implementation of the sub-component included: Regarding nutrition a) The monitoring of children's nutrition and the recording of their baseline information on monitoring cards. b) The promotion of production and preparation of micronutrients at community level in the 9 ECD provinces. Twitezimbere distributed seeds for vegetable gardens to the zone leaders (who, in turn, distributed them to every family in their zone with a child under age 6). Twitezimbere promoted household gardens as a means of supplementing the nutrition of pre-school children. It supplied seeds and instruction for growing carrots, mushrooms, cabbage and soybeans, which have been well accepted and were produced in all 34 ECD communes and beyond. About 3,700 local trainers have been trained in nutrition. 35 Regarding health c) The preparation and distribution to all the health centers and chefs de zone and members of the committees at the colline level of an educational booklet on health. d) The distribution of de-worming medicines (which, incidentally, was effective in persuading parents to bring their children to the health centers). e) IEC activities and the promotion and distribution of about 266,000 mosquito nets to 140,000 households in 20 communes. f) The start of a supply chain providing 83 rural health centers with essential drugs to prevent morbidity and mortality due to malaria, acute respiratory infections, worms, and diarrhea. g) The implementation at the commune level of community drugstores and elected health committees. Regarding the development of children's cognitive abilities h) The establishment of informal pre-school circles to support development of the children's cognitive abilities. A total of 196 informal pre-school circles (including small shelters for 178 of them) hosting 16,450 children have been built and are functioning in the 34 ECD communes, with members of mothers' committees (2-4 per circle) elected, trained and appointed as trainers. Regarding training i) The preparation of training manuals for the health, nutrition and psycho-social activities (and in particular the preschool circles). j) With technical support from the Ministries of Agriculture and Health, about 110,000 persons in 34 communes have benefited from training/IEC workshops in nutrition and health. k) The training in health and nutrition of community leaders, civil society members and local supervisors; over 12,500 kits of training materials have been distributed. l) The organization of a workshop on ECD for medical doctors, social workers and local administrators. Regarding mainstreaming m) The design of a national integrated ECD policy, with the objective of adapting and diffusing at the country level the first pre-school model developed in the country by Twitezimbere. n) The strengthening of the capacities of the ECD unit at the Ministry of Education. Regarding other aspects o) The preparation of an M&E framework and tools for the ECD sub-component. p) The procurement of the goods necessary for execution of health and nutrition activities, in particular, medicines, seeds and fertilizer. 36 q) The purchase of 30 minutes of national radio per week for delivery of messages for health, nutrition and other community activities, and for further training of the mothers' committees. r) The financing of baseline studies on health, nutrition and psycho-social development for children under 6 in the 34 ECD communes, and the setting of baseline indicators. 1.3. 5 There were implementation delays, particularly regarding the psycho-social services (preparation of curriculum for training trainers and of a guidebook for activities) and the distribution of mosquito netting. The delays can be explained by the demands that rapid execution of community mobilization put on staff. To address the issue, and in order to increase its supervision efforts, Twitezimbere recruited a staff member with skills in psycho-social development whose primary responsibility was implementation of the ECD sub-component. That staff has been posted in Ngozi, a province with many ECD communes. 1.3. 6 Senior staff of the Ministry of Education commented that Twitezimbere did not work closely enough with the Ministry and did not properly supervise activities. They also claimed that the mothers have not shown much appreciation for the work of the volunteer coordinators, who are now asking the Ministry for some form of compensation for their work. UNICEF has supported similar activities in neighboring communes, using young trainers with small wages.Their contention is that a program that relies so much on voluntarism is not sustainable. Achievements 1.3. 7 The ECD sub-component served a substantial number of children at a relatively low cost. According to school directors and teachers, the best students in the lower classes of primary school are often those that participated in the program. The approach was well accepted and the program was much in high demand. In a sense, the Government adopted it when it decided that health care for pregnant women and children of less than 5 years would be free. The ECD sub-component was an important innovation in the project, a move away from the focus on social and economic infrastructure toward delivery of social services. It showed that a social fund can mobilize communities to work with the formal health and education structures to supply social services, rather than limit itself to bricks and mortar. 1.3. 8 Based on the increasing demand and with technical support from UNICEF and from Twitezimbere, Government has launched a planning process aimed at extending an integrated ECD program at the national level. The curricula developed by Twitezimbere's consultants has been approved by the Ministry of Education and is now being used in the ECD pre-school circles as well as other private and public pre-school classes in the country. However, the Government is confronted with the challenge of resource mobilization. With the government decision for universal access to a free primary education, the government's commitment to pre-school education may be reduced. 37 Part B or Component 2 -- Poverty Monitoring Rating: Unsatisfactory 2.1 This component consists of poverty monitoring, through statistics collection, execution of surveys, and training of statisticians. The major strategic choice made was for smaller-scale statistical monitoring. BURSAP I had financed a relatively large-scale survey to produce detailed baseline information on poverty. In contrast, BURSAP II will finance less expensive surveys that can be repeated to provide up-to-date follow-up information on the evolution of poverty. Furthermore, those surveys will be oriented towards action to reduce poverty rather than simply gathering analytical information about poverty. To that end, the surveys will provide information on the condition of roads, water supply, health and education services, and local institutions that is necessary for planning reconstruction programs. The relevant indicators will be selected through a survey of statistics users in the Government among Donors and NGOs. The process 2.2 The PAD did not include much information about the content of the component. However, the project execution plan scheduled one statistical survey per year, and the training of Burundian statisticians. There have been significant delays, and actual outputs are quite disappointing. In addition to the completion of the 1998 priority survey, they consist of only two rapid surveys that were undertaken (over a period of five and a half years) and, with the exception of some short-term training, no Burundian benefited from comprehensive training as a statistician. 