IEG Report Number: ICRR14658 ICR Review Independent Evaluation Group 1. Project Data: Date Posted: 06/12/2015 Country: Guinea-Bissau Project ID: P113468 Appraisal Actual Project Name: Guinea Bissau--food Project Costs (US$M): 5.00 5.00 Price Crisis Response Program L/C Number: Loan/Credit (US$M): 5.00 5.00 Sector Board: Agriculture and Rural Cofinancing (US$M): Development Cofinanciers: Board Approval Date : 09/22/2008 Closing Date: 09/30/2011 08/31/2013 Sector(s): Crops (60%); Public administration- Agriculture fishing and forestry (20%); Other social services (20%) Theme(s): Global food crisis response (100% - P) Prepared by: Reviewed by: ICR Review Group: Coordinator: Ebru Karamete William R. Sutton Christopher David IEGPS1 Nelson 2. Project Objectives and Components: a. Objectives: The Project development objectives stated in the Grant Agreement (p. 5) is " to improve food security for the most vulnerable population of the Recipient including children, and to increase smallholder rice production in the regions of Bafata, Biombo, Cacheu, Gabu, Oio, and Bissau (the “Selected Areas”)". The Project Appraisal Document statement of objectives is (p. 6): "to improve food security for the most vulnerable population, including children, and increase smallholder rice production in project areas. This review uses the Grant Agreement version. b.Were the project objectives/key associated outcome targets revised during implementation? No c. Components: 1. Support for the Most Vulnerable Population (Appraisal Estimate : US$ 1.5 million, Actual: US$ 1.5 million). The aim of this component was to support the most vulnerable population through: (i) a school feeding program and (ii) a food-for-work program to rehabilitate land under rice. Under the school feeding program the project planned to serve 14,000 students one hot meal per day and supply approximately 930 metric tons of food for both programs (school feeding and food-for-work).Also, the school feeding program planned to provide take-home rations for female students who had an attendance rate of 80 %. The food-for-work activities aimed to rehabilitate rice land by building dikes for mangrove rice on approximately 2,000 hectares and drainage channels and anti-erosion banks for lowland rice (about 3,000 hectares). The beneficiaries of this program were expected to work on rehabilitation activities generating approximately 160,000 workdays, in exchange for food rations. World Food Program implemented the component. 2. Support for Increasing Rice Production (Appraisal Estimate : US$ 3.0 million, Actual: US$ 2.3 million). The component aimed to deliver immediate benefits to food-insecure rural communities through the distribution of inputs, tools, and other equipment, as well as training while also creating a foundation for sustainable production increases. Through matching grants under a demand driven approach, 500 smallholder farmer groups were expected to increase food production (10,000 tons of paddy rice) and their access to markets. Training was planned to Ministry of Agriculture Rural Development (MADR) in order to strengthen its technical capacity to assist smallholders and their organizations. MADR coordinated these activities with the help of a selected group of national NGOs that had been trained to support communities’ requests for and implementation of matching grants. 3. Project Management, Monitoring, and Evaluation (Appraisal Estimate : US$ 0.5 million, Actual: US$ 1.2 million). The component supported (i) the coordination, management, oversight and M&E functions of the Technical Coordination Unit set up within the Department of Rural Engineering in MADR. d. Comments on Project Cost, Financing, Borrower Contribution, and Dates: Project Costs Total project cost estimated at appraisal was US$ 5.0 million and 100 % of this amount was spent at project closing. Financing: The Project was financed by a US$ 5.0 million grant from the Food Price Crisis Response Trust Fund. Borrower Contribution : There was no Borrower contribution. Dates: The project was approved on September 22, 2008 and planned to be effective on October 13, 2008 but project effectiveness was delayed for 4 months and the actual effectiveness was on February 10, 2009. The original closing date of September 30, 2011 was extended for 23 months to August 31, 2013 through 5 level 2 restructurings. The reason for the extensions was to mitigate delays in the procurement and distribution of Component 2 goods; to allow completion of activities after political instability that caused 9 month suspension of funds from April to December 2012; and to facilitate the preparation of an Additional Financing, which had been envisioned but eventually decided to be prepared as a separate project. 