Document of The World Bank IMPLEMENTATION COMPLETION AND RESULTS REPORT (IBRD IBRD-73990) ON A LOAN IN THE AMOUNT OF US$22.5 MILLION EQUIVALENT

Transcription

1 Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Document of The World Bank IMPLEMENTATION COMPLETION AND RESULTS REPORT (IBRD IBRD-73990) ON A LOAN IN THE AMOUNT OF US$22.5 MILLION EQUIVALENT AND AN ADDITIONAL FINANCING LOAN IN THE AMOUNT OF US$22.5 MILLION EQUIVALENT TO THE STATE OF PIAUÍ FOR A RURAL POVERTY REDUCTION PROJECT July 26, 2010 Report No: ICR Sustainable Development Department Brazil Country Management Unit Latin America and Caribbean Region

2 CURRENCY EQUIVALENTS (Exchange Rate Effective January 31, 2010) Currency Unit R$1.00 = US$ US$ 1.00 = R$1.874 FISCAL YEAR January 1 to December 31 ABBREVIATIONS AND ACRONYMS AF AGESPISA CA CAS CDD CPS FECAMP FUMAC FUMAC-P HDI-M IICA INCRA IRR IBGE KPI MC MDS MIS O&M OP PAC PDO PPS RPAP SCI SEBRAE SEPLAN SINAPI SOE SP STU Additional Financing Water and Sanitation Company of Piauí Community Association World Bank Country Assistance Strategy Community-driven Development Country Partnership Strategy Federal University of Campinas Municipal Community Scheme Pilot Municipal Community Fund Municipal Human Development Index Inter-American Institute for Agricultural Cooperation Institute for Colonization and Agrarian Reform Internal Rate of Return Brazilian Institute of Geography and Statistics Key Performance Indicator Municipal Council Ministry of Social Development Management Information System Operation and Maintenance Original Project State Community Scheme Project Development Objective Physical Performance Study Rural Poverty Alleviation Project Social Capital Index Brazilian Service for Support to Small Business State Secretariat of Planning Piauí National Civil Construction Index State of Piauí` Statement of Expenditures Subproject State Technical Unit Vice President: Country Director: Pamela Cox Makhtar Diop

3 Sector Manager: Project Team Leader: ICR Team Leader: Ethel Sennhauser Edward W. Bresnyan Edward W. Bresnyan

4

5 BRAZIL Piauí: Rural Poverty Reduction Project A. Basic Information...i B. Key Dates...i C. Ratings Summary...i D. Sector and Theme Codes...ii E. Bank Staff...ii F. Results Framework Analysis...iii G. Ratings of Project Performance in ISRs...vii H. Restructuring (if any)...viii I. Disbursement Profile...ix 1. Project Context, Development Objectives and Design Key Factors Affecting Implementation and Outcomes Assessment of Outcomes Assessment of Risk to Development Outcome Assessment of Bank and Borrower Performance Lessons Learned Comments on Issues Raised by Borrower/Implementing Agencies/Partners Annex 1. Project Costs and Financing Annex 2. Outputs by Component Annex 3. Economic and Financial Analysis Annex 4. Bank Lending and Implementation Support/Supervision Processes Annex 5. Beneficiary Survey Results Annex 6. Stakeholder Workshop Report and Results Annex 7. Summary of Borrower s ICR and/or Comments on Draft ICR Annex 8. Comments of Co-financiers and Other Partners/Stakeholders Annex 9. List of Supporting Documents MAP: State of Piauí... 81

6

7 A. Basic Information Country: Brazil Project Name: Rural Poverty Reduction Project - Piaui Project ID: P L/C/TF Number(s): IBRD-46240,IBRD ICR Date: 07/27/2010 ICR Type: Core ICR Lending Instrument: SIL Borrower: STATE OF PIAUI Original Total Commitment: USD 22.5M Disbursed Amount: USD 45.0M Revised Amount: USD 45.0M Environmental Category: B Implementing Agencies: SEPLAN Cofinanciers and Other External Partners: B. Key Dates Process Date Process Original Date Revised / Actual Date(s) Concept Review: 10/10/2000 Effectiveness: 04/22/ /22/2002 Appraisal: 03/05/2001 Restructuring(s): 06/15/ /15/ /22/2010 Approval: 06/26/2001 Mid-term Review: 12/07/2004 Closing: 06/30/ /31/2010 C. Ratings Summary C.1 Performance Rating by ICR Outcomes: Risk to Development Outcome: Bank Performance: Borrower Performance: Moderately Satisfactory Moderate Moderately Satisfactory Moderately Satisfactory C.2 Detailed Ratings of Bank and Borrower Performance (by ICR) Bank Ratings Borrower Ratings Quality at Entry: Moderately Satisfactory Government: Moderately Satisfactory Quality of Supervision: Moderately Satisfactory Implementing Agency/Agencies: Moderately Satisfactory Overall Bank Overall Borrower Moderately Satisfactory Performance: Performance: Moderately Satisfactory i

8 C.3 Quality at Entry and Implementation Performance Indicators Implementation QAG Assessments Indicators Performance (if any) Potential Problem Project Yes at any time (Yes/No): Problem Project at any time (Yes/No): DO rating before Closing/Inactive status: Yes Moderately Satisfactory Quality at Entry (QEA): Quality of Supervision (QSA): None None Rating D. Sector and Theme Codes Original Actual Sector Code (as % of total Bank financing) Irrigation and drainage 15 5 Other social services Roads and highways 45 5 Sub-national government administration 8 10 Water supply Theme Code (as % of total Bank financing) Participation and civic engagement Rural non-farm income generation 17 5 Rural policies and institutions 17 5 Rural services and infrastructure E. Bank Staff Positions At ICR At Approval Vice President: Pamela Cox David de Ferranti Country Director: Makhtar Diop Gobind T. Nankani Sector Manager: Ethel Sennhauser Mark E. Cackler Project Team Leader: Edward William Bresnyan Luis O. Coirolo ICR Team Leader: ICR Primary Author: Edward William Bresnyan Anna F. Roumani ii

9 F. Results Framework Analysis Project Development Objectives (from Project Appraisal Document) To assist the State of Piaui to reduce currently high levels of rural poverty by: (a) improving wellbeing and incomes of the rural poor through better access to basic social and economic infrastructure and services and support for productive activities, using proven community-driven development (CDD) techniques; (b) increasing the social capital of rural communities to organize collectively to meet own needs; (c) enhancing local governance by greater citizen participation and transparency in decision-making, through creation and strengthening of community associations and Municipal Councils; and (d) fostering closer integration of development policies, programs and projects at the local level, by assisting Municipal Councils to extend their role in seeking funding, priority-setting and decision-making over resource allocation. Revised Project Development Objectives (as approved by original approving authority) PDO was not revised. (a) PDO Indicator(s) Indicator Indicator 1 : Value quantitative or Qualitative) Baseline Value Original Target Values (from approval documents) Formally Revised Target Values Number of families benefited from subproject investments. Incremental 36,000 81,600 families families Zero. (Original Project - (Additional OP). Financing - AF). Actual Value Achieved at Completion or Target Years Total families benefited 144,655; 124,000 without repetition. Date achieved 04/22/ /30/ /31/ /31/2010 Comments (incl. % achievement) Achieved: Combined operations reached 144,655 families (123% of target) and 124,000 without repetition (no target). OP: 82,875 fams (102%) and AF: 61,800 fams (172%). Indicator 2 : Incremental employment generated from subproject investments. Value quantitative or Qualitative) Estimated 2,550 Zero. OP: No target. AF: No target. jobs created. Date achieved 04/22/ /30/ /31/ /31/2010 Comments (incl. % achievement) Achieved: 2,550 jobs created, extrapolated from total subprojects (SP) financed. 27% of SP created jobs directly; 10% created at least one job/family; 13% created jobs for >3 jobs/family, averaging direct employment of 1.3 persons/sp. Indicator 3 : Increase in wellbeing and incomes of project beneficiaries Value quantitative or Qualitative) Zero. OP: No target. AF: No target. Wellbeing impact positive. Economic analysis suggests positive income gains. iii