2.3 One reason for the delay in starting the first rapid survey is that the ISTEEBU could not start a new survey when it had not yet published the results of the 1998 priority survey that had been financed under BURSAP I. The 1998 priority survey was to form a basis for future surveys by providing a basis for comparison. The results of the priority survey were finally published in February 2001; this delay severely limited its usefulness. More generally, there have been delays in production of routine statistics and complaints from several donors. Activities 2.4 Following the visit to the Bank of two ISTEEBU staff members to make final correction to the data, the results of the 1998 priority survey were published (in CD-Rom and through posting on the Bank's internet site) in February 2001. 2.5 Consultants from AFRISTAT visited Burundi to advise the Bank on the underlying reasons for the near paralysis of the ISTEEBU, carry out an evaluation of the ISTEEBU's training and institutional needs, and to clarify the reasons for the lack of progress and propose solutions and ways to build capacity. An AFRISTAT staff produced a report evaluating the statistical system of Burundi. The report noted that the ISTEEBU budget was clearly insufficient and that salaries were rather low so that staff had no incentive to 38 stay with ISTEEBU. Its recommendations to increase budget and salaries were beyond the scope of BURSAP II. Instead, the Bank recommended a retraining program for statisticians by the Faculty of Economics at the University of Burundi, in consultation with the ISTEEBU. 2.6 Starting in June 2001, the Bank asked the ISTEEBU to launch a short and rapid survey (Questionnaire Unifié pour les Indicateurs de Développement -- QUID) to update the main indicators on living conditions and poverty. The survey agents conducted their field work from January 20 through April 20, 2002 and all of them returned safely to Bujumbura. Late delivery of electronic scanners and training of staff in the QUID methodology were responsible for some delays. The ISTEEBU entered the data into a computer memory through scanning and ran quality checks. Finally, the ISTEEBU issued its report on the results of the survey in 2003. 2.7 In 2005/2006, another survey (Questionnaire Unifié pour les Indicateurs de Base du Bien-être -- QUIBB) was carried out, using the expertise of the Centre Universitaire d'Etudes Economiques et Sociales (CURDES) ­ a research center of the University of Burundi­ with the assistance of AFRISTAT's consultants and under the supervision and control of ISTEEBU. A final report on the QUIBB, together with the results of a complementary survey on prices, were issued in June 2006. 2.8 The training of Burundian statisticians never began, despite years of effort by Twitezimbere to work with the ISTEEBU and then the Faculty of Economics and Statistics at the University of Burundi. There was no progress in launching a training program for statisticians because of misunderstanding between ISTEEBU and the Faculty of Economics at the University of Burundi, which was pressing for an overly ambitious program. Comments 2.9 Like most public institutions, ISTEEBU was severely hit by the longlasting civil unrest; it had a very weak management and a lack of qualified statisticians. It was very difficult for Twitezimbere to undertake proper monitoring of the statistical component by directly contracting ISTEEBU's consultant services. One of the constraints to improvement was that the Bank budget for supervision did not allow adequate supervision of the component. It turned out to be difficult to find and fund a statistician to supervise the component and to provide technical advice to the ISTEEBU. 2. 10 Although there was general agreement that BURSAP resources could be used to update social indicators and the poverty profile as required by the PRSP process, the ISTEEBU failed to deliver. To some extent, this component's objectives overlap with other donor-funded operations (UNDP, European Union (EU), and the African Development Bank (AfDB)) as well as one recently approved IDA Project (Economic Management Support Project). The resulting lack of hard data on economic and social conditions is impeding project monitoring, government fiscal planning, and preparation of projects by donors. 39 Conclusions 2. 11 The results of the 1998 priority survey have been used by Twitezimbere for the selection of communes. Although results of 2002 QUID survey have been used with delay, they have proven useful for other Bank supported analytical work such as preparation of the interim PRSP (2005) and poverty assessment (2005), Education Country status Report (2006); Health Sector Note (2006). As for the QUIBB 2005 survey, the report was delivered in June 2006 and the results have not been used so far. 2. 12 There has been basically no progress in building the capacity of the ISTEEBU. Over the years, the ISTEEBU has not been very diligent to prepare a program for training of its staff or any technical proposals for future surveys as had been agreed. Part C or Component 3 ­ Administration Rating: Highly Satisfactory 3.1 Overall, the management of the project by Twitezimbere has been highly satisfactory. Twitezimbere has handled the community investments component very efficiently; its management of the poverty monitoring component has not been a great success, but this was probably something beyond its control. 3.2 The Executive Committee and the General Assembly of Twitezimbere met regularly to approve the programs of work and activity reports for the project. 3.3 To carry out activities, Twitezimbere contracted with qualified NGOs or other entities to work as coordinating agencies (Agences d'Encadrement), with small and medium enterprises for the construction of infrastructure and with consulting firms for the design and the supervision of works. 