3. Relevance of Objectives & Design: a. Relevance of Objectives: Substantial Guinea-Bissau is a low-income, food-deficit country, which ranked 176th of 186 countries in the 2012 UNDP Human Development Indicators. Chronic poverty is very high and increasing in the last decade (the poverty rate climbed from 65 % in 2002 to 69 % in 2010). Food insecurity is a chronic phenomenon in the country despite strong agricultural potential. The country’s agricultural population was able to meet only 60 % of the national demand for rice in 2007/08. At the time of appraisal due to global food price crisis, consumer prices for rice, the most important and preferred staple food, nearly doubled in international markets between 2007 and 2008 and therefore in the country as well. Since the country spent approximately 65 % of its income on food, it was extremely vulnerable to food price increases. The development objective was relevant to the Bank Group's Interim Strategy Note for Guinea-Bissau (FY 09-FY10), which identified increased rice production as a key outcome indicator under its first pillar: strengthening economic management and laying the foundations for improvements in the productive sectors. The project’s focus on the provision of seed, production tools, fertilizer, and small equipment was also well aligned with the Government's Emergency Plan for the Agricultural Campaign of 2008–2010. Yet, one shortcoming was that, although the project was prepared on an ‘emergency’ basis, the project development objective statement did not include the fact that the project support was ‘urgent’ or ‘immediate’. b. Relevance of Design: Substantial The statement of objectives was clear and measurable. Overall, the results framework presented a logical causal chain between project activities and expected attainment of the objective. The objectives of improving food security for the most vulnerable population of the country including children, and to increase smallholder rice production were to be supported through the timely provision of food rations as well as improved inputs and technical advisory services. Component 1 was designed to use World Food Program's expertise to urgently respond to food insecurity of two vulnerable populations: rural children participating in the school feeding program and the female students who received take-home rations; and also unemployed and landless rural laborers, who received daily rations through the food-for-work program. The design primarily attacked short term food security bottlenecks through the distribution of food rations and agricultural inputs and matching grants. In addition, via the land rehabilitation activity under the food for work program, medium-term measures were included. The capacity building activities under Component 2 to an extent tried to serve medium-term goals as well. Balanced by that project M&E framework was not designed to measure food security, but access to food, i.e. outcome indicators could include increase in number of food secure days for the households. Also, there could be additional indicators to measure the effects of school feeding program on weight of the students, for example. 4. Achievement of Objectives (Efficacy): The project has 2 objectives: (i) improve food security for the most vulnerable population including children; (ii) increase smallholder rice production in the selected areas. Both of the objectives were rated substantial. (i) Improve food security for the most vulnerable population including children , rated modest. Outputs: The project distributed meals to primary students one meal per day. A total of 14,102 student received meals daily against a target of 14,000 students (99 % achieved in 2008 school year and 101 % achieved in 2009 school year). Among these students 1,411 girls (approximately 10 %) received take-home rations based on their high attendance rates. The food-for-work program distributed 865.63 tons of food to approximately 29,598 beneficiaries for a total of 285,000 working days (130 % of the target) over 2 years (2009 and 2010). Women made up more than 50 % of the workforce for rehabilitating lowland fields, but only 10 % of the workforce for rehabilitating mangrove rice fields, due to traditions as well as the difficulty of working conditions in these fields. Outcomes: The achievements were modest. The project tried to respond rapidly to the food crisis by establishing a school canteen program and providing food in exchange for works to rehabilitate rice land through the use of World Food Program. However, as the effectiveness was delayed by 4 months, the Bank’s support for school feeding didn’t start until around June 2009 rather than at the beginning of the 2008-2009 school year as planned (although WFP was able to provide some support from their own resources). The key outcome indicators—the number of primary students in target areas receiving one meal daily and the number of work-days (rations) for rural producers in food-for-work activities were achieved. However, the PDO indicator measured ‘access for food’, but not food security. The ICR did not provide information regarding how much the vulnerable population’s food security improved through the food rations that were distributed. Also, it was not clear, how significant the target beneficiary of 14,000 students were among the total number of students and whether these students were poor. In addition, the ICR reported that (p. 10) the rations had to be cut considerably during the second year due to supply issues (rations decreased in quantity, as well as nutritional quality, when rice was substituted for maize and quantity of rice, oil and sugar declined). The ICR did not report on the calorie content of the food rations and how this reduction affected food security of those assisted. The ICR also did not comment on what happened to the savings when the rations were cut, as the money was spent completely. According to the ICR (p. 31), schools with feeding programs performed significantly better than those without in 2009 (national averages of average dropout rate is 14 % and an exam pass-rate is 65 %, against food program schools achieved pass-rates of 79 % (boys) and 86 % (girls) and against drop-out rates of 7 % (boys) and 4 % (girls). However, it was not clear if this data was specific to the project or included all the school feeding programs. (ii) Increase smallholder rice production in the selected areas , rated substantial. Outputs:  969,454 meters of dikes were built and 54,939 meters of canals were dug, rehabilitating 2,880 hectares of lowland fields and 2,741 hectares of mangrove rice fields (96 and 137 % of the targets, respectively) for a total 5,621 hectares of newly rehabilitated land.  131 producers and MADR staff received technical and project management training (202 % of the target).  