Document of The World Bank

Transcription

1 Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized ESSD Sector Management Unit Brazil Country Management Unit Latin America and Caribbean Region Document of The World Bank IMPLEMENTATION COMPLETION REPORT (CPL-42520) ON A LOAN IN THE AMOUNT OF US$80.0 MILLION TO THE STATE OF MARANHÃO FOR A RURAL POVERTY ALLEVIATION PROJECT December 28, 2004 Report No: 30275

2 CURRENCY EQUIVALENTS (Exchange Rate Effective June 30, 2004) Currency Unit = Real R$1.00 = US$ 0.32 US$ 1.00 = R$3.10 FISCAL YEAR January 1 to December 31 ABBREVIATIONS AND ACRONYMS APCR CAS CDD FUMAC FUMAC-P GEAGRO GEPLAN HDI IICA IRR MIS NEPE NGO NRDP O&M PAC POA R-NRDP SEBRAE SOE PCU Program of Support to Small Rural Communities Country Assistance Strategy Community-driven Development Municipal Community Schemes Pilot Municipal Community Schemes State Management Unit for Agriculture, Livestock and Rural Development State Administration for Planning and Economic Development (Borrower) United Nations Human Development Index Inter-American Institute for Cooperation in Agriculture Internal Rate of Return Management Information System State Center for Special Programs (Project Coordination Unit) Non-governmental Organization Northeast Rural Development Program/Project Operation and Maintenance State Community Schemes Annual Operating Plan Reformulated Northeast Rural Development Program/Project Brazilian Service for Small and Medium Business Statements of Expenditure Project Coordination Unit Vice President: Country Manager/Director Sector Manager/ Director Task Team Leader David de Ferranti Vinod Thomas John Redwood Luis O. Coirolo

3 BRAZIL Rural Poverty Alleviation - Maranhao CONTENTS Page No. 1. Project Data 1 2. Principal Performance Ratings 1 3. Assessment of Development Objective and Design, and of Quality at Entry 2 4. Achievement of Objective and Outputs 4 5. Major Factors Affecting Implementation and Outcome Sustainability Bank and Borrower Performance Lessons Learned Partner Comments Additional Information 26 Annex 1. Key Performance Indicators/Log Frame Matrix 27 Annex 2. Project Costs and Financing 28 Annex 3. Economic Costs and Benefits 30 Annex 4. Bank Inputs 37 Annex 5. Ratings for Achievement of Objectives/Outputs of Components 38 Annex 6. Ratings of Bank and Borrower Performance 39 Annex 7. List of Supporting Documents 40

4 Project ID: P Team Leader: Luis O. Coirolo Project Name: Rural Poverty Alleviation - Maranhao TL Unit: LCSER ICR Type: Core ICR Report Date: December 29, Project Data Name: Rural Poverty Alleviation - Maranhao L/C/TF Number: CPL Country/Department: BRAZIL Region: Latin America and the Caribbean Region Sector/subsector: Other social services (40%); Roads and highways (20%); General agriculture, fishing and forestry sector (20%); Sub-national government administration (10%); Water supply (10%) Theme: Rural services and infrastructure (P); Participation and civic engagement (P); Rural non-farm income generation (P); Decentralization (S); Other social development (S) KEY DATES Original Revised/Actual PCD: 04/15/1997 Effective: 03/15/ /12/1998 Appraisal: 06/02/1997 MTR: 06/04/ /14/2002 Approval: 11/20/1997 Closing: 06/30/ /30/2004 Borrower/Implementing Agency: Other Partners: State of Maranhão/State Administration for Planning and Economic Development (GEPLAN)/State Center for Special Programs (NEPE) Community Associations STAFF Current At Appraisal Vice President: David de Ferranti Shahid Javed Burki Country Manager/Director: Vinod Thomas Gobind T. Nankani Sector Manager/Director: John Redwood Maritta Koch-Weser Team Leader at ICR: Luis O. Coirolo Luis O. Coirolo ICR Primary Author: Anna Roumani with Task Manager; R. Caminha and Brazil Northeast Team 2. Principal Performance Ratings (HS=Highly Satisfactory, S=Satisfactory, U=Unsatisfactory, HL=Highly Likely, L=Likely, UN=Unlikely, HUN=Highly Unlikely, HU=Highly Unsatisfactory, H=High, SU=Substantial, M=Modest, N=Negligible) Outcome: Sustainability: Institutional Development Impact: Bank Performance: Borrower Performance: S L SU S S QAG (if available) Quality at Entry: Project at Risk at Any Time: No ICR S

5 3. Assessment of Development Objective and Design, and of Quality at Entry 3.1 Original Objective: The Loan Agreement states Project objectives as assisting the Borrower to alleviate rural poverty in the Municipalities by: (a) providing basic social and economic infrastructure, and employment and income generating opportunities for the rural poor; (b) supporting rural community groups in planning and implementing their own investment subprojects; (c) intensifying current efforts to create social capital in poor rural areas; and (d) leveraging revenue mobilization at the community and municipal levels. 3.2 Revised Objective: Project objectives were not revised. 3.3 Original Components: The Project s three components were as follows: (a) Community Subprojects (US$95.0 million, 90% of base cost) to support small-scale investments selected, executed, operated and maintained by the beneficiaries themselves; (b) Institutional Development (US$7.4 million, 7% of total base cost), which would provide all implementing agencies and communities with technical assistance and training to increase their capacity and improve implementation of the Project; and (c) Project Administration, Supervision, Monitoring and Evaluation (US$3.1 million, 3% of total base cost), financing project coordination and activities to provide feedback on project performance and impact. These components represented an appropriate set of actions to achieve project objectives and to benefit an estimated 198,000 families - close to 1.0 million people - having been based on an evaluated, scaled-up pilot experience in 10 Northeast states under the reformulated Northeast Rural Development Program (R-NRDP, ). 3.4 Revised Components: Project components were not revised. 3.5 Quality at Entry: Rated Satisfactory (project appraisal pre-dated the QAG). The Project emerged from the experiences and positive outcome of an innovative program which tested a different approach to rural poverty reduction. Rooted conceptually in both Bank and non-bank rural poverty programs in several regions, the program also assumed a distinct home-grown persona reflecting the realities of Northeast Brazil and the new Constitution of The centralized, administratively-driven and largely ineffective integrated rural development model of the previous 20 years was abandoned and substituted by decentralized, participatory, demand-based mechanisms conforming to the following guiding principles: (a) place money for implementing approved subprojects directly into the hands of the beneficiary communities; (b) decentralize decision-making and involve local authorities as participants/partners; (c) maintain transparent decision-making at all levels; (d) stress community participation in planning, financing, executing, operating and maintaining the investments financed by the Project; and (e) use simple, explicit poverty targeting mechanisms which can be easily monitored. The community-driven development (CDD) approach also dispensed with the intermediary agencies and bureaucratic complexities of earlier programs, thus ensuring that at least 90% of project funds were used for direct subproject investments, the remainder financing - 2 -