3.4 Various audits have documented a constant improvement in performance in technical expertise in community development and in fiduciary management. The external financial audits of the accounts have been unqualified. Procurement reviews did not find any significant problems, and the technical and social audit concluded that procurement rules have been respected. Technical audit reports assessed the quality of the infrastructure as generally good. 3.5 The project had the highest performance rating among 10 projects in the review of the fiduciary management of the IDA portfolio in February 2006. 3.6 The IDA credit and supplemental grant as well as the Belgian Grant have been fully disbursed. 40 Annex 5. Economic and Financial Analysis (including assumptions in the analysis) A cost-benefit analysis of the Project is not feasible because of the difficulty of measuring benefits in a rural economy where data are scarce and where goods, services and labor are often not traded and, therefore, are difficult to value in money terms. However, a summary analysis of a few costs shows that Twitezimbere has achieved a low cost of administration, and that unit costs of construction are lower than for other projects. The figures show that BURSAP II has been a cost efficient project. Costs of Administration Annex 3 on "Project Costs and Financing" shows that the actual/latest estimate of the cost of component 3 on "Administration" is about 10% of the total cost of the Project. The cost of administration includes operating expenses and many other expenditures such as audits and capacity building, so that operating expenses are less than 10% of the total project costs. There are no recent comparative figures readily available for social funds in Africa , but usually ratios of 10-12% are considered an indication of good performance. Twitezimbere claims that operating expenses are equal to about 2% of the amount disbursed for the infrastructure, and about 1.4% of total disbursements, but Twitezimbere has not provided satisfactory evidence of such low percentages. Nevertheless, there is no question that Twitezimbere is an efficient organization. Construction Costs The table below shows that, as was already the case for BURSAP I, unit costs of construction, per square meter, of primary schools, community colleges and health centers are lower than for other projects. Type of Cost per square meter - in FBU Infrastructure Twitezimbere ABUTIP Belgian-Burundian Fund Primary school 110,993 162,427 n.a. Community college 116,591 n.a. 168,800 Health center 184,564 300,255 n.a. Note: The areas are the same for primary schools (646.2 square meters) and health centers (666.1 square meters), but not for community colleges (544.1 square meters for a community college by Twitezimbere and 1,022 square meters for a community college by the Belgian-Burundian Fund). 41 Annex 6. Bank Lending and Implementation Support/Supervision Processes (a) Task Team members Names Title Unit Responsibility/Specialty Lending Agnes Albert-Loth Senior Finance Officer LOAG2 FMS Anne Anglio Senior Program Assistant AFTH3 Program Assistant Siaka Bakayoko Senior Financial Management Specialist MNAFM FMS Pascale Helene Dubois Senior Institutional Integrity INT Lawyer Pamphile Kantabaze Senior Operations Officer AFTH3 Health/Education/ECD/CDD Anne Kielland Consultant AFTH2 ECD Markus Kostner Country Program Coordinator MNCA4 Community Development Essimi Menye Consultant AFTQK Statistics T. Mpoy-Kamulayi Lead Counsel LEGAF Lawyer Susan Opper Senior Education Specialist AFTH3 Education/ECD Christine Lao Pena Senior Human Development Economist LCSHH ECD Menahem M. Prywes Senior Economist ECSHD TTL/Poverty Rene E. Welter Consultant AFTPC Procurement Supervision/ICR Diane Coury Consultant AFTH2 Operations, ECD Jacinthe Desmarais Consultant AFTH3 Early Childhood Development Bella Lelouma Diallo Senior Financial Management Specialist AFTFM FMS Yvette Laure Djachechi Senior Social Development Specialist AFTS3 Environment & CDD Michael P. Fowler Senior Finance Officer LOAG2 Disbursement Pamphile Kantabaze Senior Operations Officer AFTH3 Health, Education, CDD Eric Mabushi Economist AFTP3 Community Development Myrina D. McCullough Consultant AFTH3 Operations Isabella Micali Drossos Senior Counsel LEGLA Counsel Joseph Kizito Mubiru Senior Financial Management Specialist AFTFM FMS Prosper Nindorera Senior Procurement Specialist AFTPC Procurement Deo-Marcel Niyungeko Municipal Engineer AFTU2 Community Infrastructure Seraphine Nsabimana Program Assistant AFMBI Program Assistant Susan Opper Senior Education Specialist AFTH3 ECD, Education Menahem M. Prywes Senior Economist ECSHD Poverty Economist Emmanuel Tchoukou Financial Management Specialist AFTFM FMS Jean-Paul Zoyem Consultant AFTPM Poverty Monitoring 42 (b) Ratings of Project Performance in ISRs No. Date ISR Archived DO IP Actual Disbursements (USD M) 1 03/13/2000 Satisfactory Satisfactory 0.50 2 09/13/2000 Satisfactory Satisfactory 0.50 3 12/21/2000 Satisfactory Satisfactory 0.84 4 03/30/2001 Satisfactory Satisfactory 1.06 5 06/25/2001 Satisfactory Satisfactory 1.53 6 12/26/2001 Satisfactory Satisfactory 3.09 7 05/09/2002 Satisfactory Satisfactory 4.85 8 06/03/2002 Satisfactory Satisfactory 5.38 9 11/07/2002 Satisfactory Satisfactory 7.39 10 06/10/2003 Satisfactory Satisfactory 9.86 11 12/23/2003 Satisfactory Satisfactory 11.82 12 06/29/2004 Satisfactory Satisfactory 13.76 13 12/21/2004 Satisfactory Satisfactory 16.35 14 05/16/2005 Satisfactory Satisfactory 19.21 15 12/13/2005 Satisfactory Satisfactory 25.12 16 06/29/2006 Satisfactory Satisfactory 27.84 (c) Staff Time and Cost Staff Time and Cost (Bank Budget Only) Stage of Project Cycle No. of staff weeks USD Thousands (including travel and consultant costs) Lending FY99 87.44 FY00 24 45.88 FY01 0.00 FY02 0.00 FY03 0.00 FY04 0.00 Total: 24 133.32 Supervision/ICR FY99 8.02 FY00 15 27.50 FY01 26 61.29 FY02 26 82.80 FY03 28 81.77 FY04 18 0.00 FY05 18 12.71 FY06 13 19.47 FY07 5 31.64 Total: 149 325.20 43 Annex 7. Detailed Ratings of Bank and Borrower Performance Bank Ratings Borrower Ratings Ensuring Quality at Entry: Satisfactory Government: Moderately Satisfactory Quality of Supervision: Satisfactory Implementing Agency/Agencies: Highly Satisfactory Overall Bank Overall Borrower Performance: Satisfactory Performance: Satisfactory 44 Annex 8. Beneficiary Survey Results (if any) Since this is a Core ICR, there has not been any beneficiary survey at the end of the Project. However, in early 2006 an individual consultant carried out a social audit of BURSAP II. The final report of the social audit issued in April 2006, indicates that the communities are satisfied with the Project. The populations had the following comment: "Only Twitezimbere has not abandoned us. Twitezimbere stayed with us and helped us very much". The following are some of the findings of the social audit: a) The process of community participation, with the establishment of community development committees (CDC) in all the communes, is a great achievement, considering the country circumstances and the socio-political situation. b) The CDC could be used by all donors that want to rely on community