60 tons of lowland rice seed and 27 tons of mangrove seed was provided to producers on newly rehabilitated land. However, there were serious start-up delays, which is contrary to the rationale for an emergency operation such as this.  788 micro-projects were approved and implemented (534 micro-projects focused on improving production, and 267 on post-harvest handling and processing) by the end of the third year of implementation.  Production micro-projects benefitted 13,934 producers through the procurement and distribution of NPK fertilizer (115,002 kilograms), urea (70,089 kilograms), rope for delimiting parcels (30,446 meters), lowland rice seed (62,355 kilograms), vegetable seed (370 kilograms), and power tillers (13) with fuel (20,700 liters), among other items.  Post-harvest micro-projects benefitted 7,072 producers through the procurement and distribution of rice hullers (52), materials (cement, nails, roofing supplies, etc.) for building warehouses and shelters for post-harvest processing machinery, sacks (8,400), and machines to puree tomatoes (5), among others. Outcomes: In terms of yield increases, the ICR reported contradictory results: on page 19, the ICR reported that yields of lowland rice doubled, increasing from the baseline of 1.1 tons per hectare to 2.1 tons per hectare on average by 2012. Yields for mangrove rice increased from 1.6 tons per hectare to 1.7 tons per hectare. However, according to Table A2.5 and Table A3.1, the average rice yield decreased, from 2.04 tons per hectare in 2009 to 1.85 tons per ha in 2011 (and no data was presented for 2012). Total rice production for 2009–11 on land rehabilitated by the project reached 9,378 tons by the end of the third season (125 % of the target).However, the baseline was zero production, which was not realistic for rehabilitated lands, as there must be some type of crop production before the project. It was noted that tons of rice produced were calculated based on data for average yields and cultivated area collected by MADR’s Department of Agricultural Statistics. The use of this secondary data complicated reporting of this outcome. A direct survey of project beneficiaries with a control group would have been preferable. Nevertheless, based on the emergency nature of the project as well as limited project size and budget, this point is not reflected in efficacy ratings. Another point is, it is not clear if production increases were attained by small-holders only. The ICR in various parts mentioned lack of objective targeting for Component 2 activities and PAD mentioned elite capture risk, but it was not clear if this risk was adequately mitigated during implementation. The ICR could have shed light on this. 5. Efficiency: Modest. Due to the emergency nature of the project and very short preparation time, the PAD did not include a quantitative economic analysis, but presented an explanation of direct and indirect benefits of the project. Accordingly, direct benefits would be attained from incremental yields per hectare as a result of dissemination knowledge, improved seeds and fertilizer application as well as rehabilitation of degraded mangrove and rice fields. Indirect benefits would come from increased employment of rural labor and agricultural technicians to assist with project implementation. Ex-ante calculations included cost benefit analysis of project components, but there were some weaknesses. Accordingly, food support to vulnerable populations had three types of economic impacts: a food security impact, an employment impact, and a rice production impact. The benefit-cost ratios for investment in rice production on the land rehabilitated under the food-for-work program were high with 3.5 and 3.9 based on price of rice of US$ 450 per ton and US$ 500 per ton respectively (assuming that cost of production was 40 % of value of rice). However, the calculations did not seem to be including the investment cost and/or discounting the benefits and costs over time, including these aspects would reduce the benefit/cost ratios. The ICR also calculated IRR for the rehabilitation of dikes for lowland rice fields and mangrove rice fields (21 % and 29 % respectively), however, the assumptions as well as source of data were not presented. It was not clear if the overhead costs (cost of component 3) was included. Also, the ICR did not present project cost per beneficiary figures for the different project support. This review calculated the unit cost for Component 1, as US$35.5 per beneficiary per year, which was about the same level with the "industry standard" of US$ 34 per student per year, given 200 feeding days and 700 calories per daily ration. However, the calorie level and number of days could be lower for this project, and this point is not clear from the ICR. Moreover, it is uncertain if the US$ 178/ha rehabilitation cost for rice and mangrove fields was reasonable compared to similar projects (no comparators were provided). Support for increasing food production through distribution of rice seed, fertilizer, and farmer training in good agricultural practices via micro projects had 