6 mainly technical assistance. The Project was embedded in the Bank s Program of Targeted Interventions, employing a three-part methodology for reaching poor rural populations: geographically by poverty level and other characteristics of the municipalities; geographically within municipalities to target rural settlements and communities; and community-based selection by the project Municipal Councils of poor beneficiaries and particularly vulnerable groups. Important complementary objectives included: laying the groundwork for, and consolidating, participatory institutions and processes at the municipal and community levels; ensuring that project funds reached the poorest municipalities and communities; and, expanding the successful strategy of decentralization and participation tested under the previous Bank-supported program in the Northeast. * * Reformulated Northeast Rural Development Program (R-NRDP, ) Objectives and activities were internally consistent, addressing State poverty conditions, the benchmarks of the 1996 and 1997 CAS for Brazil, and the State s poverty reduction strategy at that time. Maranhão is the second poorest state in Brazil and the most rural, with exceptionally low Human Development Index (HDI) measures and gross deficiencies in education, socio-economic infrastructure and services, and institutional capacity. At preparation (1997), about half the population was indigent despite the state s freedom from the drought/semi-arid conditions typical of the Northeast region - with 75% of that cohort living in rural areas; and, over 90% of all municipalities harbored precarious living conditions affecting childhood survival.* * HDI is a summary measure of three dimensions of the human development concept: living a long and healthy life, being educated and having a decent standard of living. It combines measures of life expectancy, school enrollment, literacy and income and is closely linked to the MDG. Design and objectives were also consistent with the 1997 Bank Country Assistance Strategy (CAS) for Brazil which stressed: the centrality of poverty reduction; regional development of the Northeast as a priority; development of the poor rather than relief of poverty; improved sustainability through cost recovery; and stronger financial commitment from municipal and state governments. The State s strategy for poverty alleviation in the period targeted reductions in the unequal distribution of wealth and included technical and financial incentives for local job generation by small-scale producers; investing in improved social indicators; and, diversifying and decentralizing economic activity to the community level and establishing more efficient distribution of services. Three mechanisms with increasing degrees of decentralization are used to screen, approve and implement community requests for financing PAC, FUMAC and FUMAC-P. Under PAC (State Community Schemes), rural communities submit investment proposals directly to the State Technical Unit (STU) which screens and approves them, releasing funds to the beneficiary association. With FUMAC (Municipal Community Schemes), investment decision-making is delegated by the State to project Municipal Councils, representative bodies with 80% community members and local civil society, and 20% municipal authorities. Councils establish priorities and approve community proposals. The STU reviews them for consistency with the Operational Manual, then transfers funding for approved proposals directly to the community association bank account. Councils monitor and supervise subproject execution. FUMAC-P (Pilot Municipal Community Schemes) is a more decentralized variant of FUMAC piloted under the Rural Poverty Alleviation Program.* With an annual budget from the STU based on criteria such as rural population, poverty and previous year s performance, Councils submit an Annual Operating Plan - 3 -

7 (POA) to the STU. Approval releases budget to the Council which manages its distribution to beneficiary associations and assists with subproject execution. PAC is now only used in municipalities with little chance for whatever reason - of forming a FUMAC. * Rural Poverty Alleviation Program comprising 8 state-level CDD projects in the Northeast. Design heeded lessons learned under the R-NRDP, scaled-up successfully-piloted features and introduced further innovations and improvements: (i) mainstreamed the FUMAC Municipal Council mechanism (a pilot under R-NRDP) and introduced the even more decentralized FUMAC-P instrument, giving selected, high-performing FUMAC Councils greater control over budget resources; (ii) systematized beneficiary targeting and transparency through performance-based incentives and penalties; (iii) upgraded the Management Information System (MIS) - data is sent regularly to the Bank via Internet to support proper targeting, better project management, and monitoring and evaluation; and (iv) introduced an Operational Manual with technical, financial and administrative parameters, along with funding specifically for technical assistance to support mobilization and organization, build capacity and foster the participation of NGOs and other partners. Other improvements included even more intensive supervision, standardized subproject design and cost parameters, publicity and information campaigns, and cost-sharing requirements for the State Government and beneficiaries. Project design also reflected early lessons emerging from peer operations in the same program (approved and under supervision) introducing: (i) increased Bank share of expenditures financed for community investment subprojects, from 59% to 75% ; (ii) technical assistance (TA) funding under the Institutional Development component for the Secretariat of Planning to conduct administrative and other state reforms; and (iii) funds for technical assistance support for community investment subproject preparation and/or implementation, valued from 2-8% of subproject cost, depending on complexity. Experience in Brazil and elsewhere showed that the main risks for rural poverty projects were irregular funding, technical weaknesses and over-design, inadequate supervision, mis-targeting of resources and over-centralized decision-making. Measures were introduced to avoid or reduce those risks, which proved effective in practice and are standard features of the CDD program region-wide. 4. Achievement of Objective and Outputs 4.1 Outcome/achievement of objective: Rated Satisfactory Objective Providing basic social and economic infrastructure, and employment and income generating opportunities for the rural poor. The Project met this objective, financing 4,160 investments in basic rural infrastructure, productive and social facilities, improving the living standards and well-being of some 276,000 families plus another 10,000 indigenous and Afro-descendent families about 1.4 million rural people as well as the local economy of rural Maranhão. Aggregated data shows the Project financed the recuperation of 8,130 km of access roads; built or renovated about 2,900 m of bridges; installed some 6,570 km of electricity distribution line; constructed or renovated 9,253 houses; installed 42 km of water supply distribution systems; established 1,300 ha of communal - 4 -

8 farm plots; and installed 156 small-scale agro-processing units. The more than 1,300 ha of orchards and manioc, corn, rice and bean fields, as well as the introduction of small-scale fowl-raising (1,950 sq. meters) goat-raising (840 breeding animals), cattle (535 breeding animals) and 154 community fish farming facilities are not only increasing family incomes and absorbing labor but also improving family food intake and nutrition. These investments, and the many others financed, are responding to beneficiaries needs and are having positive impact on community health, education, housing, employment and income, productivity, transportation, communication, migration and other elements of human development in rural areas. Infrastructure investments predominated (76% of total) with rural electrification, community access roads and simplified water supply systems accounting for 93% of that category. As in other participating Northeast states, communities desire the basics for their immediate impact on daily life and for their potential to jump-start employment and income-generating activity, followed by productive and social investments (16% and 8% respectively). Installation of electrification for some 76,000 families is powering homes, schools, streets, village shops, small-scale agro-processing units, and pumping water. Among its many benefits are night school and adult literacy classes, conservation of food and medicine, a dramatic increase in access to domestic appliances, improved productivity from small-scale irrigation and processing systems, and access to information. Access roads benefiting about 98,000 families are especially vital in rural Maranhão where distance and isolation combined with exceptionally poor, frequently impassable roads/tracks limit the economic, social and education opportunities of poor communities. Particularly striking, Project-supported access roads represented 50% of all such investments in the State in the period. Water supply investments benefiting about 46,000 families are facilitating domestic labor, promoting better family hygiene, sharply reducing the high incidence of water-borne disease (amoebic dysentery, cholera, malaria). Communities surveyed reported only isolated cases since their water supply subproject was installed, the healthcare burden on municipalities has been reduced, as well as the daily burden on women and girls of having to collect water from distant and risky locations. Evidence strongly suggests that water investments, when combined with complementary investments in power for example, can create productive jobs and increase incomes. Community tractors and equipment were about 25% of approvals in the productive category, and significantly expand area cultivated by small farmers, ensure timely preparation of land to exploit rainy periods, improve quality of soil preparation, and facilitate cultivation and post-harvest activities, contributing to better incomes. Research repeatedly demonstrates the positive direct and indirect impacts, including interesting multiplier effects, of communal tractor subprojects under the Northeast CDD program. Within the social cohort of investments, housing improvement investments substituting cement/tile dwellings for deteriorated wattle and daub construction with earthen floor, were about - 5 -