participation. However, the functioning of the CDC is sometimes difficult, particularly in cases where they are not fully accepted by local authorities. The beneficiaries commented on the fact that there are many committees, but they prefer the CDC because of their concrete realizations. In some cases, the lack of diversity with an insufficient representation of women and youth was cited as a problem. Another problem was the workload and lack of resources of community leaders who volunteer their time for the benefit of the community. c) All infrastructure financed by Twitezimbere are priorities of the populations, which contributed on average about 8.5% of the sub-project costs in the form of local construction materials or in cash. However, because of the lack of resources, the completed sub-projects are not sufficient; there is a need for more infrastructures in view of the increase in population. The beneficiaries criticize the absence of water, particularly in health facilities and the pre-school circles. The communities would like to be better informed of the day-to-day monitoring of construction done by the Agences d'Encadrement; they would also like to see a greater field supervision done by Twitezimbere staff. d) Efforts have been made by the beneficiaries to maintain the completed sub- projects, through the establishment of a maintenance sub-committee (of the CDC) and a maintenance fund to pay for small repairs, or through the recruitment of a guard paid by the community. Also, communities have tried to establish the principle that "whoever destroys or breaks should pay". However, the importance of maintenance is not understood and appreciated by everybody. The early child development activities are greatly appreciated by the populations for the obvious reasons that they contribute to a satisfactory physical and intellectual development of children and for the few advantages that they provide to families (drugs, vaccines, and agricultural inputs). The pre-school circles are overpopulated; there is a need for more circles and more encadreurs. Questions related to the "compensation" to be given to the volunteer coordinators (encadreurs ) and provision of snacks for the children have been recurring problems because of the extreme poverty. 45 Annex 9. Stakeholder Workshop Report and Results (if any) Not applicable - Since this is a Core ICR, no stakeholder workshop was held at the end of the project. 46 Annex 10. Summary of Borrower's ICR and/or Comments on Draft ICR REPUBLIC OF BURUNDI MINISTRY OF DEVELOPMENT PLANNING AND NATIONAL RECONSTRUCTION MANAGEMENT UNIT OF TWITEZIMBERE ASBL (A Not-For-Profit Organization) SECOND SOCIAL ACTION PROJECT COMPLETION REPORT SUMMARY FINAL VERSION JUNE 2006 47 SECOND SOCIAL ACTION PROJECT SUMMARY OF COMPLETION REPORT I. INTRODUCTION I.1. Mission framework The Government of the Republic of Burundi has benefited primarily from IDA financing for the implementation of its Second Social Action Project (BURSAP II), which is governed by the following financing agreements: (i) IDA credit agreement No. 3287-BU of October 28, 1999 in the amount of SDR 9 million; and (ii) IDA grant agreement No. H033-BU of April 14, 2004 in the amount of SDR 10.3 million. The project seeks to: (i) encourage the various community groups to work together, within the context of crisis recovery, for the selection, partial financing, monitoring and evaluation, and conduct and sustainability of priority and common interest activities, and (ii) help the poorest and most vulnerable groups gain access to better economic and social services, including social welfare, through incomes, regular employment, and social facilities, thereby reducing the risk of being condemned to poverty. The credit agreement signed between the World Bank and the Government stipulates that the borrower draft the completion report in order to evaluate the project's performance and learn lessons from its implementation. I.2. Work methodology The consultant worked closely with the management team and technical departments of Twitezimbere ASBL [association sans but lucratif--nonprofit association]. He first consulted the main project-related documents in order to compare their content, at the appropriate time, with what had been observed on the ground. This was followed by thematic discussions and interviews with the various partners involved in the project's design and implementation with a view to supplementing, confirming, or refuting the findings of the project documentation. II. PRESENTATION OF THE SECOND SOCIAL ACTION PROJECT II.1. Project components The social action program itself has three components: · Social mobilization and community participation based on the design and implementation of a participatory approach for the selection, financing, execution, monitoring, and maintenance of community activities and facilities; · Monitoring of poverty, which includes capacity-building of the Institut de statistiques et d'études économiques du Burundi (ISTEEBU) [Burundi Institute of Statistics and Economic Research] and other statistics organizations to enable them to monitor and analyze poverty; and · Administration and management, which includes capacity-building of the entity 48 responsible for implementing BURSAP II and of the partners (authorities, community organizations, NGOs, and supervisory agencies) for the execution, coordination, oversight, and monitoring of project implementation. II.2. Project financing Project financing came from two primary sources: (i) IDA credit No. 3287 BU in the amount of SDR 9 million, and (ii) the IDA grant in the amount of US$14.2 million as supplementary financing. II.3. Project implementation strategies The participatory approach, which involves consultation, mobilization, and capacity- building through discussions with the people first and then with their elected representatives, will be used throughout the life of the project. This process aims at identifying development problems experienced at the grassroots level (collines), selecting the priority activities, prioritizing them, designing projects to be implemented, and delegating the responsibilities of the various partners regarding the execution, monitoring, and maintenance of the resulting subprojects. II. ACHIEVEMENTS OF BURSAP II III.1. Social mobilization Twitezimbere ASBL intervention helped to establish Community Development Committees (CDCs) and to build awareness among the people for effective participation in the planning, implementation, and sustainability of the socioeconomic infrastructure. III. 2. Rehabilitation and construction of social infrastructure Construction of infrastructure focused in particular on social sectors such as education, health, and drinking water. During the period covered by BURSAP II, the following infrastructure was constructed: (i) 159 preschool circles, (ii) 141 schools, (iii) 131 communal schools (schools designed