10 Date achieved 04/22/ /30/ /31/ /31/2010 Achieved: (i) Wellbeing: More project families gained access to water, electricity Comments than control group. Marked reduction in infant mortality, Chagas disease, (incl. % Hepatitis. (ii) Income: IRRs and Cost/Ben analyses suggest positive income achievement) gains (see Annex 3). Indicator 4 : Increase in social welfare of rural communities. Value quantitative or Qualitative) Zero. OP: No target. AF: No target. Social welfare benefits substantial. Date achieved 04/22/ /30/ /31/ /31/2010 Comments (incl. % achievement) Achieved: (i) Higher family incomes from est. 2,550 jobs; (ii) 36,000 fams with access to electricity; (iii) 26,000 fams with access to clean water and to basic health centers/services; and (iv) 8,600 fams with better sanitation facilities. Indicator 5 : Increase in social capital index (SCI) of project Municipal Councils (MC) Value Strong evidence of quantitative or Qualitative) Zero. OP: No target. AF: No target. social capital formation/growth. Date achieved 04/22/ /30/ /31/ /31/2010 Comments (incl. % achievement) Indicator 6 : Value quantitative or Qualitative) Achieved: MCs effective in representing community interests; better interface b/w communities and local auths; greater transparency, social control over Mun. decisions. 81% Mun. auths. saw MCs as positive influence on Mun. admin. (SCI not re-calculated). Number of Municipal Councils participating in priority-setting and decisionmaking on resources allocation of project and non-project funded development activities. 133 Municipal Councils established by previous project (RPAP). OP: No target. End-project, 221 MCs (66% increase) deciding and allocating AF: No target. project resources. About 25 MCs deciding allocation non-project resources. Date achieved 04/22/ /30/ /31/ /31/2010 Comments (incl. % achievement) Achieved: Binswanger (2006) found about 39 MCs discussing other programs of which 25 MCs (11% overall total) were actually deciding allocations from those programs to their communities. Indicator 7 : Increase in total project and non-project financing allocated through the Municipal Council mechanism. Value quantitative or Qualitative) Previous project (RPAP), MCs represented 63% of total SP financing (US$20.9 m). MCs' OP: No target. deliberation/allocation of non-project resources still incipient. (i) By end-op and AF, aggregate 95% of project financing for SP had been AF: No target. allocated by MCs. (ii) US$5.7 m. in State and federal resources also allocated via MCs. iv

11 Date achieved 04/22/ /30/ /31/ /31/2010 Comments (incl. % achievement) Partially achieved: The US$5.7 m. in non-project resources were incremental, and discussed/decided in the MCs. Project appraisal estimated ratio of 1:5 (Loan to non-project funds). Outcome fell short but a positive contribution. Indicator 8 : Number of communities graduated from the program. Value quantitative or Qualitative) Zero. OP: No target. AF: No target. 403 community associations graduated from further productive SPs. Date achieved 04/22/ /30/ /31/ /31/2010 Achieved: (i) Project permitted an assn. grant financing for one productive SP; Comments (ii) 403 assns with productive SP graduated (but not from project as a whole); (incl. % and (iii) as customary, informal graduation also via project's targeting/allocation achievement) mechanisms. (b) Intermediate Outcome Indicator(s) Indicator Indicator 1 : Value (quantitative or Qualitative) Original Target Values (from Baseline Value approval documents) # Community subprojects implemented. 1,344 subprojects financed under previous (RPAP) project. OP: Target 1,200 SP. Formally Revised Target Values AF: Target 1,200 additional SP, reduced to 500 due higher average cost per SP. Actual Value Achieved at Completion or Target Years Aggregate OP and AF subprojects financed: 1,902 (112% of aggregate target). Date achieved 04/22/ /30/ /31/ /31/2010 Comments Achieved: 112% of aggregate (revised) target of 1,700 SP, and 79% of original (incl. % aggregate appraisal targets of 2400 SP. achievement) Indicator 2 : # Municipal Councils created and operational. Value (quantitative or Qualitative) 133 MCs established under previous project (RPAP). OP: 77 additional MCs (aggregate 210) created and operational. AF: Target fully achieved under OP. No new target. Incremental 89 MCs created and operational (116%). Date achieved 04/22/ /30/ /31/ /31/2010 Comments Achieved: By end-op, 189 MCs in 222 project-eligible municipalities. By end- (incl. % AF, 222 Councils, aggregate gain of 89, and 100% coverage. achievement) # Training courses offered to beneficiary associations, Councils, STU and others Indicator 3 : to build capacity. Value (quantitative or Qualitative) Zero. OP: Targeted 192 training courses for associations, AF: Targeted 87 courses (zero for OP: 545 courses; AF: 88 courses. v

12 MCs, STU MCs). staff/technicians. Date achieved 04/22/ /30/ /31/ /31/2010 Comments (incl. % achievement) Achieved: Project reached 284% and 100% of OP and AF targets respectively. Despite zero target for MC training under AF, STU delivered 33 courses. See Main Text and Annex 2. Indicator 4 : % women in Municipal Councils and community associations. Value (quantitative or Qualitative) Zero. Under RPAP, women's participation noted but not quantified. OP: No target. AF: No target. Project fostered women's participation. Date achieved 04/22/ /30/ /31/ /31/2010 Achieved: (i) MCs: of 190 MCs, 26 (14%) presided by women; in 169 MCs, Comments min. one female board member; (ii) Assns: of 1,026 assns, 187 (18%) had (incl. % female president; 893 (87%) had female Treasurer; 100% had women on Fiscal achievement) Committee (Batista 2009). Indicator 5 : Cost-effectiveness and quality of basic infrastructure and social subprojects. Value (quantitative or Qualitative) Zero. But, previous project (RPAP) ICR reported high indices for quality and costeffectiveness. OP: No target. AF: No target. High cost effectiveness, quality and satisfaction levels, OP and AF. Date achieved 04/22/ /30/ /31/ /31/2010 Achieved: Phys. Perform. Studies confirmed avg. costs 30% lower than public Comments works of similar type/quality. High % SPs rated tech. satisfactory/good quality (incl. % due demand-led selection, standard tech. designs, competitive contracting, achievement) community execution. G. Ratings of Project Performance in ISRs No. Date ISR Archived DO IP Actual Disbursements (USD millions) 1 10/29/2001 Satisfactory Satisfactory /21/2002 Satisfactory Satisfactory /26/2002 Satisfactory Satisfactory /21/2003 Satisfactory Satisfactory /16/2003 Satisfactory Satisfactory /19/2003 Satisfactory Satisfactory /18/2004 Satisfactory Satisfactory /08/2004 Satisfactory Satisfactory /29/2005 Satisfactory Satisfactory /09/2006 Satisfactory Satisfactory /24/2006 Satisfactory Highly Satisfactory /28/2007 Satisfactory Highly Satisfactory /12/2007 Satisfactory Satisfactory /07/2007 Satisfactory Moderately Satisfactory /12/2008 Moderately Satisfactory Moderately Satisfactory vi

13 16 11/26/2008 Moderately Unsatisfactory Unsatisfactory /21/2008 Moderately Moderately Unsatisfactory Unsatisfactory /26/2009 Moderately Moderately Unsatisfactory Unsatisfactory /17/2009 Moderately Satisfactory Moderately Satisfactory /28/2010 Moderately Satisfactory Moderately Satisfactory /29/2010 Moderately Satisfactory Moderately Satisfactory /08/2010 Moderately Satisfactory Moderately Satisfactory H. Restructuring (if any) Restructuring Date(s) Board Approved PDO Change ISR Ratings at Restructuring DO IP Amount Disbursed at Restructuring in USD millions 06/15/2009 N MU MU /15/2010 MS MS /22/2010 N MS MS Reason for Restructuring & Key Changes Made Under Level 2 Restructuring, Closing date extended to 01/31/2010 to permit State to continue accessing Loan funds and to restructure resources to respond to Flood Emergency. Under Level 2 Restructuring, US$1.973 m. reallocated to Area 1 Subprojects, and to Incremental Operating Costs. Under Level 2 Restructuring, US$0.914 m. reallocated to Area 1 Subprojects. vii