9 93% of the total - despite the State Special Projects Unit (NEPE, the Project Technical Unit) having approved barely 20% of actual demand - and benefited about 9,250 families. Families reported much cleaner air and fewer respiratory ailments, combined with more sanitary, safer and normal living conditions. NEPE notes in its end-project completion report* that housing subprojects have a strong structural significance to poor communities and are not seen in purely social terms. * Projeto de Combate a Pobreza Rural do Estado do Maranhão PCPR-MA (Comunidade Viva), Relatório Final, November 2004 (Borrower Completion Report). Objective Supporting rural community groups in planning and implementing their own investment subprojects. At Closing, about 3,000 community associations had participated in the Project (with another thousand organized and waiting to participate), and a large proportion had received technical assistance and/or training. The training methodology was client-driven and participatory. NEPE and its Regional Development Units (RDU) trained the Municipal Councils in the selection and priority ranking of community needs and their translation into sound investment proposals. Development of standardized subproject designs has helped associations with technical aspects and construction of the more common types of investments (housing improvement, simple water supply systems, manioc mills). This has many benefits for the communities: simplified the preparation and evaluation of subprojects by poor communities and Councils, fostered more competition and better quality of execution of works, reduced subproject cost and communities dependence on private firms, and facilitated procurement. Of equal importance, community-based research during project implementation showed that the mobilization, organization, training and technical assistance associated with the subproject cycle fosters the steady, substantial and measurable accumulation of social capital in beneficiary communities, boosting inter alia, their will and readiness to continue pressing for their rights as citizens. Objective Intensifying current efforts to create social capital in poor rural areas. Fostering social capital in rural municipalities and communities has been a watershed achievement of the Northeast CDD projects, including Maranhão. As participatory, open forums for discussing and establishing local development priorities, the Municipal Councils have reduced political interference and patronage-based management; created partnerships between the Councils, communities and municipal governments; given communities a voice in public resource allocation; fostered citizenship through greater awareness of social responsibilities; and strengthened local capacity to select, establish priorities for and implement, investment decisions. Social capital has expanded through the evolution of successive CDD-based rural poverty programs and through the increasing decentralization of the mechanisms themselves. Gains in social capital within communities and municipalities are evidenced by the significant number of Municipal Councils which have moved beyond subproject decision-making to become participatory venues for broader municipal planning, i.e., the discussion goes beyond the Project, and the allocation of non-project resources; and community associations which are leveraging social capital acquired under the Project to obtain other, previously inaccessible, investment resources. While this phenomenon holds great promise for the future sustainability of Councils as an integral part of the rural socio-economic and political establishment, it should not be over-stated. It is an evolutionary - 6 -

10 process with 212 Councils in Maranhão at various stages/levels of capacity. Most Northeast rural communities have significant, established levels of social capital and thus the Project s participatory mechanisms are not starting from scratch. One of the core challenges, as defined by a decade of Bank research into the impact of the CDD projects on social capital formation in the Northeast, is to build community trust in the state. Thus, each subproject deliberated, implemented and completed by the community is a positive component in the process of changing the traditional relationship. Community access to accurate information (through publicity campaigns, and as a core element of the CDD methodology) at all stages of their subproject cycle is crucial to this process and to leveraging existing social capital for further developmental gains. Objective Leveraging revenue mobilization at the community and municipal levels. The Maranhão Project and its peer projects are among the few programs in Brazil which have successfully leveraged revenue mobilization at the local level. Beneficiary cost-sharing (mostly family labor and materials) under the Project s matching grant feature totaled US$9.5 million by end-project, or about 10% of the costs of all subprojects financed through Closing, and about 9% of total Project cost. Financial contributions by municipalities not required under the Project - were negligible. It is common throughout the participating states of the Northeast to find mayors contributing to the operations of Municipal Councils with logistical support in lieu of financial (e.g., equipment, premises, transportation for meetings), but this was not as common in Maranhão where mayoral support has been slower to evolve as a generalized phenomenon. Under the new Project (Maranhão Integrated Program: Rural Poverty Reduction Project, Loan 4735-BR), Councils will match community investment proposals with available funding including non-project resources from other public programs, while abiding by established CDD mechanisms and rules, in this way leveraging both additional funding for Project-generated activities and scaled-up use of the methodology itself outside the Project. 4.2 Outputs by components: Community Subprojects (US$95.0 million or 90% of total base cost) Rated Satisfactory This component financed small-scale investments selected, implemented, operated and maintained by the beneficiaries. Investments of up to US$50,000 each were financed through matching grants (public grants matched with community contribution), with legally constituted community associations identifying their priorities and committing to their operation and maintenance. It was estimated the Project would finance 3,392 community-managed subprojects over five years covering basic infrastructure, productive and social investments benefiting 198,000 families organized in 1,696 associations. An estimated 116 FUMAC and 20 FUMAC-P Councils were to be established. These were largely indicative targets due to the demand-based, flexible nature of the Project, and were estimated on the basis of demand/behavior patterns under the previous project. In six years of implementation, the Maranhão RPAP exceeded most of its primary physical targets, as follows: (i) implemented 4,159 subprojects (123% of target) from a total demand of - 7 -

11 around 10,000 proposals, a stark reminder of the massive deficits in basic infrastructure and services in this state and pent-up demand awaiting the follow-on project; (ii) financed at least one subproject for 276,145 beneficiary families (140% of target) representing about 1.4 million people in 212 municipalities, and including 5,952 indigenous families and another 4,102 Quilombola (Afro-descendent) families; (iii) organized some 4,000 community associations (236%) of which around 3,000 obtained financing for subprojects (177%); (iv) established 212 FUMAC Councils (183%), but no FUMAC-P (see 4.5). The NEPE completion notes that the Project was viewed by the State Government as a full partner in its execution of policies and public works in poor rural areas. Around 70% of all investments were financed through the FUMAC mechanism and the remainder through PAC. The effort to mobilize, analyze demands and supervise subproject implementation was shared by the 212 Councils, NEPE and its 18 Regional Development Units (RDU), as well as partnerships arranged with public and private entities to support Project execution. An average 66 families benefited per subproject with 74 families averaged for infrastructure and 38 for social investments. Numbers of beneficiaries per subproject varied little between the proposal and reality for productive subprojects, since income generated must be divided between the association members. The same was true for housing improvements, built on a per family basis. In the case of socio-economic infrastructure, mainly access roads and electrification, numbers of beneficiaries were significantly above original estimates. Subprojects per municipality averaged about 20, but some 24% or 979 investments were concentrated in just 10 municipalities, indicating significant differentiation between municipalities in levels of organization, leadership and/or political clout. Average cost per subproject was around R$43,000 (about US$14,000, significantly less than the permitted ceiling of US$50,000) and average cost per family was R$600 (about $200). Community demand for basic socio-economic infrastructure has traditionally emphasized rural electrification, housing improvement, water supply and other basics of human well-being. Nevertheless, as in all participating states, there is enormous pent-up demand for income and job-generating activities which will be a strong focus of the new Maranhão Integrated Project. Communities presented proposals for close to 2,000 productive investments of which 663 were approved and financed, representing 37 different types of activity. Stimulating greater investment in productive activities also implies focused attention on economic and financial feasibility, and marketing. Efforts will be made to expand communities participation in the ongoing Fair Trade initiative (Comércio Ético e Solidário) which establishes commercial links between small-scale Northeast producers and European supermarkets and other buyers for fruits, honey, crafts and other products. The discussion of subproject demand and selection of priorities before submitting proposals for funding proceeded as envisaged at appraisal, through the communities. Choices were made in open assembly with liberal participation of local people, not only association members. In a wise move designed to facilitate communities handling of the subproject cycle and enable them to make informed decisions, an Orientation Manual with basic Project rules, requirements and procedures (e.g., how to present a proposal to the Council, submit statements of expenses, contract service and other providers, and execute and maintain a subproject) presented in - 8 -