to improve local access to secondary education), (iv) 23 apartment buildings (homes), (v) 94 health centers, (vi) 619 rehabilitated springs, (vii) 37 water supply systems, and (viii) seven trails and small bridges. II. 3. Early child development program (ECD) The ECD covers 34 communes chosen from among the poorest communes and includes activities to combat poverty by: · sensitizing people to the importance of balanced nutrition, hygiene, maternal and infant health, and child cognitive development; and · providing material and organizational support to set up mothers' committees and 49 health committees. To achieve these objectives, Twitezimbere ASBL carried out the following activities: · Construction of preschool circles; · Training of supervisors at the preschool circles; · Distribution of agricultural inputs (1,250 kg of cabbage, 1,250 kg of carrots, 125 metric tons of KCl, 125 metric tons of urea, and 500 liters of Dursban insecticide) to households with children attending the preschool circles, to supervisors, and to households with children under six from communes benefiting from ECD activities; · Distribution of medicines for the most common diseases among young children (0 to 6 years); · Distribution of 266,608 mosquito nets; and · Establishment of mothers' committees and health committees. III.4. Monitoring changes in poverty During implementation of BURSAP II, the project financed the finalization of the priority survey launched during the 1998-1999 period, the 2002 QUID survey (development indicators survey), the 2005-2006 CWIQ survey (Core Welfare Indicators Questionnaire), and the price survey supplementing the CWIQ survey. III.5. Administration and management (i) Administrative management The Government of Burundi, in agreement with the World Bank, has entrusted the execution of the social action project to Twitezimbere ASBL with which it signed an agreement outlining the corresponding administrative, financial, and managerial procedures. To implement BURSAP II activities, Twitezimbere sought assistance from NGOs and supervisory agencies to mobilize the people and monitor execution of the projects, and from SMEs to construct infrastructure. Twitezimbere also called on consulting firms to conduct various studies. With respect to the functioning of the entities, the executive committee and the general assembly met regularly to approve programs and activity reports. (ii) Financial management Generally speaking, the audits concluded that there was sound financial management of the project. The technical improvements proposed by the auditor were implemented and Twitezimbere was able to obtain disbursements without any major difficulties. There were no financing-related problems of note. The Government counterpart funds were disbursed on a regular basis. The social and technical audit confirmed that contracts were awarded on a regular basis, 50 despite a number of delays in the timetable owing to the difficulties people experienced in gathering the local materials that represented their contribution. Agreements with supervisory agencies were also signed regularly. The initial project financing, which primarily included IDA credit No.3287 (EUR 12 million), was fully disbursed, and the disbursement ratio for supplementary financing, IDA grant No. H 033 in the amount of US$14.2 million, was 94 percent. As of April 30, 2006, disbursements and pledges accounted for 98 percent of the grant. The undisbursed balance is estimated to be US$575,000. World Bank aide-mémoires and reports from various external appraisals conducted indicate that the Twitezimbere management unit achieved a relatively high level of performance. According to the World Bank assessment, the project obtained the highest performance rating of the ten projects evaluated during the latest review in February 2006 of the fiduciary management of the IDA portfolio in Burundi. IV. END OF PROJECT EVALUATION IV.1. Evaluation of the initial objectives regarding project development, design, and quality i. Initial objectives. The objectives were not modified during the project period. ii. Initial components. The components were not modified during implementation of the project. iii. Initial quality. The social action project responds to needs and is based on the participation of beneficiaries supported by liaison entities such as NGOs and local governments. An agreement was concluded between the Government and Twitezimbere ASBL, an independent organization tasked with managing the project in accordance with the criteria and procedures set out in an operational manual. The second social action project was launched under favorable conditions: The people were already accustomed to mobilizing in order to participate in and help execute the subprojects in collaboration with NGOs; · The project operational manual provided specific details on the implementation methods for subprojects to be financed by Twitezimbere; and · The monitoring system of the first social action project was improved, with the implementation of surveys to define reference situations. 2. Achievement of the objectives and outcomes (i) Achievement of the objectives Objective No. 1, which sought to encourage the various community groups to work together, within the context of crisis recovery, for the selection, partial financing, monitoring and evaluation, and conduct and sustainability of priority and common interest activities, was achieved. 51 Objective No. 2, which sought to help the poorest and most vulnerable groups gain access to better economic and social services, including social welfare, through incomes, regular employment, and social facilities, thereby reducing the risk of being condemned to poverty, was also achieved. (ii) Results by component 1. Results achieved in the community development process (a) Consideration of community priorities in selecting the projects to be implemented Decisions regarding the consideration of community priorities were made by triangulating the information sources. According to these sources, all projects executed were selected from among the priorities identified by the people. (b) Participation of beneficiaries in project implementation and maintenance of infrastructure constructed under the project Beneficiaries participate meaningfully in the implementation of the activities, accounting on average for 8 percent of the cost of the subprojects. This includes site development for the construction of infrastructure, the provision of local construction materials, and cash contributions. The sustainability of the projects' achievements depends largely on how the infrastructure is used and maintained. According to the economic and social audit report, 85 percent of the infrastructure is well maintained. Maintenance is either carried out by the people themselves (construction and maintenance of drains around schools) or with a monetary contribution to pay