14 I. Disbursement Profile viii

15 1. Project Context, Development Objectives and Design 1.1 Context at Appraisal This Implementation Completion Report (ICR) describes the experiences, achievements and lessons of the Rural Poverty Reduction Project in the Brazilian State of Piauí, implemented in two stages: the original project (OP - Loan 4624-BR) approved June 26, 2001 and closed January 31, 2007, and an Additional Financing (AF - Loan 7399-BR) of equal value with the same project development objectives, design and implementation arrangements, approved July 20, 2006 and closed January 31, The AF was designed to scale up project activities/investments in small-scale socio-economic infrastructure and services to enhance the impact of a successful first-stage communitydriven development (CDD) project Piauí is exceptionally poor, with 63% of total families and 77% of all rural families living in poverty at the time of project appraisal in About 39% of rural families were surviving on less than one minimum salary per month (about US$58 at that time) and 67% on less than two. Official data showed massive deficits in basic socioeconomic services in rural areas 80% of rural households lacked piped water, 77% were without sanitation, and 53% lacked electricity - compared to urban Piauí and national averages. 1 Some 60% of the State is semi-arid. Other, Bank-supported Rural Poverty Reduction Projects in the Northeast region, including Piauí, had helped to alleviate these conditions using a participatory, community demand-driven (CDD) approach but the sheer scale of rural poverty in Piauí remained challenging Government s Strategy and Bank Involvement. At the time of project preparation, the Federal Government had just launched the Projeto Alvorada framework for poverty reduction in municipalities nationwide with lowest Municipal Human Development Index (HDI-M). The Bank-supported CDD projects in the rural Northeast including Piauí - had piloted important delivery mechanisms which fit well with the Alvorada strategy of local empowerment and self-help, and the projects themselves were regarded as key components for its implementation At the State level, government had, in addition to successful experiences under the CDD rubric, partnered with the Bank in education (FUNDESCOLA II), agricultural research and development (PRODETAB), investments in water access and irrigation (Northeast Irrigation I) and land access (Land-based Poverty Alleviation I). The Northeast Competitiveness Initiative sought to expand the region s commercial agriculture by identifying and exploiting regional competitive advantages in global markets. 1 These data compare with 25%, 21% and 4% respectively in urban Piauí, and 7%, 3% and 0.9% respectively in urban Brazil. 2 Reformulated Northeast Rural Development Projects ( ), and Northeast Rural Poverty Alleviation Projects ( ). 1

16 1.2 Original Project Development Objectives (PDO) and Key Indicators (as approved) As stated in the PAD and reiterated in the Project Paper for the Additional Financing, the project sought to assist the State of Piauí to reduce high levels of rural poverty through four sub-objectives the achievement of which was to be measured by Key Performance Indicators (KPI). The table below shows the relationship between each element of the PDO and its KPIs. 2

17 Project Development Objective Over-riding PDO: To assist the State of Piauí to reduce high levels of rural poverty. Sub-objectives: 1. Improving the wellbeing and incomes of the rural poor through better access to basic social and economic infrastructure and services and support for productive activities, using proven communitydriven development (CDD) techniques 2. Increasing the social capital of rural communities to organize collectively to meet own needs; 3 3. Enhancing local governance by greater citizen participation and transparency in decision-making, through creation and strengthening of CAs and MCs Key Performance Indicators 82,000 families benefited from subproject investments; Incremental employment generated from subproject investments; Increase in wellbeing and incomes of project beneficiaries; and, Increase in social welfare of rural communities; Increase in social capital index (CPI) of project MCs; and, No. of communities graduated from the program. 4 No. of MCs participating in priority-setting and decisionmaking on resource allocation of project and non-project funded development activities; 4. Fostering closer integration of development policies, programs and projects at the local level, by assisting MCs to extend their role in seeking funding, priority-setting and decisionmaking over resource allocation. 5 Increase in total project and nonproject financing allocated through the Municipal Council mechanism Revised PDO (as approved by original approving authority) and Key Indicators, and reasons/justification 3 Social capital is defined as, a stock of knowledge, behavioral practices and attitudes held by the members of a social group, which guides the social activities in which they participate, so as to resolve a community problem identified by the community as a priority. Social capital provides citizens with appropriate reasons and motives to act collectively on behalf of a community they personally identify with. (See Van Zyl, Sonn and Costa (July 2000). 4 For purposes of this ICR, graduation is interpreted as an indicator of social capital formation. 5 Integration was defined as the project's capacity to leverage additional, complementary financial resources from other public/private programs using the project's participatory mechanisms: community associations acting within their Municipal Councils. 6 Target ratio for leveraged non-project financing was US$5.00 for every US$1.00 of project financing. 3

18 1.3.1 While the PDO remained unchanged, there were changes in some physical targets under the AF. For example, the number of beneficiary families was increased by 36,000 and numbers of subprojects by 1,200 (later reduced to 500). 1.4 Main Beneficiaries The geographic coverage was the same for both the OP and AF: 221 municipalities (excluding only metropolitan Teresina). Resources were to be concentrated in 122 municipalities with lowest Municipal Human Development Index (HDI-M). The OP intended to reach with at least one subproject - an estimated 82,000 families living mostly in remote, low density areas with scarce infrastructure and services, and deriving their main income from farming and agricultural wage labor or, on the urban periphery of municipal centers typically showing a high incidence of acute poverty. The AF targeted an incremental 36,000 families with the same profile. 1.5 Original Components Project components were as follows (see also Annex 2): Component 1: Community Subprojects (US$27.0 m., 90% of total estimated cost) financed matching grants to organized rural CAs for 1,200 small-scale infrastructure, productive and social investments under each stage (total 2,400 investments). Component 2: Institutional Development (US$1.50 m., 5% of total estimated cost) financed technical assistance and training to build capacity in implementing agencies, state modernization activities affecting poverty reduction, and further piloting of information technology for transparency and to create market linkages for productive investments. Component 3: Project Administration, Supervision, Monitoring and Evaluation (US$1.5 m., 5% of total project cost) financed the incremental costs (excluding salaries) of project administration (State Technical Unit) and coordination including supervision, monitoring and impact evaluation. 1.6 Revised Components New types of investments under the AF included education, health, environment and culture. 7 The AF narrowed its institutional focus mainly to capacity-building. Resources under the AF brought total project cost to US$60.0 m., total targeted subprojects to 2,400 and total targeted beneficiaries to 118,000 families. 1.7 Other significant changes Additional Financing: As noted above, the AF loan was designed to scale up financing, activities and beneficiary coverage to strengthen the impact of the successful 7 The Datasheet Section D (Sector and Theme Codes) shows unrealistic coding, e.g., 45% of total Bank financing was expected to finance roads and highways, unlikely under a demand-driven project with huge deficits of more vital services. Similarly the Theme Code estimate of only 33% for rural services and infrastructure was equally unrealistic. 4