12 accessible language with graphics and illustrations, was distributed to all associations and Municipal Councils. Meetings were held to discuss this manual and training was provided. Councils adopted fairly uniform procedures for analyzing proposals. Upon receipt, Councils sent a committee of members to the respective communities to verify legitimacy of the proposal. Proposals, with committee opinion attached, were then read, discussed, voted and sent forward (or not) during a scheduled Council meeting. Communities contributed their 10% share, generally as labor or materials, and usually came to the RDU and Councils for help in preparing the technical proposal, installation of the subproject and for advice on submitting accounts. Technical assistance firms and autonomous professionals were contracted to prepare technical proposals and support construction. Communities elected a supervision committee for subproject execution, as per the Operational Manual. While this group lacked specialized technical knowledge of their subprojects, they had the advantage of knowing the locale and were vigilant about the quality and performance of contractors. The associations informed NEPE that they had not experienced difficulties throughout that phase but that the final, required submission of statements of account (prestação de contas) was complex and bureaucratic. Rural workers unions usually stepped in to assist communities at that point and Bank supervision missions worked with NEPE to simplify those procedures. Besides the discussion and interaction with the RDU in the selection and submission of demands, NEPE (2004) cites the views of Councils and associations on how subproject cycle efficiency might be further improved: (i) reduce documentation requirements; (ii) eliminate required legal sign-off for authentication purposes which is expensive and time-consuming; (iii) accelerate subproject analysis and release of resources to the associations; (iv) toughen NEPE s pre-qualification of technical assistance firms especially their understanding of local realities; and (v) increase NEPE/RDU technicians for Project supervision and monitoring, and resources, materials and trained personnel to support the Project. In its completion report, NEPE alludes to the comparatively high rate of rejection/devolution of proposals, mainly due to documentation issues, lack of available financing to meet the high level of demand and issues concerning the legitimacy of demand in some cases. The report also notes, inter alia, the excellent performance of the Councils in analyzing community proposals carefully and screening out those whose community-based credentials seemed suspect. While political pressure was a reality on occasion, whether from the mayor, councilmen, state deputies, or even INCRA technicians, the overall results were positive for the communities. NEPE acted to limit, to the extent possible, these attempts to influence the process, conducting awareness-building training with Councils and associations and following up on specific complaints. Indigenous and Afro-descendent Beneficiaries: Maranhão has an indigenous population of some 19,000 families (IBGE) distributed across 17 indigenous lands officially demarcated by the National Indigenous Foundation (FUNAI), and an Afro-descendent (Quilombola) population of some 56,000 families including city-dwellers. Based on core principles of social inclusion, the Project made special efforts to stimulate the participation of these groups, initially with training and awareness-raising activities and the preparation of sustainable development strategies for each community involved, followed by investments. NEPE staff includes a specialist in indigenous - 9 -

13 affairs. Some 5,950 indigenous families in three municipalities benefited from 109 subprojects, a level of capture three times their weight in the general population. Types of investments were mainly electric power, agricultural equipment, housing improvement and access roads. Similarly, the Project also benefited about 4100 members of Afro-descendent communities with investments mainly in electric power and access roads. In financial terms (around R$3.7 million) and number of beneficiaries, Maranhão represents the largest Bank-supported intervention in the Northeast region targeting indigenous communities. It should be emphasized that in each case, the communities themselves set priorities and executed their investments, typically contributing their share in labor. Important lessons were derived from this experience (see Section 8). A recent diagnosis revealed that while 25% of the indigenous population benefited, their communities continue to harbor high levels of poverty and social exclusion. Traditional methods of production (fishing, hunting and subsistence agriculture) do not adequately fulfill their survival needs or address new demands generated by their contact with non-indigenous populations, e.g. access to consumer goods, infrastructure and better quality of life. The new project in Maranhão further intensifies actions on behalf of these communities. Targeting. The Maranhão Project was accurately targeted; beneficiaries conformed to the intended profile. Project inclusion of indigenous and Afro-descendent groups, typically among the very poorest, demonstrated the state s commitment to sharp targeting of Project funding and benefits. Studies of targeting effectiveness across all participating states indicate that the CDD program methodology for reaching poor families works well. Beneficiaries are extremely poor by most key measures: educational level, occupation, location of dwelling and work, living conditions, domestic and productive assets, income, and the identification of a portion of beneficiaries as being subject to food insecurity and hunger (about 50% of all beneficiaries were estimated to have endured hunger and food shortages prior to implementation of Project investments in their community). Estimates also show that a large majority of Project beneficiaries have incomes below the poverty line. Studies performed in 2004 indicated that 73% of beneficiaries had an average monthly family income below R$ Another 25% had income slightly above this level but below R$ per month. Bearing in mind that average family size is five members, income per capita is below R$37.00 or less than US$1.00/day for over 73% of beneficiaries. In addition to income, baseline studies for selected participating states in 2000 indicated extremely low levels of access to basic services: 41% with electricity, 36% with septic tank, 28% with piped water, 24% with access to a health post, and 40% of families were illiterate. Given that Maranhão is the 2nd poorest state in Brazil, these averages are almost certainly higher than Maranhão in most if not all indicators. Institutional Development (US$6.9 million, 4% of total base cost) Rated Satisfactory This component financed training, technical assistance and consultancies to consolidate, refine and strengthen community participation and the operational capacity of community associations, the Municipal Councils and the Technical Unit itself. A wide variety of activities were developed and executed in connection with preparing the annual training program, community mobilization, training and technical assistance. In basic numbers, around 17,000 community members received

14 training; leadership skills were enhanced for some 4,050 FUMAC Council members; and close to 3,400 courses were delivered to community associations supporting subproject execution. Types of training included: Council and subproject management, local development planning, awareness-building, cooperative management, training for productive activities including bee-keeping, cattle-raising, vegetable production, fruit-growing, cheese-making and fish farming. Training methods stressed integrated, participatory planning targeted at the effective management of diverse subproject investments and more effective interaction with public authorities. Training covered indigenous and Afro-descendent communities as well as women s groups. A survey of the effectiveness of community training by NEPE saw 84% respond that the courses helped a lot (NEPE, 2004). NEPE personnel received training in 34 different categories, the emphasis always being on upgrading staff technical skills and qualifications to support the Project. The scope of training was broad and included: public accounting and financial management; gender in sustainable development; communications skills; project monitoring and evaluation; public procurement; information technology and management (including as a result, establishing a Project website); technical cooperation methodology; environmental management; a wide range of courses on administrative and organizational issues; and courses designed to support community productive activities, e.g., fish farming. The above activities/programs received skilled, effective support from the Brazilian Service for Support of Small and Medium Business (SEBRAE), NGOs, municipal mayors and governments, the State University of Maranhão (UEMA), and the National Indian Foundation (FUNAI), all of which provided instructors, as well as technical, administrative and methodological support. Project Administration, Monitoring and Evaluation (US$3.1 million, 3% of base cost) Rated Satisfactory This component performed as expected, supporting project coordination and activities to assess/measure project performance and impact; incremental operating costs (excluding salaries) of NEPE to enhance supervision and strengthen project operations; reformulating and upgrading the project management information system (MIS); materials/programs and outreach supporting the state s information diffusion campaigns to raise awareness of Project objectives, benefits, rules and methodology; monthly, six-monthly and annual reports on physical, financial, strategic and qualitative aspects of the Project; and an end-project evaluation. NEPE conducted close to 14,000 supervision visits to subprojects under execution and 1,700 visits to Municipal Councils. 4.3 Net Present Value/Economic rate of return: The SAR did not estimate an economic or financial rate of return for the Project. Instead, an analysis was provided for illustrative purposes, of the kinds of benefits achieved under the preceding reformulated Northeast Rural Development Project (R-NRDP) for the most common types of investments program-wide. Similar benefits were expected, on a larger scale, from the follow-on projects, including Maranhão. While estimates were made at appraisal of the types of investments most likely to be demanded, in practice this is a decision made by communities over the life of the Project and precise ex ante estimation of resource allocation, cost-effectiveness, rate of return and fiscal impact is not possible. It is evident for example, that appraisal significantly over-estimated likely demand for productive investments and this has been the case in most