workers hired for this purpose. (c) Participation of women in the community development process The project encourages the full participation of women in the community development process. The participation of women is essential, be it at the level of meetings of the collines or the sectors. (d) Establishment and functioning of community development committees During participatory rural evaluations, the people elected community development committees for the different levels of their communes. These committees function relatively well, because they mobilize the people for community development activities and convene meetings to provide information and updates on the progress of the projects being executed. However, the mission of the community development committees should be expanded to include the other projects financed by other donors in the commune, provided all these 52 projects have the same objectives. Moreover, the new communal law stipulates that CDC members shall be designated by the communal authorities. This could hinder the participatory approach process which has already achieved results at the grassroots level. (e) Quality of the participatory rural evaluations The people greatly appreciate the opportunity to participate in making decisions pertaining to their development. The local government is also pleased, because participation by the people in the selection of projects to be executed facilitates their mobilization for building infrastructure. 2. Improvement of the quality of the infrastructure in place The technical audit was designed to review the quality of the infrastructure constructed under the project. It concluded that the quality was good and noted a number of improvements: sheet metal replaced tiles for roofing and steel frames replaced timber frames. Water supply systems and rehabilitated springs were well constructed. Although works are monitored by supervisory agencies, oversight must be strengthened to provide more support to enterprises building the infrastructure. Finally, it bears noting that the cost of the infrastructure constructed under the project is much lower than that of other stakeholders. 3. Evaluation of the ECD component ECD activities were launched in four pilot communes in the north of the country with the establishment of mothers' committees, the training of supervisors, the distribution of mosquito nets, vermifuges, and other medicines, as well as the provision of seeds and the launch of preschool circles. 4. Monitoring poverty and the living standards of the people During implementation of BURSAP II, the project financed the finalization of the priority survey launched during the 1998-1999 period, the 2002 QUID survey, the 2005- 2006 CWIQ survey, and the price survey supplementing the CWIQ survey. 5. Project management Twitezimbere ASBL efficiently managed the resources made available under BURSAP II. The various audits noted steady improvements with regard to technical expertise in community development and the fiduciary management of the project. Before closing the project, a grace period will be granted to settle all invoices pertaining to the goods, services, and works that will have been delivered or received by June 30, 2006, with the exception being the financial audit. 53 (iii) Impact on institutional development The project proved useful to NGOs and associations. It also had a significant impact on the Ministry of Development Planning and National Reconstruction as it provided assistance to ISTEEBU, training for statisticians at the University of Burundi, as well as training abroad for central administration officials from the ministry. (iv) Impact of the project on the beneficiary populations Health centers: Health centers were constructed using the health care map published by the Ministry of Public Health. This was a response to the concern regarding close contact with patients at the shelters, as it reduced distances covered by persons seeking health care services, and thus increased access to basic health services for the people, and the most vulnerable in particular. Primary schools: The construction of primary schools contributes to the effort to achieve universal education, a goal the country hopes to accomplish by 2010. The construction of schools in the poorest communes has helped to increase enrollment rates among children in disadvantaged areas. Shorter distances, upgraded facilities, and smaller class sizes have also helped to increase school attendance and improve success rates. According to the findings of the social audit, parents and local authorities are satisfied with the achievements in the education sector and are requesting more primary schools. Water supply systems and rehabilitated springs: Rehabilitated springs and water supply systems help to reduce the distance covered by people seeking drinking water. An increase in the number of springs reduces the time spent waiting to collect water, and, according to users, the rehabilitation of springs has improved the quality of the water, thereby considerably reducing the incidence of diseases caused by water pollution (e.g., verminoses and dysentery). Communal schools: The communal schools have helped to facilitate the move from primary to secondary schools. By constructing these schools in the most disadvantaged regions, Twitezimbere encouraged parents to send their children to school, thanks to shorter distances and lower school fees. Homes for teachers: The construction of homes for teachers established the permanent presence of teachers, thereby providing these teachers with improved working conditions to better prepare the subjects to be taught. (v) Impact on the poorest and most vulnerable groups By focusing on the poorest and most ill-equipped communities, as well as on those with the highest number of repatriated and displaced persons, the second social action project is having a positive impact on poverty reduction. Young children are also included in the category of most vulnerable persons and ECD activities are having a positive effect on their health and psychosocial development. However, the poorest people still have great difficulty accessing social welfare services, owing to their low incomes and to the fact that the cost of essential services is still high. Actions should be specifically carried out to provide social welfare services directly to the poorest and most vulnerable persons. 54 (vi) Impact on the partners In addition to the grassroots communities, the beneficiaries of this project, the main project partners are the Government, the technical departments, the authorities, the supervisory agencies, and the NGOs. All these partners welcome the participatory approach retained for mobilizing the people. The technical departments have also received support for capacity-building. The supervisory agencies have gained experience in promoting and mobilizing development, and have benefited from training sessions specifically organized for them. Small- and medium-sized enterprises have been established and have strengthened their capacity to construct social infrastructure. Considerable funds have been allocated to associations of local artisans and producers. 