19 OP while preserving its basic design, methodology and implementation arrangements. 8 Loan size was identical and the allocation of resources between components was similar. The Board approved the AF on July 18, 2006 and the new Closing Date was June 30, As noted above, the AF planned to finance an additional 1,200 subprojects for 36,000 families (180,000 people) Restructuring: The project had three Level Two Restructurings under the AF. The first, effective June 15, 2009, extended the Closing Date for seven months to end- January 2010 to permit the State continued access to AF resources and to restructure remaining AF resources to support the State s Flood Emergency relief efforts in priority project areas. 10 The second, approved by the Bank on January 15, 2010, reallocated US$1,973,750 to Category 1-A, Area 1 Subprojects (US$1,773,750), and to Category 3- A Incremental Operating Costs (US$200,000). Under the third, approved by the Bank on April 22, 2010, US$914,384 was reallocated from Category 2, Consultants Services and Training to Category 1-A, Area 1 Subprojects Extension of the Closing Date: The OP had two extensions of the Closing Date: from June 30, 2005 to June 30, 2006 (Bank approval June 1, 2005); and, June 30, 2006 to January 31, 2007 (Bank approval June 28, 2006). The first was designed essentially to recoup delays and fully disburse the Loan, taking advantage of the large number of approved subproject proposals awaiting financing. The second was a bridge to the AF which could not be approved by the Board unless the OP was still open Subproject Target: Under the AF, the target of 1,200 subprojects was reduced to 500 coinciding with the restructuring which reallocated loan funds, approved January 15, 2010 (see 1.7.1). This 42% reduction was a function of the appreciation of the Real/US Dollar exchange rate; generalized cost inflation in the AF period relative to the OP; and, larger scale of subprojects and numbers of beneficiary families per subproject. The reduction was a realistic assessment of what was likely to be fundable by closing, given these conditions Regional Offices of STU: The State decentralized project coordination, a measure recommended by the Bank for many years given the territorial spread of the CDD projects, establishing an initial office in Picos (2003), and three additional offices in Piripiri, São Raimundo Nonato and Floriano (2009). Offices were strategically located, 8 The Federal Government approved borrowing by the State of Piauí for the OP with subsequent authorization to scale up, provided project performance was satisfactory. Given strong OP progress, the State sought the scaled-up AF in the full, Federally-authorized amount of US$22.5 m. 9 The AF planned greater focus on productive investments, which tend to have fewer direct beneficiaries than socio-economic infrastructure. 10 The AF was approved by the Board in July 2006 with a Closing Date of end-june The OP had been extended 6 months - to end-january to fulfill the requirement that the OP still be under implementation at the time of Board approval. This left room to compensate for the few months cut off the maximum allowable 3-year period of the AF, which was extended 7 months to end- January 2010, permitting the full three years. 5

20 although, with the exception of Picos, all were minimally-staffed, representing a nominal presence in each region. 2. Key Factors Affecting Implementation and Outcomes 2.1 Project Preparation, Design and Quality at Entry Soundness of the Background Analysis: The analytical basis and justification rested on the documented successes and lessons of similar operations in all Northeast states, including Piauí, since Further, the Bank had financed studies, analytical work, and state economic memoranda to identify the dimensions, characteristics and causes of rural poverty and to develop strategic options for policy and programmatic actions. 11 The project was viewed as an instrument for intensifying economic activity within the viable small farm sector, a growth enhancer for the non-farm sector, and as a safety net for poor rural families in drought-prone areas with natural resource limitations. The Bank s Country Assistance Strategy (CAS BR, dated May 24, 2001, FY 01-03) identified poverty and inequality reduction as the core of Bank assistance efforts and stressed well-targeted, decentralized programs, social capital formation and local integration of programs. The AF, in turn, accorded with the CAS BR, dated November 10, 2003 (FY04-07), which called for successive projects under the Northeast CDD program to finance basic infrastructure for the rural poor, support income generation, and promote closer integration of State and Federal rural initiatives in participating municipalities. Assessment of project design: Objectives: Project objectives were rational given conditions on the ground in rural Piauí, were aligned with country and sector strategies, and remained consistent with the Borrower s rural priorities during State Government turnover in 2006 and the subsequent AF. The higher-level development objective not expected to be measured within the life of the project sought a direct impact on rural poverty through four subobjectives for which the project could reasonably be held accountable. A decade of piloting, scaling up and evaluation of the Northeast CDD projects provided reasonable assurance that the PDO was achievable, albeit quite difficult to measure and implying a well-formulated and implemented M&E plan Indicators: Key Performance Indicators (KPI) were relevant but numerous. Most lacked targets, even under the AF when this - and other data issues noted below - might have been remedied, and in most cases were stated as an increase without citing a baseline (planned for the first year). 12 Graduation was characterized as communities moved out of the program when actual design called for associations with productive 11 State Economic Memorandum - Piauí (World Bank, 2000); Rural Poverty Reduction in Brazil: Towards an Integrated Strategy (World Bank, 2003). 12 Baselines were not obligatory in 2001 (appraisal). Such study was planned for the first year of implementation, but never realized. 6

21 subprojects to be graduated from further productive matching grants. 13 Wellbeing and incomes were combined in one indicator. The PDO focused on community social capital but the related KPI called for an increase in the social capital index (SCI) of the project MCs, for which there was no baseline (and no SCI). 14 The KPI on graduation seems only loosely aligned with the PDO, but is interpreted by the ICR as a social capital factor Components and organization: Project components were few, clearlyformulated and appropriate for achieving project objectives, importantly due to their internal flexibility, permitting innovations and adjustments within a demand-led, administratively straightforward framework and methodology. The organizational structure remained simple, linking relevant specialist agencies to the STU/project via partnership agreements, avoiding the complications inherent in their direct, daily involvement in project execution and consequent need for an inter-agency coordinating body. A key design feature of the Northeast CDD projects has been the direct transfer of resources for approved subprojects to beneficiary associations and this continued. The established STU, experienced and embedded in its parent Secretariat of Planning, retained its coordination role. Implementation of components was to be decentralized, but neither the PAD (OP) nor the Project Paper (AF) mentioned decentralization of the STU itself, a need discussed with the Borrower under previous projects Strategic choices: The OP made and the AF retained, several strategic choices in project design: (a) diversification of income and employment opportunities; (b) maintaining the cost-effective, responsive community-led approach; (c) boosting the capacity of project MCs to engage in broader local planning, leveraging resources from agencies and programs outside the project; (d) more precise targeting framework delineating priority areas based on relative HDI-M; (e) rigorous monitoring and evaluation; and (f) introduction of a graduation or exit strategy for the beneficiaries of productive investments Graduation: Graduation was a new design element responding to internal Bank concerns about the use of grant financing for productive investments. The project team maintained that grants were an appropriate mechanism for jump-starting incomegenerating activities given the barriers often faced by poor, small borrowers seeking formal credit. The PAD notes prior consultations with Banco do Nordeste (BN) which 13 A community can have more than one association, and an association can be just part of a community. The association owns the subproject, not the community. An association with one grant-financed productive subproject is graduated to other financing for productive activities, but the association retains eligibility for other types of investments. 14 A Social Capital Index was developed (Van Zyl et al, 2000) to measure the evolution of community social capital under the Northeast Brazil CDD program but it was not formally updated for this project and did not, in any case, contain variables to measure evolution of the Municipal Councils per se. 15 With the exception of the Picos Regional Office, established by the new state government in 2003, the STU operated in a largely centralized manner until 2009, with STU technicians having to physically oversee hundreds of small, dispersed communities over an extensive geographic area from headquarters (see Map, end-document). 15 7