4 types of impacts in project areas—improved knowledge and capacity, increased farm income, increased employment, and increased rice production. Benefit/cost ratios for these investments were also substantial, varying from 1.64 to 2.74 (with 50 % and 75 % increase in yield respectively). The ICR reported that (p. 20) cost savings was achieved on training programs, i.e. more beneficiaries and MADR staff benefitted from the training than the originally targeted (131 vs. target of 65) by revising the training plan to bring international experts to the country allowing them to train many more beneficiaries than could have been trained outside the country as originally planned. Training planned for Europe was switched to regional locations, further stretching the original budget. However, the ICR did not present cost per beneficiary figures from similar projects as comparators. There were some administrative and operational inefficiencies. The project closing date was extended for a total of 23 months from the original closing date. The extension was partly due to political instability that caused a 9 month of suspension of funds, but there were also delays in procurement and delays in distribution of Component 2 goods. The emergency nature of the project and therefore rapid project preparation resulted in insufficient attention for procurement, staffing and M&E activities, which caused delays in implementation. In addition, the overhead costs (Component 3) ended up being 240 % of appraisal estimate, almost a quarter of total disbursements, which seems high. It was not clear why this substantial cost over-run was encountered and how it affected outcomes and outputs. The ICR does not mention how the cost overruns for Component 3 affected the other components from which the resources had to be deducted. a. If available, enter the Economic Rate of Return (ERR)/Financial Rate of Return (FRR) at appraisal and the re-estimated value at evaluation : Rate Available? Point Value Coverage/Scope* Appraisal No ICR estimate No * Refers to percent of total project cost for which ERR/FRR was calculated. 6. Outcome: In light of ‘substantial’ relevance of objectives and relevance of design but modest achievement of the objective of improving food security for the most vulnerable population due to lack of data on improved food security of beneficiaries, 'substantial' achievement for the objective increasing smallholder rice production, and ‘modest’ efficiency due to lack of sufficient evidence on significant economic rate returns or benefit/cost ratios, as well as high administrative and operational inefficiencies, the overall outcome is 'moderately satisfactory'. a. Outcome Rating: Moderately Satisfactory 7. Rationale for Risk to Development Outcome Rating: Political Risk : High. Political turmoil in the last 4 decades with frequent military coups has left the country with low prospects for macroeconomic growth, and lack of stable and accountable institutions; therefore provision of essential public services and infrastructure entirely depends on resources from donors. Nevertheless government ownership and commitment to address food security issues is high, which should have a positive impact for sustainability. Institutional and Financial Risk : High. Institutional and financial risk are interlinked as the institutional development activities are under risk due to lack of financial resources. Training provided under Component 2 increased the capacity of MADR to carry out the extension, advisory, and project implementation roles. However, MADR departments are unable to function without donor support. Personnel in the field lack operational budgets, equipment, transportation; and donors have to finance nearly all operations. For Component I activities, there is no expectation that school feeding and food-for-work programs will be taken on by government institutions rather than donors in the near future. Also, the lack of improved inputs (seed, tools, and small production and post-harvest equipment) and lack of funds to maintain dikes and canals, jeopardizes the improved production rates and therefore threatens any future growth in production. Balanced by that at project closing a new GFRP-financed operation was being prepared to follow this project and was expected to reinforce its sustainability. Social Risk: Moderate. The ownership developed over project goods in the process of implementing micro-projects is likely to strengthen the already existing tendencies to cooperate. a. Risk to Development Outcome Rating : High 8. Assessment of Bank Performance: a. Quality at entry: Emergency procedures under OP 8.0 were used and preparation of the project from initial concept review to board approval took only 2 months. However, effectiveness was then delayed by four months longer than what was planned, thereby undermining the sense of urgency with which the operation was initially prepared. Background data used in designing the project was adequate and included World Food Program surveys and analyses of food insecurity and malnutrition, documentation for the Coastal and Biodiversity Management Project, and the analysis that the government had undertaken for the Emergency Action Plan for Agricultural Campaign 2008–2010. Project's technical design was in general sound. Project’s components focused on a set of urgently needed activities to mitigate particularly the short-term impacts of the food crisis. Direct contracting with the experienced World Food Program for the school feeding and food-for-work programs not only addressed