15 participating states. The end-project numbers demonstrate vividly what is now well-known; poor communities opt for basic infrastructure before any other type of investment and in the Maranhão context of massive deficiencies, the investment outcome was rational. Overall, subprojects financed by the RPAP, including Maranhão, did not differ markedly from the previous program and thus a similar analysis was conducted of program-wide investments under the RPAP as an indication of cost effectiveness and sustainability, and of the financial viability of productive subprojects (see below and Annex 3). Cost Effectiveness and Sustainability About 77 percent of all subproject investments under the RPAP were infrastructure while social investments comprised less than 3 percent (Maranhão data shows 76 percent and 8 percent respectively). Most fall into the category of core public services (water supply, electrification, sanitation, access roads, housing renovation) with benefits supporting their universal provision. The CDD approach to rural infrastructure and service delivery for poor rural communities works cost-effectively in the Northeast due to several basic design features: (i) demand-driven subproject priority-setting and selection in a democratic forum such as the FUMAC Councils; (ii) standard technical designs and cost parameters, ensuring the use of least-cost models and providing established patterns of initiating and completing works; and (iii) delegating subproject implementation directly to community associations, thus generating cost-savings relative to similar works executed by public agencies. The contracting procedures established in the Operational Manual require direct, competitive contracting for all investments; associations solicit at least three bids, selecting the least-cost. RPAP costs were from 30-50% cheaper than publicly-executed subprojects of similar quality, based on an analysis of a random sample of subprojects including 10 categories comprising 80 percent of types financed. State-level evaluation of the RPAP in six Northeast states found that the quality of overall finish and quality of operation were technically and physically satisfactory for a large majority of subprojects in all states. Operation and maintenance was generally satisfactory with a majority of associations paying user fees where appropriate. A separate evaluation found that 89 percent of a sample of subprojects completed in 1995 and 1997/98 were still fully operational. Sustainability did not vary greatly between infrastructure, productive and social subprojects (89%, 87% and 88% respectively). 4.4 Financial rate of return: Internal rates of return were estimated for the most common types of productive subprojects, and exceeded on average, 30%. Benefits are largely concentrated in the beneficiary communities. The analyses assumed constant benefits over a ten-year project life. Sensitivity analysis performed on these illustrative subprojects showed that they tended to be reasonably robust vis a vis decreases in output and prices and increases in production costs. Benefit cost ratios are high (greater than 2.0) for the main types of productive investments analyzed and are generally financially sustainable. Although beneficiary associations do receive a one-time matching grant, the investments tend to be sustainable because cost recovery through user fees by the average beneficiary association is normally adequate to cover both O&M and replacement of original investments before the end of their useful economic life. All subprojects

16 show satisfactory internal rates of return with some quite high. In all cases, less than six years is required to recover the investment and less than 3.5 years in half of all cases. Fiscal Impact Evaluations show that subproject investments make a significant fiscal contribution through cost savings to state and municipal governments. For example, in most Northeast states, RPAP investments in water supply eliminated the need for water trucks to supply communities in time of drought (not generally applicable in Maranhão). Provision of clean, reliable water reduces public health costs and saves time spent by family members collecting water in often remote locations. Cost savings also derive from the community-driven Project design; as stated earlier, subprojects implemented by communities, directly or contracted, have consistently proven less expensive than publicly-executed works of similar type and quality. In terms of revenue generation, direct impacts are likely to be small, but indirect impact can be significant. Installation of electric power results in surging demand for domestic appliances, significantly increasing the ICMS (sales tax) revenue to municipalities.* Based on a sample of the 990-odd rural electrification subprojects as well as water supply investments financed in Maranhão between January 1998 and December 2000, NEPE projected the acquisition of over 16,000 television sets, about 13,000 refrigerators, 11,000 music sound systems, 6,900 blenders, 6,700 parabolic antennas, 4,800 flatirons, 870 washing machines and about 870 electric motors. *ICMS is collected by the state and all municipalities receive 25% of collections regardless of where collected. 4.5 Institutional development impact: As a participatory, demand-oriented mechanism, the Maranhão project involved many actors and stakeholders. All community associations and Municipal Councils received orientation and training about how the Project functioned, community organization, associative action, operations and responsibilities of Councils, and specific technical themes. Community Associations: Community associations created under the Project generally performed well. Members typically met their collective responsibilities and benefited from the efforts of NEPE and its RDU, as well as civil society entities to transmit information about the Project and create conditions for poor communities to participate. Membership is inclusive with women increasingly assuming leadership roles. Importantly, the benefits of subprojects have tended to reach non-members as well as members. Technical assistance and training are clearly critical for the proper functioning and sustainability of associations at their formative and longer-term operational phases. The NEPE report (2004) cites the opinions of surveyed association leaders and members, registering the universal call for even greater emphasis on the principal themes affecting their ability to manage their roles. Associations also noted the importance of a monthly contribution from members to maintain the association and especially to cover expenses such as the preparation and registration of documents pertaining to the association itself (cost R$200) and other key services/documents ranging from R$30.00 to R$ (cost of lawyer to review and sign the statute establishing the association). Other expenses include transportation and food for trips to the municipal or state capital to resolve land-related issues, tax issues or regularization of

17 documents. Associations which collect regular monthly fees from members have no trouble covering such expenses. Another issue raised by community leaders included the need for regular, more frequent meetings to maintain community interest and involvement. Surveys showed that participation is relatively high about 80%. Average membership of an association is 67, with an average 56% men and 44% women. An important finding from interviews with heads of household and rural leaders, is that associations can and should gain critical mass by acting jointly. Further, it was found that associations which had received productive subprojects continued to be active, while in some communities a degree of passivity set in following the acquisition of infrastructure which is associated in communities mind with public sector authority and responsibility. The report concludes that awareness-building and training are critically important to galvanize communities to continue to press for their needs both from within and outside the Project, over the medium to long-term. Municipal Councils: All 212 Municipal Councils received training in 2002 focused on basic operational aspects of the Project and their own functions/responsibilities. NEPE also initiated an evaluation of Council performance in 2002 assessing basic elements such as meetings, priority-setting, assistance provided for the organization and legalization of community associations, checking of community Carta Consultas; adherence to Project directives; relationship to public sector and elected officials, and level of member participation. Councils are not necessarily uniform, having different situations and characteristics, e.g., Councils exercising complete sovereignty over their deliberations and collaborating productively with authorities and local power groups; Councils where it is the active collaboration of the mayor and his local infrastructure which sets a positive stage for the Council s proper functioning and successful execution of the Project; some Councils which were moribund due to constraints on their proper functioning and where the intervention of NEPE and more training was required to re-constitute them; and cutting edge Councils where members are actively involved in mobilizing and organizing communities, supervision and evaluation of subprojects, providing technical assistance, introducing new themes into the discussion of local development (including environmental), preparing local development plans and active efforts to integrate the Bank-supported Project with other municipal, state and federal public policies and programs. NEPE (2004) cites the results of field surveys defining the attitudes of Council members. The message which emerges is a healthy self-confidence in their work and its significance, but lacking the resources in some cases to fully realize their potential and not having, in a significant number of cases, adequate support from the mayors, despite the substantial resources injected into their municipalities by the Project. A majority of the mayors however, are supportive, even paying for the registration and legalization of burgeoning community associations. The principal expenses of a Council include regular meetings to analyze subprojects, implying long journeys often on foot for members and participants to attend, food while they attend since many are away from their homes for several days, and the costs of subproject supervision, again signifying distance from their homes/communities. Members frequently bear these costs personally. This situation results in some Councils meeting less frequently than they should, delays in subproject analysis and approval, and reduced supervision activities. There was broad consensus on the need for an