3. Main factors affecting project implementation and outcomes The second social action project launched its activities amid widespread insecurity in the country. Generally speaking, this situation slowed the pace of the implementation of activities at the grassroots level. In spite of these difficult circumstances, Twitezimbere ASBL staff endeavored to maintain contact with the grassroots communities. 4. Sustainability of Twitezimbere's activities The project's sustainability can be analyzed objectively at two levels: at the level of Twitezimbere ASBL and at the level of the completed subprojects. It should be recalled that during the period covered by BURSAP I, Twitezimbere continued to function during the long period of suspension of IDA credit disbursements and of funding from donors such as the HCR and the African Development Fund (ADF). The viability of Twitezimbere ASBL was therefore tested, and there was no reason why Twitezimbere should not continue to function as an NGO responsible for collecting funds from donors. With respect to the subprojects, it bears noting that none of the infrastructure constructed or restored by Twitezimbere was damaged or destroyed. This speaks volumes about the people's feeling of ownership toward the facilities they helped to build and have gone to great lengths to protect. The contribution made by the people and supported by the NGOs and local associations is a key factor for the sustained use and maintenance of the infrastructure. 5. Performance of the World Bank During the first three years of the project, the climate of insecurity adversely affected the frequency and scope of the oversight activities of Bank headquarters staff. However, the Bank's country office in Bujumbura established an ongoing presence on the ground and maintained good working relations with the Government, Twitezimbere, and the other partners. On a whole, the Bank has demonstrated its readiness to help Burundi during this difficult period, believing that it would be wise to essentially use the social action project as a means of providing assistance to the poorest people affected by the crisis. During 55 the remainder of BURSAP II, the Bank increased the number of oversight missions which helped the project achieve its objectives. 6. Performance of the Government The Government demonstrated its strong commitment to BURSAP II and Twitezimbere ASBL by not intervening in the project's daily operations as it does with certain projects, and specifically, by disbursing the agreed counterpart funds in a timely manner. Twitezimbere ASBL, the implementing entity for BURSAP II, carried out the activities for this project in a very efficient manner. With respect to the functioning of the entities, the Twitezimbere ASBL management unit produced programs and activity reports in a timely manner, and the executive committee met regularly to approve them. These documents were submitted to the Government and the donors. The general assembly also met annually to provide guidelines for the Twitezimbere ASBL management unit. With regard to compliance with the project's management procedures, the technical audit concluded that the selection procedures for the intervention zones as well as the monitoring procedures for the eligibility criteria had been followed. Selection of project infrastructure and the awarding of contracts were carried out on a regular basis. Independent auditors concluded that the project's financial management was good, as evidenced by the content of audit reports for the entire project period. In the area of oversight of project activities, Twitezimbere ASBL monitored the quality of the project's achievements. To that end, this entity called on NGOs and supervisory agencies to spearhead the participatory process and provide technical oversight of the projects. In light of the large number of projects implemented, it can be concluded that the intermediary agencies were well chosen. 7. Lessons learned The first lesson learned is that by responding to the needs of communities in a country in crisis, the BURSAP II project contributed to community solidarity, and thus to national reconciliation. Second, the implementation of BURSAP II contributed to the effective construction of infrastructure (schools, health centers, and water supply systems), to increased access to essential social services, and to job creation. Finally, it is essential that effective community participation not only ensures the timely participation of beneficiaries in the costs of the subprojects, but also the maintenance and proper functioning of Bank-financed infrastructure. 8. Observations of the partners (i) The Government and Twitezimbere ASBL On a whole, the Government is satisfied with the BURSAP II project and with the performance of Twitezimbere ASBL. This latter entity is also pleased with the role it played in increasing people's access to basic infrastructure. However, it regrets not being 56 able to consider and finance all the requests that it received. (ii) The beneficiary population Beneficiaries are in general satisfied with the implementation of BURSAP II. In terms of its performance, persons interviewed found that the various partners, namely the World Bank, Twitezimbere, and the Government, pooled their efforts to set up many essential community facilities. The lack of equipment for certain infrastructure that had been put in place by Twitezimbere ASBL, such as the lack of desk seats in schools, and of beds and mattresses at health centers, was cited by persons interviewed as the main weakness of the project. They also cited the existence of incomplete infrastructure such as health centers without morgues or incinerators, and schools without teacher housing or latrines. In this case, the resources made available to Twitezimbere ASBL for implementing BURSAP II were inadequate to construct these additional facilities. The persons interviewed also cited contractor noncompliance with the timetables for the construction of community infrastructure. Extensions of the agreed deadlines were due primarily to the delay in collecting the beneficiaries' contribution. The last negative point raised was that the representatives of associations were not always involved in the contract awarding process. In this instance, the participation of beneficiaries in the awarding of contracts was affected not only by the unavailability of a number of the representatives of the beneficiaries, but also by a lack of resources and the long distances between their locations and Twitezimbere's head office. All agreements were, however, signed by the representatives of