the government’s implementation capacity constraints but also created quick implementation set up. The matching grant model was adapted from successful community - based development projects from Angola and Nigeria as well as 3 on-going projects in Guinea-Bissau (funded by the World Bank, Global Environment Facility (GEF), the European Commission and the African Development Bank). However, there were also the following moderate shortcomings due to accelerated time for project preparation: (i) The criteria for selecting areas and beneficiaries for the school feeding and food-for-work programs was not defined rigorously; i.e. beneficiaries of the land rehabilitation activities were identified based on regional level analysis that did not permit to objectively target specific plots owned by the most vulnerable people; (ii) the proposed implementation timeline for Component 2 proved optimistic; (iii) although project management arrangements were in general sound, with the Technical Coordination Unit set up in MADR, this was balanced by the lack of coordination and communication arrangements across the Technical Coordination Unit, World Food Program, and Project Steering Committee; (iv) Insufficient time was available to develop critical implementation tools, such as those related to M&E, recruitment, and procurement; (v) The PDO did not have a time element, although it was an emergency operation; (v) The PDO indicator measuring food security was output oriented rather than outcome (i.e. instead of measuring improvement in food security, it measured access to food rations). Overall, the operation took five years to implement and required five extensions, which is long for an emergency operation, and a reflection of inadequate planning. The risks identified and mitigation measures were adequate. They included: exogenous economic shocks, the political fallout from delayed elections, weak local capacity for implementation, governance risks, technical/design risks resulting from inadequate oversight and limited availability of inputs, limited implementation capacity, elite capture of matching grants, potentially inadequate fiduciary management. The measures proposed to mitigate those risks were generally appropriate and adequate. Quality-at-Entry Rating: Moderately Satisfactory b. Quality of supervision: The project remained under the leadership of the same task team leadership from preparation to closing, providing continuity on project supervision. Field supervision missions were carried out regularly during the life of the project and aide-mémoires examined achievements and administrative and technical implementation issues, and they were always reviewed with stakeholders at the end of each mission. ISRs adequately described project progress, and identified issues requiring management’s attention. Balanced by that, early in the life of the project, delays in filing ISRs prevented management’s timely follow-up of emerging concerns with Implementation. Furthermore, the team failed to conduct an MTR despite the fact that the operation lasted for 5 years, and the lack of a mid-term review delayed identification of emerging M&E challenges related to Component 2 There was no baseline surveys and the issues with the M&E specialist was not resolved. It was not clear if the team adequately monitored the issue that the overhead expenditures more than doubled during implementation. The operation also lasted for five years and required five extensions, which is long for an emergency operation. Quality of Supervision Rating : Moderately Satisfactory Overall Bank Performance Rating : Moderately Satisfactory 9. Assessment of Borrower Performance: a. Government Performance: The government’s performance rating is based on the performance of: MADR in its role as policy maker and facilitator of development programs and the Technical Coordination Unit (TCU); as well as the Project Steering Committee in its oversight role. MADR appointed the Director of MADR’s Department of Rural Engineering to work with the Bank team in designing the project. By bringing extensive knowledge of rural development in Guinea, the director contributed significantly to the technical design and the structure of implementation arrangements. After effectiveness, MADR encouraged qualified government staff to accept positions with the project. The Department of Rural Engineering was given responsibility for coordinating all project activities and implementing Components 2 and 3. The TCU was experienced on rural operations and knowledgeable on rural community needs. It needed strengthening on Bank procedures and a stronger emphasis on M&E system design and utilization. This would have reduced delays at start-up, and would have helped capture and document the project’s results more extensively. Nevertheless, the staff quickly mastered the application of the Bank’s fiduciary management mechanisms, procurement, reporting, filing financial reports, and organizing supervision missions. Balanced by that, there were significant delays in hiring and training NGOs that delayed Component 2 activities and led to extension of the project closing date. Also, there was substantial cost over-run for Component 3, unexplained “supply issues” in purchasing food, and quite a few ineligible expenditures that the Government did not repay promptly. In addition, M&E weaknesses were not addressed and an M&E specialist was not hired. The Project Steering Committee, which included representatives from 5 ministries, was created in a timely fashion but met only once over