18 operating fund (similar to the associations). Even so, the overriding concern expressed by both NEPE and the associations, relating to the functioning of the Councils, was the paucity of new leaders coming up from the communities/grassroots, and the systematic effort needed under the new Project to address this. FUMAC-P: The Project was slated to establish 20 FUMAC-P Councils but closed without action, although in 1998 several municipalities were selected as candidates and the issue of their creation remained a subject of discussion with supervision missions virtually to end-project. The Mid-term Review mission devised an alternative plan to gradually decentralize certain responsibilities to selected FUMAC Councils based on criteria such as sense of autonomy, deliberative ability, and level of participation in Council meetings and decision-making. They would assume certain activities including training and technical assistance and subproject supervision. Financial resource transfers were also contemplated and such Councils would be designated advanced FUMAC Councils. Lack of action is explained by: (i) NEPE s concern that the transfer of resources to such Councils would be viewed by non-fumac-p municipalities as a political victory, provoking their strong calls to be similarly benefited. However, the FUMAC-P pilot was never intended to be available on demand, but rather as a selective, performance and capacity-based promotion from FUMAC; (ii) many FUMAC Councils at the time were judged by NEPE to still need the more advanced organizational and administrative capacity permitting the independent and efficient management of Project resources required under FUMAC-P; and (iii) three changeovers in NEPE leadership and the reform of the State administration substantially modified NEPE s institutional structure and its willingness to pilot FUMAC-P. The new Project does not include this mechanism. NGOs/Civil Society Organizations: Substantial effort went into identifying civil society bodies with potential to support the Project (and the new Maranhão Integrated Project). The main streams/forums of civil society organizations in Maranhão are associated with poverty alleviation, environmental issues, gender and with Afro-descendent and indigenous groups. While their focus and primary objectives differ substantially, much of their effort goes into training social groups and activities associated with income-generation and improved wellbeing within a sustainable framework. Supervision and Mid-term Review missions found that there was a marked consensus among these bodies on what should be priority actions for rural areas: training and technical assistance; their own greater involvement in supervision, monitoring and evaluation of the Project; horizontal integration of public policies; lining these policies up more closely with sustainable development objectives/initiatives; and opening up specific instances where dialogue and cooperation could be fostered. As a group, their broad reach and ability to collaborate has proven valuable, their credibility in the communities is strong, they bring specific skills to the Project and accordingly, the State s plans to foster their continued and increased support is strongly-endorsed by the Bank. Technical Unit: NEPE administers and coordinates the Project, acts as a vehicle/conduit for communication with the Bank and State Government, and as main liaison with the Municipal Councils and in PAC municipalities, with the community associations. NEPE has significant administrative and financial autonomy within GEAGRO. Through a combination of Project-funded training, and the complementary impact of the State s mandatory administrative

19 reform program, NEPE was modernized, decentralized to 18 Regional Development Units which assumed responsibility for analysis, monitoring, technical assistance and subproject supervision along with training and data entry in the Management Information System (MIS), and strengthened in terms of information technology, plant and equipment. These measures brought the RDU into close contact with the Councils and associations, to the benefit of the Project. While admitting that even more could have been done in terms of awareness-building and training, NEPE holds that the Councils and associations fulfilled their roles well, acting as intermediaries between beneficiaries and NEPE managers. 5. Major Factors Affecting Implementation and Outcome 5.1 Factors outside the control of government or implementing agency: There were no major factors in this category. 5.2 Factors generally subject to government control: While there were several important administrative and organizational events instigated by Government, they had either positive or negligible impact on the Project and are outlined in Sections 7.5 and Factors generally subject to implementing agency control: No major factors subject to implementing agency control affected Project implementation and outcome. 5.4 Costs and financing: Total Project cost was US$106.7 million, on par with the appraisal estimate. There were two reallocations of Loan funds. An amount of US$4.8 million was reallocated from Category 1(b) Pilot Grants (FUMAC-P) as follows: US$3.0 million to Category 1(a) (FUMAC Grants); US$1.5 million to Category 1(c) (PAC Grants); and US$300,000 to Category 3(a) (Incremental Operational Costs). This was followed by a second reallocation of US$1.1 million as follows: US$500,000 from Category 1 (b) (Pilot Grants) and US$600,000 from Category 4 (Unallocated) to Category 2 (Consultants services and training for Parts B and C of the Project). The Loan was fully-disbursed. The first reallocation from Category 1(b) was an acknowledgment (and early indicator) that FUMAC-P was unlikely to progress. The second reallocation was used to benefit Afro-Brazilian and indigenous communities comparatively numerous in Maranhão - financing training in leadership, project rules/substance and impact evaluation in advance of actual subproject investment activity. A six-month extension of the Closing Date to end-june 2004 provided continuity for preparation of the new Maranhão Integrated Project, using a small, remaining Loan balance. Procurement: All procurement of goods/works was carried out in accordance with Bank guidelines, and in the CDD context was both innovative and effective. Acquisition of mainly goods and works under the community subproject component was handled directly by the community associations in all cases, and subject to at least three price quotations. Supervision

20 missions vetted procurement through random reviews of subprojects both in the field and analysis of NEPE documents. NEPE handled procurement of consultants to provide technical assistance to the communities, and to support its own responsibilities, with satisfactory performance and outcomes. In addition, the use of standard subproject designs and unit costs, required by the Operational Manual, and providing associations with basic information on how to proceed, simplified procurement, reduced the incidence of under- or over-dimensioned investments, and made supervision and monitoring smoother. The community cost-sharing requirement (10% minimum) under these CDD projects correlates directly with communities seeking competitive prices through three quotations, fostering economy and efficiency. 6. Sustainability 6.1 Rationale for sustainability rating: Sustainability is rated Likely Institutional: The Project re-affirmed that delegation of responsibility and influence to local-level organizations and involving local government can successfully support decentralized resource allocation and social capital creation in rural Maranhão. Linking the Councils to local municipal planning and budgeting processes and programs, as well as alternative sources of credit and financial services, fosters sustainability while scaling up Project impact. NEPE s sustainability is intimately related to the skills and experience of its core technical team and its broad involvement with other state and Federal programs, including the Bank-supported Crédito Fundiário land reform program. Programmed actions are underway to stimulate greater interest and commitment from municipal mayors and to foster strategic partnerships between the mayors and NEPE s RDUs to strengthen the Councils and realize their potential. Similarly, NEPE is committed to working even harder to fortify the communities, strengthening their capacity to manage the subproject cycle, and especially to administer Project-funded goods and structures. Despite major efforts to publicize and mobilize communities, NEPE believes that an even more intense capacity-building program would enable weaker community associations in some areas to fend off political encroachment and to better manage their investments. Financial: Financial analysis and field investigation confirm the sustainability of investments financed by the R-NRDP and RPAP projects. A sample of over 8,000 subprojects in Northeast states which executed Projects as of 1995 (R-NRDP) and 1997/98 (RPAP) was reviewed in Some 89% of all of them were still fully operational. Analysis of productive subprojects revealed that cost recovery through user fees by beneficiary associations was adequate to cover both operation and maintenance and replacement of the original investment well before the end of its useful economic life. NEPE (2004) comments at some length on operation and maintenance performance under the Project, especially of productive subprojects. As noted, a high proportion of communities are expanding their initial investments, increasing planted area, re-investing to expand small-scale agro-industries, maintaining goods acquired under the Project (tractors, equipment, processing facilities), and augmenting their businesses (e.g., selling ice-cream following Project financing of an ice-making facility). Physical: Participation by beneficiaries throughout the subproject cycle, together with significant community cost-sharing, fosters a senses of the value of the investment in both social and financial