the beneficiaries. It is nevertheless important to improve this process through decentralization and capacity-building of the community representatives. (iii) Other partners (NGOs, private sector) Overall, NGOs and associations that worked together to implement BURSAP II gave a positive evaluation of Twitezimbere's work, and expressed the wish that the Government continue to use the participatory approach in the selection, execution, and monitoring of investment actions in the country. They also commented on the standardization of grassroots development approaches used by the various stakeholders. This would be a part of the Government's general policy on grassroots community development. For example, the level of community participation should be the same for all stakeholders. The last observation made concerned the current relationship between the CDCs and the communal administrative entities established after the last political elections. According to the prevailing communal law, CDC members must be designated by the communal authority even though they had been elected using nonpolitical criteria based primarily on integrity, dedication, and competence. The NGOs and associations are asking the Government to recognize the importance of these CDCs in promoting grassroots development, to comply with procedures for their establishment, and to accord legal status to this entity. 57 V. CONCLUSIONS AND RECOMMENDATIONS 1. Capitalize on the achievements of BURSAP II In preparing future development projects, and in particular, PRADECS (Community and social development project), it is desirable to adopt the participatory and decentralized approach to development. This project should specifically reintroduce the components related to social infrastructure, reinsertion of vulnerable persons, and capacity-building in order to meet the needs of the communities in these areas. 2. Design a common grassroots development policy The various interventions carried out in grassroots communities are not harmonized and confuse the people about the need to contribute to the financing of social facilities being constructed for them. 3. Give legal status to the community development committees Pursuant to the prevailing communal law, the CDCs face an increasingly precarious future, despite their importance in spearheading development within the communes. These committees should be independent and provide technical support to the communal authorities. To date, there is no legislation that recognizes them as independent, nonpolitical entities established by communities to initiate development at the decentralized level of the commune. The CDCs should play a role in all development sectors and no similar entity established as a committee should be set up in a commune. 4. Promote rural microcredit There is a great demand by the people for microcredit to finance income-generating and job-creating activities. This component, for which Twitezimbere is responsible, is financed by the AfDB. In order to decentralize development at the grassroots level, the Government should develop a package of interventions to promote microcredit in order to increase access for households to credit. 5. Build the capacity of partners at the decentralized level Discussions with partners revealed their desire to strengthen in particular the training activities for the CDCs, the SMEs, the health committees, and the educators at preschool circles. There is also the possibility of including the grassroots administrators. The projects to be launched in rural areas should include in its priorities capacity-building activities at the grassroots level which focus on simplified modules for managing, monitoring, and evaluating projects. 58 6. Strengthen ISTEEBU departments It is desirable for ISTEEBU to be strengthened so that it can produce, on an ongoing basis, databases necessary for drafting development and poverty reduction policies. For future projects, it would be important to be able to determine, over a period to be agreed upon, the impact of the project on poverty reduction. 59 Annex 11. Comments of Cofinanciers and Other Partners/Stakeholders Not Applicable. 60 Annex 12. List of Supporting Documents 1) Project Concept Note, [month year] 2) Integrated Safeguards Data Sheet 3) Project Information Document (PID) ­ July 28, 1998 4) Project Appraisal Document on a Proposed Credit in the Amount of SDRs 9 Million (US$ 12 Million Equivalent) to the Repu blic of Burundi for the Second Social Action Project - July 7 1999 5) Development Credit Agreement (Second Social Action Project) between Republic of Burundi and IDA ­ October 28, 1999 6) Project Agreement (Second Social Action Project) between IDA and Twitezimbere ­ October 28, 1999 7) Manuel d'Opérations Techniques ­ Octobre 1999 8) Analyse Environnementale des Activités Futures du Second Projet D'Actions Sociales ­ Avril 1999 ­ Damas Nduwumwami 9) Implementation Completion Report (24940) on a Credit in the amount of SDRs7.5 million (US$10.4 million equivalent) to t he Republic of Burundi for a Social Action Project ­ June 20, 2000 10) Belgian Grant for co-financing for BURSAP II (Cr. IDA 3287-BU) ­ Early Childhood Development - Grant No. TF 024242 ­ O ctober 2001 11) Audit technique et Social ­ 2000/2001 ­ Mars 2002 ­ Bernd Dobberstein 12) Unité de gestion de Twitezimbere ­ Deuxième Projet d'Actions Sociales ­ Rapport d'évaluation à mi-parcours ­ Octobre 2 002 13) IDA ­ Proposed Supplemental Financing (In Grant) of SDR 10.3 Million (US$ 14.2 Million Equivalent) to the Republic of Burundi for the Second Social Action Project (Credit 3287-BU) ­ March 14, 2003 14) Development Grant Agreement (Second Social Action Project ­ Supplemental Financing) ­ April 14, 2003 15) Agreement Amending Agreement (Second Social Action Project ­ Supplemental Financing) between IDA and Project Twitezimb ere ­ April 14, 2003 ­ Credit No. 3287-BU & Grant No. H033-BU 16) BURSAP II ­ Financement Supplémentaire ­ Plan de Gestion des Risques Environnementaux pour les Sous-Projets du BURSAP ­ Mars 2003 17) BURSAP II ­ Financement Supplémentaire ­ Cadre d'Acquisition de Terres pour les Sous- Projets du BURSAP ­ Mars 2003 18) Rapport d'Audit Technique ­ Juillet 2003 à Août 2005 ­ 24 novembre 2005 ­ Eddy Bynens 19) Ministère de la Planification du Développement et de la Reconstruction Nationale ­ Unité de Gestion de l'ASBL Twitezim bere - Deuxième Projet d'Actions Sociales ­ Rapport d'Achèvement ­ Version Définitive ­ Juin 2006 20) Rapport Annuel d'Activités ­ 2000 21) Rapport Annuel d'Activités ­ 2001 22) Rapport Annuel d'Activités ­ 2002 23) Rapport Annuel d'Activités ­ 2003 24) Rapport Annuel d'Activités ­ 2004 25) Rapport Annuel d'Activités ­ 2005 26) PSR/ISR ­ Numbers 1 to 16 27) Back-to-office reports (BTO) and aide-mémoires of supervision / implementation support missions 28) Rapport Final d'Audit Social ­ Avril 2006 ­ Ms. Scholas Murebwayire 61 MAP 62