the course of the project. The critical functions of the Project Steering Committee, such as technical reviews and budget approvals, were performed by the TCU in collaboration with the Minister of Agriculture. Government Performance Rating Moderately Satisfactory b. Implementing Agency Performance: The Implementing Agencies performance rating is based on the performance of: (i) World Food Program (WFP) responsible for implementation of Component 1, and (ii) the national NGOs responsible for activities supported under Component 2. Overall both WFP and NGOs performed satisfactorily. Component 1 was 100 % disbursed by July 2009. WFP resolved implementation issues in a timely way such as supply issues and logistical delays. It also coordinated and collaborated with other donors and stakeholders. Minor shortcomings in fiduciary management regarding financial reporting delays occurred. WFP's reporting system was different than the Bank's reporting requirements. (Although, WFP’s project reporting system, normally establishes the Standard Project Report as the reporting tool for all donors to WFP, this format was not agreed prior to signing the Grant Agreement). Under Component 1, national NGOs performed M&E functions adequately and mitigated the risks of corruption in ration distribution. Under Component 2, national NGOs performed satisfactorily. NGOs’ adequate facilitation of participatory preparation of proposals for micro-projects are evidenced by the relevance, and comprehensive details of the micro-projects prepared under Component 2. They provided continuous guidance in the field to the beneficiary communities that contributed to the sustainability of the results. ICR mission on the field observed that tools and equipment distributed by the project was very well maintained and still being used by the farmers. Implementing Agency Performance Rating : Satisfactory Overall Borrower Performance Rating : Moderately Satisfactory 10. M&E Design, Implementation, & Utilization: a. M&E Design: M&E Framework did not include an adequate outcome indicator to measure improvement in food security of project beneficiaries. Rather, the designed outcome indicator was output oriented that measured access to food rations.. The M&E design assigned responsibility for collecting data on indicators to World Food Program and NGOs. The functions of compilation, analysis, and presentation of results remained with MADR. The project would have benefited from a better approach for calculating the total number of direct project beneficiaries, as the same beneficiaries could have received support for land rehabilitation and also received support through micro-projects to improve production on that land. Due to the emergency nature of the project, no baseline field studies were planned; although the project had substantial time during implementation to do baseline surveys. Also, all indicators relied on secondary data. Also, MIS system had be developed during implementation. b. M&E Implementation: M&E data was collected during World Bank supervision missions and for the quarterly and final reports prepared by Technical Coordination Unit. The overall quality of project M&E showed variations based on the Component. Monitoring of Component 1 was better than that of Component 2, as few M&E specialists were available in the government or private sector to work for the project. When M&E activities had to be managed by the project coordinator, as no replacement was found for the M&E specialist, whose contract was terminated, the data collection, compilation, and analysis improved. This experience showed that the government has to have adequate capacity for project M&E. The secondary data sources could not provide adequate production or post-harvest data for evaluating the supply response, which complicated the impact assessment. c. M&E Utilization: Although the project did not have an adequate M&E system, the ICR notes that (p 15) it could collect a significant quantity of data to inform the project indicators and design of follow-up activities. M&E Quality Rating: Modest 11. Other Issues a. Safeguards: The project activities triggered 3 safeguard policies: OP 4.01 (Environmental Assessment), OP 4.04 (Natural Habitats), and OP 4.09 (Pest Management). As the project was an emergency operation using the expedited procedures under OP 8.00, the preparation and public disclosure of safeguards instruments were not required until 120 days after effectiveness. An Environmental and Social Safeguards Framework (ESSF) was prepared and disclosed but with a 4 months delay. Implementation of safeguard policies were adequate. The project completed a Pest Management Plan before distributing pesticides under Component 2. NGOs supporting project beneficiaries in each region were trained in safe pest management techniques, and an environmental safeguards checklist accompanied each micro-project application. Public disclosure events were conducted via forums, radio programs, and brochures in areas where the project operated. b. Fiduciary Compliance: The three audits for the project were submitted to the Bank with considerable delays, but all confirmed that the fiduciary management of the project was adequate. One audit flagged a number of ineligible expenditures, which the government reimbursed in full, with some delay. In terms of procurement, ex ante and ex post procurement reviews turned up no major shortcomings. However, there was persistent weaknesses in file management and archives. c. Unintended Impacts (positive or negative): d. Other: Gender: The