21 terms. In addition, the democratic process intrinsic to the representative Municipal Councils ensures better selection and priority ranking of investments by beneficiaries, contributing to their longer-term sustainability. Nevertheless, in Maranhão, research observations towards the end of the Project noted the lingering perception by some communities that O&M was the responsibility of the mayor. Maintenance of access roads presented particular difficulties in terms of who was responsible and equipped to do the work; similarly, water supply systems need focused attention to ensure that communities understand and are committed to the payment of adequate user fees. Domestic connection inside the dwelling has been repeatedly found to promote regular payment. Even so, NEPE s completion report finds that most communities find solutions to O&M issues and administer their investments effectively. The new project introduces performance incentives for community O&M, to be monitored by the Councils and RDU. NEPE (2004) states that a sizeable proportion of beneficiary associations are operating their investments in a professional manner, effectively managing their production, marketing, distribution of profits between association members and capital reinvestment. Surveys showed, in the case of tractors, that users pay on average, R$25/1.5 hours if members and R$30 if non-members, plus fuel costs. Users also pay a fee to use the tractor for hauling produce to market. For manioc processing, a fee of 10% of the value of flour produced is levied, i.e. 3 kg of flour from each 30 kg. In the case of rice, the average fee is R$3.00/60 kg sack or R$0.05/kg. Members pay R$200/month for use of boats and nets acquired from fishery subprojects. However, the maintenance of reserve funds is not generalized across all subprojects and payment of user fees is still not uniformly adopted (where relevant) for all subprojects. The main issue affecting Project-funded productive subprojects in Maranhão is inculcating a disciplined approach to establishing and maintaining a reserve fund for re-stocking raw materials, maintenance and supplemental investments. Association leaders interviewed mentioned members who still need an appreciation of their subproject as an asset requiring long-term care, or as something to which they need to contribute financially or with their labor. There are also issues with some socio-economic infrastructure investments. In the case of water supply investments, the association charges members R$3.00 per residence, and while the vast majority pay, not all pay regularly, and this can compromise O&M and even payment for the electric power or fuel to run a system. Some associations still need to develop mechanisms to secure payment per se and regular payment, by every member/user. Maintenance of roads and access routes is also problematic due to constant rains, sandy lands, and intensified vehicle usage especially by heavy trucks which do not come from the community. Communities which have a tractor patch the roads to the extent possible, but the real problem lies with the lack of long term action by the mayors and by the State Department of Roads (DER-MA), to maintain what the Project finances. Electricity networks are transferred to the Maranhão Electricity Centers (CEMAR) which regularly handles supervision and maintenance of lines, and charges for electricity consumption by users. Public lighting has been installed in all beneficiary communities. Not surprisingly, there are some technical difficulties such as interruptions in energy provision which can affect water supply systems and operations of small processing facilities, or artesian wells which operate at less than full capacity, but generally, communities manage by administering the resource on a fixed time schedule and with schemes for saving or storing the resource. In general, positive experiences predominate. The developmental issues raised above have been

22 well-documented and are resolvable; measures are being taken by NEPE and the RDUs to establish more systematic, sustainable ways of handling them. 6.2 Transition arrangement to regular operations: Transition arrangements to regular operations mean, in this case, execution of a subproject, its release to the community association (which legally and technically owns the investment) and its becoming operational under pre-established rules and procedures for this phase. In the case of electric power investments - and similarly with all other participating states - operation and maintenance are handled by the local power agency/concession company. 7. Bank and Borrower Performance Bank 7.1 Lending: The Bank s performance during identification, preparation and appraisal is rated Satisfactory. The Bank recognized that the R-NRDP represented a successful shift away from the traditional rural development model and that it merited follow-up based on lessons learned. In addition, as one of the last projects prepared under the RPAP, Maranhão benefited from early experiences/lessons flowing in from its RPAP predecessors which were already under implementation, in some cases for two years or more. The preparation process was marked by openness to further innovation, a hallmark of the Northeast CDD program since its inception. 7.2 Supervision: Bank supervision is rated Satisfactory. About 86 staff weeks were expended on supervision, rather than the 35 estimated at appraisal, a significant under-estimation given characteristics of the state and estimated supervision requirements for other states in the program, which tend not to drop below 75 weeks. Project execution benefited from stable Bank task management for the full duration of the Project (with similar stability in NEPE) and regular supervision missions. The Bank s Recife Office once again proved its value in terms of proximity, experienced Team leadership, Task Manager and administrative personnel, long-standing relationship with the State Government and NEPE, knowledge of the Northeast CDD mechanisms and their evolution, and decentralized structure in keeping with the nature of the Northeast rural poverty program. The Bank defined and addressed implementation issues as they arose, in cooperation with NEPE, and based on time-bound action plans when indicated. Supervision missions used a prepared advisory document showing updated performance indicators and highlighting any apparent issues, plus a MIS-generated random sample of municipalities for field visits. The Project was correctly rated Satisfactory for Development Objectives and Implementation Progress throughout. Principal issues of supervision included: (i) the potential impact on the Project of the State institutional reform (which suspended subproject approvals and financing for about three months, without noticeable effect on the Project ); (ii) upgrading the quality of services provided to communities by subproject planning firms and construction contractors; (iii) fostering communities understanding of Project rules including the importance of establishing operation and maintenance funds; (iv) capacity-building/training for Councils, associations and the RDU, using NEPE professionals and contracted NGOs, to review the Operational Manual, design tools for identifying and reaching poor communities, and reinforcing understanding of roles and

23 functions of community leaders; (v) improving through better up-front analysis, the financial viability of productive subprojects; (vi) postponed creation of FUMAC-P Councils and promoting greater community understanding of the role and responsibilities of FUMAC Councils; and (vii) leveraging, through the use of autonomous entities, professional assistance to supervise water supply, small-scale dams and community road improvement investments. NEPE also simplified its overly-bureaucratic requirements for communities submission of statements of expenses (prestação de contas), substantially reducing the number of associations technically in default due to incomplete documentation or other issues. The consolidation of RDU offices and their local level supervision capacity contributed to this reduction. Related to this, supervision also effected a significant and positive revision in Project procedures, permitting the release of subproject financing to the community bank account in one installment instead of several. This meant that subproject expenses could be paid directly, with authorization by the associations and on-site verification by the RDU. Advantages included greater sense of local control and transparency, better relations with service providers, and streamlined bureaucratic procedures for getting resources to the beneficiary communities. 7.3 Overall Bank performance: Rated Satisfactory The Rural Poverty Alleviation Program of eight projects including Maranhão received a LAC Region Performance Award and a Bank Social Development Award for Excellence in Project Quality in In addition, the Northeast rural poverty reduction program was selected for presentation as Best Practice in scaling up CDD approaches, at the 2004 Shanghai Poverty Conference. Borrower 7.4 Preparation: The Borrower s performance during preparation is rated fully Satisfactory. Preparation was straightforward with full support of State authorities, the strong commitment of NEPE whose experience went back to the POLONORDESTE program of the 1970s and early 80s, and broad recognition that the R-NRDP had successfully changed the way rural development could be done. At the time of Project appraisal, the State Governor increased NEPE s autonomy, revised its organizational structure in line with the State s decentralization strategy for the public sector, as well as the specific needs of the Project, and provided NEPE with new office premises adapted to its operational needs. 7.5 Government implementation performance: The State Government did not have a direct, day to day role in Project implementation, except through its counterpart funding/cost sharing responsibilities, which were satisfactory throughout, and indirectly through the role and performance of NEPE. The intervention and support of the State Governor when needed, was also important to the success of this Project. The profound and ambitious State institutional reform of early 1999 caused a 3-month cessation of subproject approvals and processing while important changes and adjustments were made, but this did not

24 materially affect the Project. As part of this reorganization, Government s creation of 18 regional offices of NEPE in fact, greatly benefited the Project by bringing NEPE closer to the associations and Councils. 7.6 Implementing Agency: Quality Project implementation and outcomes correlate closely with NEPE s experienced coordination and administration, Project-funded training opportunities in areas directly related to the Project, and the modernization exercise of The unit produced a high quality completion report which enriched the ICR. NEPE worked willingly and effectively with the Bank in areas needing focused attention and/or improvement. Three turnovers in NEPE leadership did not materially affect the Project because the core technical team and administrative procedures were not modified. The Coordinator at project outset withdrew to pursue a political career and became a strong advocate for the Project in the State Legislative Assembly. In 2002 the replacement Coordinator resigned for health reasons, was followed briefly by a third person who was then substituted by the new State Governor s nominee, the current Coordinator. In May 2002, NEPE was detached from the State Secretariat of Planning (GEPLAN) and came under the umbrella of the State Management Unit for Agriculture, Livestock and Rural Development (GEAGRO), the result of a mini-reorganization instituted by the new Governor. NEPE continued as a relatively autonomous body and actually benefited from coming under GEAGRO due to that agency s stronger operational culture in contrast to GEPLAN s more strategic/planning orientation. Following the State Reform Program, NEPE decentralized subproject analysis, supervision, technical assistance and monitoring to its 18 Regional Development Units (RDU, which operate in conjunction with 18 Casas de Agricultura providing special assistance to NEPE s field operations), along with training activities for Councils and communities, and project data entry in the MIS. While NEPE acknowledges that mobilization and organization of communities and Councils could have benefited from even greater intensity the scale and dispersion of rural poverty, and poor transportation conditions in this state represent a daunting organizational challenge the 4,159 subprojects contracted were 20% above target (3,392) and NEPE/RDU made some 23,000 visits to communities for monitoring, technical assistance and supervision. NEPE s efforts also helped Councils and associations to fulfill their roles effectively and become effective intermediaries between beneficiaries and project coordination. RDU staff were, for practical reasons, recruited from their own area even though this meant in some cases that the staffing cohort and hence capacity - varied between zones. Some RDU had up to 17 permanent technicians agronomists, vets and agricultural technicians, as well as temporary contracted workers. In its end-project report, NEPE points to additional efforts it is making to further improve its operational efficiency and to obtain more consistent results state-wide. When asked to list the types of frustrations they encountered, technicians noted: (i) delayed processing of subproject proposals due to overwhelming demand; (ii) limited access to remote areas in the rainy season for subproject supervision and need for more vehicles; (iii) rotation of technicians in NEPE and the RDU; (v) sizing and labor problems with some access road subprojects due to executing companies inadequate knowledge of local conditions; (vi) need for more specialized technical personnel experienced in handling productive subprojects; and (vii) heavy bureaucratic requirements of the subproject cycle