22 expressed interest in cooperating to facilitate group credit for poor communities through various bridging arrangements, but nothing came of this plan. See Annex Targeting: Targeting was further refined based on the delineation of municipalities into two cohorts based on relative HDI-M. This design was carried into the AF with an even tighter focus on the very poorest municipalities and especially vulnerable groups such as women and ethnic communities. This strategy was quite effective in practice but implementation showed that, even within such schemes, inequities can exist, requiring vigilance and re-calibration of the approach. See Annex 2. Adequacy of Government s commitment: Government s strong commitment to the PDO and methodology at appraisal was sustained through a change of government soon after effectiveness. The incoming state administration was also solidly behind AF preparation. Another electoral cycle in 2006, before AF effectiveness, kept the existing administration in power. However, there was a massive lag between the two phases due to the Federal Government s strict oversight of the Fiscal Responsibility Law, affecting the entire Bank portfolio in Brazil at the time. Further, a politically-driven change in the Project Coordinator and the STU s management team in this period negatively affected AF performance until its last year of implementation (see 2.2 and Annex 2). Risk assessment: Risks were correctly projected by the PAD based on previous experiences with similar projects in Piauí and other Northeast states. Political risk at the state and local levels was acknowledged appropriately for a project in rural Piauí - but in practice, the Bank and Borrower needed to do more than demonstrate project success to build ownership and reduce opposition. The risk management and mitigation framework of the OP was adequate and considered to remain relevant for the AF operation. The AF Project Paper mentions that the planned scale-up was not expected to create additional risk, and therefore kept the risk assessment and mitigation framework status quo. However, given that Bank supervision under the OP had consistently attempted without success to have the MCs informed annually of an indicative budget to support their annual planning, this deficiency should have been identified as an additional risk with appropriate mitigation measures for the AF to secure compliance with this project rule. 16 The reality of the project s graduation vision vs. the experience on the ground suggests that it was largely Bank-driven. Most community associations in Piauí, a desperately poor state, simply do not qualify for graduation as envisaged. A much denser level of investment is needed to make the planned transition and the project Municipal Councils have only a limited annual budget envelope. As noted elsewhere, the vast majority of associations received one subproject in the project period. 8

23 2.2 Implementation Factors Affecting Project Implementation: Mid-term Review: The OP benefited from a strong Mid-term Review (MTR) for which a quality study combining a general evaluation, survey-based physical performance study (PPS) and case studies (Civitas 2004) was prepared, 17 and a welldesigned and documented Bank review mission. The STU was alerted to important findings on targeting, training, the status and performance of the MCs and CAs, partnership formation and effectiveness, and inclusion. Some key MTR recommendations spilled over to guide the AF operational framework, e.g., strategic decentralization of public programs, additional Regional Offices, tighter targeting of benefits, institution of an indicative annual budget for the MCs, and renovation of the MIS. See Annex Delayed AF effectiveness and execution: The State s difficulties with the Fiscal Responsibility Law (which affected the Bank s entire Brazil portfolio at the time) and other issues caused protracted delays between Board approval and AF signature and effectiveness. This was followed by a hiatus of one year while newly-appointed state authorities decided what to do with the project (see 2.2.3), during which no subprojects were financed. Based on the original closing date of end-june 2009, the available implementation period at that point was about one year Extension of the closing date: The AF made dramatic progress under a 7-month extension of the closing date governed by strict conditions, and designed to support the State s emergency flood relief program in rural areas. This was the culmination of efforts since 2006 to counteract the negative effects of leadership turnover in the STU and an unjustified and disruptive reorganization of the project MCs. The experienced technical team had lost influence over the project during the AF and it drifted with minimal activity for some time. However, once STU leadership sensed that the project might fail, the technical team regained authority and the project TTL advocated a closing date extension under a strictly monitored action plan. The State Governor intervened personally, and the Bank's Country Management Unit (CMU) provided strong support. Management decided to complete the project and not cancel the undisbursed 77% of the Loan, a risk which paid off but not without some challenges. The Bank s fiduciary specialists closely monitored subproject implementation in the final months to detect any potential/actual risks. 18 This convergence of actions/events brought the AF to a successful close, fully disbursed and benefiting almost double the planned number of families Targeting: Project targeting was a source of important lessons. 19 The distribution of subprojects within municipalities showed that poverty targeting 17 Relatorio para Revisão de Meio Termo: Instituto Civitas Cidadania e Políticas Publicam/Inter-American Institute for Cooperation in Agriculture (IICA), September Much of the investment in the final months was for tractors for which there is ample information indicating cost-effectiveness and large numbers of beneficiaries, as well as preliminary indications of economic returns and job creation. 19 Under the OP, the two main priority groups were financed through the three mechanisms - PAC, FUMAC, FUMAC-P - with allocations for subprojects specified accordingly in the Loan Agreement Section IV. Under the AF, the Loan Agreement allocated subproject resources explicitly as Area 1 and 2 Grants, since all subprojects were by then decided by FUMAC Councils. 9

24 mechanisms could be affected by local/internal inequalities. Targeting was based more on the spatial distribution of demand than the pre-established poverty criteria - in practice, relative HDI-M was not a major factor in approval and financing decisions. Pressure to maintain implementation pace also led to resources flowing to the better-organized/more receptive communities and some concentration of community matching grants occurred under both phases. Finally, the requirement - virtually unique to Piauí - that the municipal mayors share counterpart funding obligations (for which they were not allocated budget) proved a handicap to the poorest municipalities, slowed project execution and further skewed targeting to better-off mayors/municipalities Intensive mobilization and organization steadily improved access of the poorest communities, the cost-sharing obligation on the mayors was lifted and successful efforts were made to establish partnerships to leverage counterpart/additional funds. Targeting overall was acceptable evaluation showed that families entering the project in more recent years were poorer than their predecessors but the project demonstrated the realities of designing and implementing an effective targeting strategy Technical and operational capacity: While the project exceeded its aggregate training targets under both the OP and AF, closer assessment of the impact of this training on the ground, e.g., organizational capacity of associations and Councils, O&M quality and the quality and sustainability of productive investments reveals weaknesses in the STU's approach to training. Aggregate expenditures for TA and training were around 52% of the aggregate allocation. 20 To leverage support for training and technical assistance services, the STU contracted a large number of private and public entities without an underlying strategy and with uneven quality control. TA and training were seen by the STU as a general operational issue and not a long-term capacity-building and sustainability factor. Training was a one-off event for many stakeholders including the MCs, mostly informational, and secondary to the physical/financial aspects Geographic factors: Piauí's geography made project monitoring and supervision difficult; the Bank had pressed the STU for many years to establish a regional presence. Cost, political reasons, resistance to decentralization of control to field offices, along with headquarters staff eager to travel long distances to supplement very low wages (travel per diems were equally modest) may explain the STU's reluctance to open such offices. While not formally planned at appraisal of either the OP or the AF, four, strategicallylocated regional offices were established: Picos with eight staff (2003), and Floriano, Sao Raimundo Nonato and Piripiri (2009) with one staff each. Responsibilities of the regional offices were not well-defined, their staffing was not project-tailored, and they seem to have been only marginally involved in project coordination Municipal Councils: The STU put a major effort into establishing and providing basic training to MCs, achieving blanket coverage. The MCs were generally wellstructured, committed to their responsibilities, and ran open, lively, participatory 20 This is partly explained by the short effective implementation period of the AF. The emphasis on and allocation for training was actually doubled for the AF but the time available to deliver these services ran out and US$2.1 million or 66% of the AF allocation under Component 2 were reallocated for subprojects. 10