project specifically aimed to support women and girls. Under Component I, take-home rations were provided to girls with superior attendance rates, thus reducing the burden of their absence on their families and therefore trying to increase girls' school attendance. Female beneficiaries were 47 % of total beneficiaries under this program in 2008/09 school year. Also, the ICR reported the following (p. 31): "..National statistics estimate an average dropout rate of 14 % and an exam pass-rate of 65 % . In contrast, World Food Program schools achieved pass-rates of 79 % (boys) and 86 % (girls) against drop-out rates of 7 % (boys) and 4 % (girls)". The micro-projects implemented under Component 2 also targeted women through support to market gardens and post-harvest processing; the majority of the 267 post-harvest micro-projects supported women’s activities. 12. Ratings: ICR IEG Review Reason for Disagreement/Comments Outcome: Satisfactory Moderately In light of ‘substantial’ relevance of Satisfactory objectives and relevance of design but modest achievement of the objective of improving food security for the most vulnerable population due to lack of data on improved food security of beneficiaries and reduced rations and delayed startup, 'substantial' achievement for the objective increasing smallholder rice production, and ‘modest’ efficiency due to lack of sufficient evidence on significant economic rate returns or benefit/cost ratios, as well as high administrative and operational inefficiencies , the overall outcome is 'moderately satisfactory'. Risk to Development High High Outcome: Bank Performance: Satisfactory Moderately Both QAE and Supervision are rated Satisfactory ‘moderately satisfactory’ due to weak M&E, insufficient beneficiary targeting for Component 2, and not being able to conduct an MTR to inform implementation. Borrower Performance : Satisfactory Moderately The Government was rated ‘moderately Satisfactory satisfactory’ due to cost over- run under Component 3, weak M&E, and delays in hiring and training of NGOs that significantly delayed Component 2 implementation that led to project extension. Quality of ICR: Satisfactory NOTES: - When insufficient information is provided by the Bank for IEG to arrive at a clear rating, IEG will downgrade the relevant ratings as warranted beginning July 1, 2006. - The "Reason for Disagreement/Comments" column could cross-reference other sections of the ICR Review, as appropriate. 13. Lessons: The ICR offers a number of lessons of which the following are the most important (with some reformulation of language):  Capacity constraints in fragile countries must be properly mitigated during design and implementation . The project experience showed that Bank teams shouldn't assume that the recipient is familiar with Bank methods and procedures, including accounting, procurement and M&E. A thorough analysis of local capacity for all aspects of project management must be a priority during preparation. Training to mitigate weaknesses at start-up, as needed, should be planned.  A minimum set of project “readiness conditions ,” coupled with intense implementation support , is necessary for emergency operations to succeed . The flexibility allowed in preparing emergency operations facilitates rapid preparation, but it may compromise the quality of implementation and jeopardize the sustainability of impacts if minimal implementation arrangements and conditions are not in place. These arrangements and conditions include: (i) a clear, appropriate action plan for start-up; (ii) intense Bank supervision and “handholding” following effectiveness, including on-the-ground procurement and M&E support; and (iii) a rapid baseline survey or equivalent information gathering to ensure appropriate targeting and M&E.  Community participation and the involvement of national NGOs enhances the sustainability of impacts . According to the ICR, one result of strong community contributions to the design and implementation of micro-projects was that equipment distributed by the project has been relatively well maintained at project closing. Involvement of NGOs in providing technical assistance and on the job training to producer groups increased the likelihood of continued capacity building and guarantees that new development management skills stay in the communities that need them most.  The scale of an operation such as this should be consistent with the objectives . The resources available seem to have been too limited to have a major impact on food security. 14. Assessment Recommended? Yes No 15. Comments on Quality of ICR: The ICR provided a description of implementation progress and challenges. However, the following points needed attention: (i) The ICR could have reported more on M&E system data resources and how they affected reporting on results; (ii)The ICR could have provided more explanation on what the initial procurement and recruitment issues, how they affected implementation and how they were resolved; (iii) The ICR was silent about the risk of elite capture for micro-grants and if and how Component 2 activities really targeted small-holders. (iv) The cost over-run under Component 3 was not explained; (v) Very limited information was given on the reasons of the 5 closing date extensions; (vi) There was no explanation on the effect of cutting of food rations; (vii) Text and data in Annexes were not consistent (such as yield data); (viii) Economic analysis assumptions were not provided, unit costs and comparisons were missing. a.Quality of ICR Rating : Satisfactory