25 Despite these constraints, which are normal for a Project involving thousands of small investments in a large and exceptionally poor state, NEPE/RDU staff expressed gratification in accompanying communities progress and the improving infrastructure in the interior of the state. Where there was backwardness there is now economic progress; where there were earthen roads and sandy, impassable paths, there are today asphalt roads with electric poles along the margins; where there were difficulties in obtaining potable water, it is now alongside or inside homes; where there was darkness or kerosene lamps, homes, schools and shops are now lit; and where there was only a subsistence plot, there is now flourishing production including for market (NEPE 2004). 7.7 Overall Borrower performance: Rated Satisfactory The Borrower performed exceptionally well and was always committed to and supportive of, the Project. Counterpart funding availability was never an issue. 8. Lessons Learned Project lessons reinforce and validate previous experience in this state and in other Northeast states with similar projects, and have influenced the design of the follow-on-project (Loan 4735-BR): Decentralization and participation: Decentralization of fiscal and investment decision-making and implementation to the state and local levels including community organizations, results in administrative efficiency and superior results. Involving beneficiaries in the selection, financing, execution, and O&M of subprojects ensures that investments meet genuine community needs, leads to cost savings, boosts community ownership and increases the sustainability of investments. CDD approaches: By assigning responsibility to local communities and their organizations, and by involving municipal government as a partner, CDD approaches can successfully support decentralized resource allocation and create social capital in even the poorest rural areas. This method also reduces bureaucracy and administrative obstacles, and reinforces accountability for project performance by eliminating the distance between decision-makers and beneficiaries who, at key points in the subproject cycle, are synonymous. Self-selection of beneficiaries results in strong targeting. The socio economic profile of project beneficiaries in Maranhão indicates that a high proportion are illiterate, with insecure access to basic foodstuffs; the vast majority have incomes well below the poverty line. Poverty targeting that is simple, verifiable and based on these kinds of objective criteria, fosters transparency, reduces political interference in project resource allocation and ensures that the vast majority of project resources reach the poorest. Alignment of community investments with the Municipal Human Development Index (HDI-M) as a targeting and mobilizing goal, establishing mechanisms for local level integration of project activities with other programs to increase targeting effectiveness and poverty impact, is both a logical evolution of the Northeast CDD approach and an important way to further scale it

26 up. In practical terms, this means strategies which concentrate financial resources and public policies on the poorest municipalities, based on HDI-M, while preserving the participatory institutional mechanisms already established. Under the new Maranhão Integrated Project, 60% of total funding for subprojects is allocated to the 80 municipalities with lowest HDI-M. Natural resources management (NRM) and sustainability issues require proactive attention in environmentally fragile states such as Maranhão, located in a transition zone between the semi-arid Northeast and the Amazon region. Environmental protection checklists, periodically updated, are essential even though the small scale of investments virtually eliminates significant environmental effects in most cases. Standard subproject designs have special importance where environmental concerns are relevant, while environmental TA and training to FUMAC Councils and community groups delivered by in-house environmental expertise in the Technical Unit if possible fosters broad awareness.* * The new Project gives communities the option of investing in environmental subprojects, piloting regional or territorial investments across several municipalities (e.g., watershed investments) and demonstration subprojects in NRM themes. Productive subprojects are likely to assume greater prominence under the new project. They require rigorous selection, preparation, technical assistance and supervision criteria and linkage to a graduation strategy which facilitates communities access to formal credit. The idea of a one time matching grant for productive ventures, as intended under the new Maranhão project, makes sense under the following conditions: services cover a majority of community members; use is regulated by strict operational guidelines; O&M is assured by payment of user fees; meets impact criteria for HDI and environment; and, provision for technical assistance is built in. Calculations made by supervision missions on the basis of cost and income data collected on-site showed that a well-thought-out productive subproject with adequate technical assistance could double or treble family income in months. Municipal Councils can play a broader role in local planning. FUMAC Councils can establish links to local municipal planning and budgeting processes and programs as well as alternative credit and financial services, including state and federal. Carefully nurtured with technical assistance, information and funding support, and maintaining transparent, comprehensible rules of operation, the Councils could become one of the most valuable and durable institutions for rural development in the Northeast. Attempts to establish FUMAC-P Councils have not been equally successful in all states and should not be pursued if/when a state has coherent reasons for opposing them. Indigenous community organizations can establish and express priorities, and administer resources received. Project investments for these groups were successful due to the role of FUNAI and other public sector personnel including NEPE, in organizing and training communities in advance of their actual investments. This experience has encouraged a general re-orientation of indigenous public policy in Maranhão, moving away from an assistance approach towards one of integration and participation. The vast majority of indigenous communities, when compared to non-indigenous, showed greater receptivity to participation in community activities, in part due to their tradition of collective action

27 Intensive supervision which includes the use of local entities closest to the communities, is an indispensable determinant of success and sustainability for CDD projects. FUMAC Councils have the capacity to assume greater responsibility for supervision, with appropriate training, TA and information technology. The Bank s Recife Office remains vital for project coordination, supervision, procurement reviews and field visits, with task managers dedicated exclusively to the Northeast region. In this manner, Bank support meshes with the decentralized structure of the Northeast program, and signals Bank commitment to the region. Technical assistance of diverse type and delivery -- is an essential ingredient for CDD projects and presents particular challenges for its systematic organization, location, delivery and supervision in poor rural regions. Formal/informal partnerships between project technical units and diverse entities (public and private agencies, NGOs, Church groups, local rural workers unions) capable of delivering quality technical assistance can make an important contribution to resolving this issue. The appointment of technical advisors recruited by each FUMAC Council and trained by the TU to manage and coordinate local technical and training needs, can have tangible benefits and has been introduced under the new project in Maranhão. The project management information system (MIS) is part of a standardized system established in all participating Northeast states, capable of generating real time data/reports on the full subproject cycle and supporting monitoring and evaluation activities. The system is flexible, permits continuous upgrading and under the new Maranhão project for example, will: (i) link project physical implementation data with financial information; (ii) integrate with State information systems to facilitate M&E of the State s multi-year development plan (PPA); and (iii) measure income, welfare and social capital gains through a comprehensive impact evaluation component. Standardized subproject documents, technical designs and unit costs periodically reviewed, updated and expanded -- simplify subproject preparation and evaluation, improve subproject quality, facilitate procurement, prevent over- and under-design and promote the participation of poorer communities. Checks and balances, clearly-defined and well-disseminated, are essential to ensure proper targeting and use of resources and to boost and institutionalize transparency. This project confirmed the practical utility of rules and procedures -- including guidelines for performance incentives and penalties, periodically reviewed and updated set down in a Project Operational Manual. 9. Partner Comments (a) Borrower/implementing agency: The Borrower provided the following comments on the Bank s draft ICR:

28 - 25 -