25 meetings on a regular basis. But they lacked more systematic/continuous training, needed more detailed knowledge of the project s objectives, design, standards and rules, and for much of the project period, lacked financial means to ensure operational sustainability. Dependence on municipal authorities for basic operating support reduced the MCs autonomy and opened the way to external influences. 21 Also, without an indicative annual budget as a framework for planning and prioritizing investments, the MCs voted many subprojects with little chance of financing, eroding community motivation and support. This was finally resolved under the AF Demand vs. Supply: Subproject demand greatly exceeded supply under both the OP and the AF. Certain municipalities were highly effective in presenting proposals, a result of better organization, access to technical assistance, and in some cases support from political/other entities. The requirement that the project operate in 221 municipalities was unrealistic given the limited project resources. State-wide promotion of the project created expectations which could not be fulfilled. The PPS-1 and Borrower Completion Report for the OP (2008) also note the poor quality of many proposals and time expended by STU technicians to process, analyze and in many cases, completely redesign them. The registration of subproject proposals with structural/other problems caused tension and operational bottlenecks within the STU, resolved over time by training and TA for communities and the STU. Project at Risk: The AF was a problem project rated MU/U for Development Objectives and Implementation Performance in 2008 and 2009, and thus a project at risk. The rating was restored to Moderately Satisfactory in late 2009 as performance rapidly improved and disbursement accelerated. 2.3 Monitoring and Evaluation (M&E) Design, Implementation and Utilization The monitoring and evaluation plan, as detailed in Annex 2, envisaged a modern, upgraded Management Information System (MIS) and a full program of evaluation studies starting with a baseline. (a) MIS: Performance of the MIS was mixed over the period in part due to limitations on what was actually monitored, and inconsistencies in the data collected. The MIS was renovated under the AF by: (i) standardizing the nomenclature used for fieldbased (real time) data input; (ii) data formatting adjustments including the grouping of subprojects into categories with homogeneous characteristics; and (iii) upgraded capacity to permit automatic emission of IFRs as per Bank requirements. 21 The Operational Manual already covered these requirements but they were not honored until the AF phase and then not consistently: MCs were authorized to have 1% of subproject cost to support operating costs, and receive an indicative annual budget. 22 In 2004, MCs were permitted to directly and democratically elect their members/president in lieu of the mayor s office automatically assuming such leadership. This caused mayors to withdraw their operational support, virtually immobilizing many MCs but inducing over time, growth in their social capital and survival skills. Even so, many MCs remained dependent on the mayors' willingness to help. 11

26 (b) Evaluation: The project financed quality studies for the Mid-term Review (survey-based field study with controls, a Physical Performance Study (PPS) and case studies) and two follow-up PPSs using beneficiary surveys for end-op and the AF. The Bank team was also able to insert Piauí, with the States of Ceará and Rio Grande do Norte, into the quasi-experimental impact evaluation conducted by Binswanger et al. in 2004/05 (see results, Annex 5). 23 However, there was a hiatus of about five years with little evaluation activity, the result of proactive but frustrated efforts by the Bank to agree with the State on terms and conditions for a baseline study. Under the AF, the short effective implementation period preempted a full evaluation, although the STU did contract a third and final PPS to gauge the quality and sustainability of AF-financed subprojects. 2.4 Safeguard and Fiduciary Compliance Safeguards compliance: The project was a category B with an Environmental Management Plan (EMP). Subproject proposals were systematically and routinely screened for consistency with environmental provisions of the Operational Manual. The STU retained an in-house environmental specialist for the duration of the project and Bank supervision missions were vigilant regarding potential cumulative environmental effects of multiple water supply subprojects and the characteristics/practices of other subprojects with potential environmental risks. The STU had access to state and federal environmental expertise as needed. No issues arose requiring specialized resolution. The only caveat was the observation by the Civitas Mid-term Review study (2004) that community leaders and beneficiaries appeared largely unaware of environmental issues, which ties into other survey-based observations on the uneven design, quality and coverage of technical assistance and training Financial Management: Financial management performance was generally satisfactory for most of the project period with no major issues requiring special action by the Bank. In 2009 however, a Financial Management supervision mission downgraded FM management from Satisfactory to Unsatisfactory due to findings related to the outdated Operational Manual, delayed IFRs and the need to improve the STU s physical environment and internal controls. Actions taken subsequently by the STU to address related recommendations were assessed by a follow-up FM mission in April The FM rating was upgraded to Moderately Satisfactory and the risk rating kept at Moderate. At the time of ICR preparation, further efforts were ongoing, with close oversight from the FM specialist, to ensure the resolution of residual issues Audit: Audit performance was generally satisfactory with five of six audit opinions Unqualified for Financial Statements and Special Opinions. 24 Management Letters (Carta Gerencial) found the project s internal control systems satisfactory for 23 Rural Poverty Reduction in Northeast Brazil An Evaluation of Community Driven Development: Binswanger, Amazonas, Barbosa, Costa, Menezes, Pazello and Romano, World Bank The 2008 audit was carried over and combined with

27 project activities. The Bank team followed up closely with the Borrower on auditors recommendations and there were no audit issues pending at closing Procurement: Procurement performance was mixed. The Brazil CMU contracted an independent procurement review of several projects in 2006, revealing deficiencies. A follow-up mission in 2007 to assess advance procurement upon which a retroactive financing decision would be made, downgraded the rating to Unsatisfactory and increased the risk rating to High, including due to the STU s weak follow-up of the Bank s earlier Action Plan. Further supervision at end-2009 however, found major improvement and upgraded the rating to Moderately Satisfactory, with a risk rating of Moderate In the final stages of the AF, the a Bank supervision mission examined 65 contracts for the purchase of tractors whereby, to obtain better prices, reduce potential risk and save time, 65 CAs had delegated procurement responsibility to the STU. Four local suppliers were invited to submit quotations for a shopping process. As reported by the Bank procurement specialist, this centralized procurement turned out to be economical (prices were reduced by 35%), efficient (NCB could not be completed in time and there was no assurance of a closing date extension), and transparent (all concerned CAs participated in the process and signed individual contracts). This successful procurement totaled US$2,360,000, with total estimated savings of US$826,000. Two similar exercises were subsequently conducted, the second obtaining even lower unit prices. Different suppliers won each shopping and offered prices were different including more competitive in the subsequent round. 2.5 Post-completion Operation/Next Phase Transition arrangements to regular operations in the context of the Northeast CDD projects mean execution of a subproject, its formal release to the beneficiary CA - which, with some exceptions such as rural electrification systems, legally owns the investment - and its operation under pre-established rules and procedures for this phase, described in the Operational Manual. Institutional capacity of the STU is likely to be sustained given the experience and longevity of its well-qualified technical and administrative staff. Experience shows however, that political turnover and changed strategies/focus can affect the performance of even the most experienced and committed staff Operation and Maintenance: All three physical performance studies rated O&M as generally satisfactory and the sustainability of most types of investments as positive but variable. High ratings for materials and construction quality, user satisfaction and welfare impacts, along with the use of standard designs with a proven track record, bode well for users incentives to maintain their subprojects. O&M procedures were mandatory in subproject proposals and criteria for approval. O&M variability is typical of more complex productive subprojects negatively affected by lack of TA for the operational phase, but also depends on whether equipment/facilities are for: (a) Collective use and maintained by the CAs: These include water supply, tractors, animal herding and processing facilities, about 40% of all investments. In the case of tractors, the 13

28 payment of hourly or daily user fees with non-association members paying higher amounts is common, but fees paid tend to be nominal (from R$35 to R$60/hour and as high as R$80/hour) not market-referenced. The practice in other Northeast states of building reserve funds for more costly replacement parts, as opposed to basic running costs (fuel), remains uncommon in Piauí and communities mostly resort to the mayors. Water supply charges surveyed averaged a symbolic R$2.00 per household/month for the electricity to run pumps and O&M effectiveness is questionable at these rates. Processing facilities tend to charge in kind, e.g., a kilo of manioc flour or a liter of manioc gum per unit of processing time, and O&M ratings were satisfactory; (b) Family/individual use and maintenance: These include household sanitation systems, goat herds, and housing improvements, representing about 25% of all investments. The O&M of most individually-managed subprojects received high ratings except for subprojects with no prior tradition in a particular area, i.e., where communities invested in a new/unfamiliar activity and TA/training was scarce (a finding in many NE states). More rigorous screening of subproject feasibility and/or ensuring appropriate and continuous training, are essential in such cases; and (c) Facilities for collective use maintained by state concession firms: This covers electricity investments - about 35% of all subprojects - whose O&M is typically handled by the local power agency/concession firm with users paying a minimum usage fee and with generally satisfactory ratings. However, the reported frequent interruption of service in rural areas, and low power value of mono-phasic connection (the most common system financed) for productive activity, remain issues requiring attention Next Phase: The State is already mainstreaming the project mechanisms, procedures and standards into its regular rural poverty reduction programs and will continue to finance similar activities with own funds. A new rural operation stressing competitiveness and innovation, better use of existing water resources, regional development and public/private partnerships is under consideration by the State. 14

29 3. Assessment of Outcomes 3.1 Relevance of Objectives, Design and Implementation Rating: High overall relevance Project development objectives remain relevant and are consistent with the Country Partnership Strategy (CPS BR, , discussed by the Board on May 1, 2008), explicitly via the challenge of reducing endemic poverty in the Northeast region through economic inclusion which strengthens communities productive potential and activities and their economic integration. Basic infrastructure delivered costeffectively to poor rural communities is essential for productive activity, as is the social capital formation promoted by the project s participatory institutions, and by the design of the subproject cycle itself whereby communities prepare, implement, operate and maintain their own investments. The integration objective has potential to boost impact and be a key element in regional/territorial approaches to scaling-up productive activities. Project design remains relevant and appropriate to these objectives, with demonstrated flexibility to incorporate innovative new elements and experiential adjustments/improvements. 3.2 Achievement of Project Development Objectives Evaluation results show that most project objectives were achieved and that the project had many positive outcomes. Survey-based studies are summarized in 3.6 and Annex 5. Objective 1: Improving wellbeing and incomes of the rural poor through better access to basic social and economic infrastructure and services and support for productive activities, using proven community-driven development (CDD) techniques This objective achieved its wellbeing goals while results for income generation are positive but preliminary. Results reported in project studies show the following: An aggregate 1,902 community subprojects were financed benefiting 144,655 families. Of these, 79% were basic socio-economic infrastructure/services (water supply, electricity, sanitation) benefiting 124,610 families and 21% were productive (honey production, small agro-industry, communal tractors) benefiting about 20,045 families; Some 2,550 jobs were created directly, based on very conservative estimates. 25 Civitas (2004) estimated that about one-third of all investments generated employment and the average was about 1.3 jobs per subproject; Results for income and accumulation of assets were positive but statistically inconclusive. However, using the benefit-cost ratio and IRR of subprojects as 25 Not all subprojects have direct job creation capacity but they satisfy other, important basic human needs which may be even more important in the first instance than income, e.g., sanitation, water supply, health posts essential for survival, and a sense of citizenship/belonging. Further, even subprojects with an assumed high potential to generate productive activity and jobs, such as electricity connection, cannot overcome communities limited ability to respond economically due to acute resource constraints. Associations need support in transitioning to investment resources once they have basic infrastructure. The potential productive impact of electricity connection is also limited if systems installed are mostly mono-phasic (see Annex 2). 15

30 surrogate indicators (see 3.3.1) provides a degree of security that additional income is flowing to beneficiaries; Family wellbeing improved: 36,000 families gained access to electricity, 26,000 families acquired clean water supply, 8,600 families now have regular sanitation and another 26,000 families can now access a health center. Clean water supply resulted in an average 38% decline in infant mortality and 70% decrease in diarrhea. Chagas disease, asthma, hepatitis and dengue fever also declined among project beneficiaries. Targeting of benefits was sound: some 75% of project beneficiaries had income below US$1.00/day before subproject implementation and targeting/access improved over time; more recent beneficiaries tended to be even poorer. Women and Quilombola (Afro-descendent) communities accessed project benefits via the CDD methodology, exercised within a culturally and socially appropriate framework. Objective 2: Increasing the social capital of rural communities to organize collectively to meet their own needs This objective was successfully achieved: Social capital formation was strong (Binswanger et al, 2009) 26 derived mainly from the project-supported creation of over 2,400 CAs, of which 1,584 gained experience in managing and executing over 1,900 subprojects (see Annex 5); Results ranging from 83% to 98% were obtained for indicators such as growth in institutional capacity of CAs since 2002, sustainability of community subprojects, and better results at lower cost than supply-driven development; Some 91% of respondents rated highly the CAs institutional capacity to represent the community and resolve its problems, and CAs were seen as influential in community life and selected democratically; Ratings above 80% were recorded for CAs' capacity to bring subprojects that benefited the community, represented community's interests, promoted collaboration between communities and strengthened ties of friendship between residents. Importantly, CAs growing strength and role in the MCs was slowly changing relations between rural communities and State authorities, as measured by the rise in the responsiveness of the latter to community demands. Over 400 CAs graduated from further grant-financed productive investments, assumed by their MCs to possess sufficient social capital and experience to independently seek alternative financing for further, similar investments; All these achievements were made despite evidence from the MTR evaluation (Civitas 2004), subsequent physical performance studies and other sources, that deficient TA and training, uneven access to project information and the ever-present influence of political/other interests undermined to varying degrees, the institutional growth of many CAs. 26 Fieldwork for the Binswanger study was conducted in 2004 in three states. The methodology was rigorous and outcomes for social capital positive. Other studies however, suggest the need to balance these findings against a series of other factors affecting the evolution and capacity of the Municipal Councils and community associations. 16

31 Objective 3: Enhancing local governance by greater citizen participation and transparency in decision-making, through the creation and strengthening of community associations and Municipal Councils This objective was achieved: An additional 2,400 CAs (over baseline) were created, all of which were, by definition, members of the 221 participatory MCs operating at end-project, statewide. An incremental 89 MCs (over baseline) were also established. All CAs with approved subprojects received basic training in associative activity, transparent decision-making and subproject management, operation and maintenance. The MCs were overhauled, permitting the popular, democratic election of all officeholders, doing away with the custom of automatically assigning these posts to the mayors/municipal authorities. While the MCs level of maturity and capacity varied, some 25% were routinely discussing a range of public programs affecting rural life and about 11% were actually deliberating the allocation of non-project funds by end-project. In 79% of municipalities, surveys showed that organized communities were positively affecting municipal administrations, and in 81% of municipalities, establishing the MCs was seen to be having the same effect A more nuanced assessment would balance these gains against certain realities in rural Piauí, where strong party politics and vested local interests remain a fixture on the social/economic landscape. Several project analyses suggest CAs and MCs felt the weight and influence of these interests, affecting their operations and potential sustainability to varying degrees. The MCs are clearly still evolving as instruments for implementing public policy, but this does not detract from their critical role in representing poor rural communities traditionally excluded from established channels for representation. Objective 4: Fostering closer integration of development policies, programs and projects at the local level, by assisting Municipal Councils to extend their role in seeking funding, priority-setting and decision-making over resource allocation This objective was achieved but not to the degree sought at appraisal: By end-project, 100% of all subprojects were being financed through the MC mechanism. MC leadership received training on the vision and objectives of integration, supported by the STU which established formal links through partnerships with appropriate programs, facilitating communities access to their resources through the regular MC debate and decision-making processes. Some US$5.7 m. of additional, complementary funds were leveraged - well short of the ambitious 5:1 ratio sought but a good performance. 27 These funds were discussed 27 The project planned a resource leveraging ratio of US$5.00 of non-project resources derived from partnerships/integration with Federal, State and local programs, for every US$1.00 of Loan funds. Sources of actual funding were: Ministry of Social Development 17

32 and decided in the normal manner by about 25 MCs (11%) and financed 150 additional subprojects. Binswanger also found that in about 38% of municipalities, MCs were routinely discussing other federal, state and municipal programs and how they could be used to benefit MCs member communities. 3.3 Efficiency An economic and financial analysis was performed based on data and information provided by a case study of 20 productive (honey production and communal tractor), water supply and rural electrification subprojects responsible collectively for about 68% of all subprojects financed by end Results are shown below with details in Annex Cost-benefit ratios (applying a discount rate of 10%) were equal to or greater than 1.68 in four of the five subprojects analyzed. In calculating Internal Rates of Return (IRR), the base scenario (IRR-1) draws on actual data collected post-implementation from the respective CAs, assuming a time horizon of 10 years. Sensitivity analysis (IRR- 2) refines this analysis and assumes a 20% cost increase and a simultaneous 20% reduction in subproject benefits. In both cases, with the exception of household water supply, IRRs of 15% or higher are obtained. In sum, evidence suggests that financial returns are robust for honey production, tractor and electrification, even under very conservative assumptions. Table 3.3: Financial Analysis of most Commonly-Financed Subprojects Type of Subproject PVBenefits (R$) PV Costs (R$) B/C ratio IRR-1 IRR-2 1. Honey production and processing 187,129 40, % 43% 2 Farm tractor 333, , % 17% 1. Rural Electrification-1 204, , % 16% 2. Rural Electrification-2 219, , % 26% 3. Household water supply 95,944 79, % na Project IRR: The indicative financial and economic performance of the project under both phases was also analyzed (see Annex 3). The estimated, weighted IRR for the four types of subprojects is around 39%, a robust result. A sensitivity analysis charged the cost of total subprojects financed to the four selected types, i.e. the net benefit stream was calculated on the basis of all subprojects and benefits for the selected four types only, a fairly conservative assumption. The resulting weighted IRR is about 10%, similar to the discount rate. This exercise, despite its limitations, suggests the project was economically and financially viable. 3.4 Justification of Overall Outcome Rating Rating: Moderately Satisfactory (MDS, $0.8 m), Caixa Econômica Federal (CEF, US$3.75 m.), Water and Sanitation Company of Piauí (AGESPISA, US$0.63), and the National Institute for Colonization and Agrarian Reform (INCRA, US$0.49 m). 18

33 3.4.1 The overall outcome rating of Moderately Satisfactory is based on the following: Achievement of project development objectives substantiated by solid results for most Key Performance Indicators and outputs; Aggregate results meeting/exceeding aggregate targets in most cases; Preliminary indications of economic and financial efficiency; Survey-based evidence of physical quality and sustainability of most financed investments, especially basic socio-economic infrastructure; Promising integration experience, leveraging an additional US$5.7m. for investments; Generally satisfactory targeting of project benefits to the poorest cohorts; Blanket coverage of MCs state-wide, indicating strengthened governance; Original and continuing relevance of project objectives and design (factoring in the lessons of this operation) to the needs of the rural sector in Piauí This rating also takes into account the following: (i) marked reduction in the number of subprojects financed by the AF compared to appraisal estimates, although the numbers of families benefited greatly exceeded estimates due to the larger scale of certain subprojects financed; (ii) while income data from formal evaluation are inconclusive, the IRR and cost-benefit ratios for subprojects suggest positive gains; (iii) the sustainability of a large proportion of subprojects is satisfactory except for certain more complex productive subprojects; (iv) the institutional development aspect is of some concern not only due to training and TA deficiencies throughout, but because accelerated implementation in the final stages left insufficient time for institutional consolidation of the newer Councils and associations. 3.5 Overarching Themes, Other Outcomes and Impacts (a) Poverty Impacts, Gender Aspects, and Social Development Poverty impact and targeting: The potential impact on poverty is suggested by the results of the economic and financial analyses outlined in 3.3 and detailed in Annex 3, but these results are preliminary, and certain sustainability issues temper the outlook. Access to water and electricity make possible more complex productive activities - farm and non-farm for many poor families, but these do not follow automatically. Additional investment resources and good quality technical assistance are required, but difficult for the rural poor to access in Piauí. Targeting overall was acceptable: evaluation showed that families entering the project more recently were poorer than their predecessors Gender: Some 35% of the project s direct beneficiaries were to be women, but the MIS was not designed to capture such data. In practice, women s associations received 29 subprojects benefiting 976 families. Of these, 26 were infrastructure (health services investments known as Birthing Units) and three were social (open university facilities). Women were active participants in many productive subprojects but not direct owners. The project also provided opportunities for female leadership. Batista (2009) found that of 190 MCs, 26 (14%) were presided by women; in 169 MCs, there was at 19

34 least one woman on the board. In 100% of cases, there were female members on the MCs mandatory Fiscal Committees. Furthermore, 187 (18%) of the CAs established had a female president; 87% of these had women occupying the role of Treasurer and/or Secretary. Overall however, the project did not institutionalize a gender policy and the subprojects primary clientele were male heads of household Ethnic groups: The project financed eight infrastructure subprojects specifically for four associations of Quilombolas (Afro-descendent groups), benefiting 436 families. Subprojects comprised five water supply and one each of oilseed processor, cashew nut cultivation and mixed sheep/goat herding. 28 The STU sought dialogue with social movements representing such groups although this did not constitute a policy directed to them. (b) Institutional Change/Strengthening Institutional change under the project was satisfactory. Over the course of the project, the STU improved its overall efficiency, for example, by reducing the time from receipt of a community proposal to its analysis by 80% and from analysis to financing by 26%. Increasing capacity to establish partnerships to leverage specialized skills and additional resources also indicated growth, as did its management of the bulk procurement of tractors. The creation of Regional Offices however, did not have much influence on the STU s overall capacity. More influential was the State s generalized endorsement of decentralization, integration and regional planning, which are slowly improving local access to technical assistance, training and credit services with feedback effects on the STU s capacity to coordinate rural poverty programs Institutional development of the MCs was influenced by the political climate in Piauí and both the creation of new MCs and several renovations of existing MCs in the eight-year period were politically-driven to some extent. The Civitas (2004) study described MCs established with little/no reference to the Operational Manual procedures and by local contracted entities only marginally capable of effectively do this. The decision to authorize MCs to democratically elect their president/board had immediate negative effects but longer-term, the MCs, learning by doing, showed evidence of burgeoning independence and capacity, albeit variable. The final outcome for the CAs was equally mixed, with various elements/influences both promoting and constraining their growth. A period of organized consolidation is needed. See Annex 2. (c) Other Unintended Outcomes and Impacts (positive or negative) The project was a vehicle for several innovations including participation of project-financed producers in the pilot Comercio Justo (Fair Trade) program promoting sales of small-farm products - honey in the case of Piauí - to national and international 28 Piauí, unlike Pernambuco, Bahia and Maranhão, had few if any African slaves due to the culture of intensive cattle ranching which did not require large quantities of labor like sugarcane and cotton. 20

35 markets; and Salas de Parto, which are small, equipped birthing units adjacent to a basic health center specifically to ensure good conditions for normal births, a major advance in rural Piauí. Women are evaluated by a public health agent and, if difficulties arise requiring more specialized facilities, arrangements are made for their prompt transfer to a larger center. 3.6 Summary of Findings of Beneficiary Survey and/or Stakeholder Workshops The project financed three physical performance studies, a mid-term analysis with case studies and Piauí s participation in the Binswanger study. Survey-based studies are selectively summarized below and in more detail in Annex 5. There were no workshops Binswanger et al. (2009): Using a quasi-experimental methodology, surveys covered 18 treatment communities (beneficiaries) and 18 control communities (waiting groups). Findings were as follows: positive effects for the accumulation of household assets for all models estimated; marked increase in electricity access in Piauí for treatment group vs. control; significant health impacts including: 38% decrease in infant mortality and 70% reduction in incidence of diarrhea from water supply investments; all results for social capital formation were positive, reflecting the effects of creating CAs and executing subprojects; and income effects need further study, due to the relative newness of the subprojects studied Physical Performance Studies (PPS, 2004, 2009 and 2010): The following synthesizes key themes/findings of all three PPSs with separate summaries in Annex 5: high ratings for beneficiary satisfaction, subproject completion, operational status after several years, quality of construction/materials and capacity to satisfy objectives; costs of the most commonly demanded investments were consistently lower than similar investments by other agencies/programs (see Annex 2); O&M generally satisfactory but specific caveats on productive/some other investments; TA and training insufficiently strategic, and coverage/quality variable; barely one-third of subprojects under the OP were implemented directly by the CAs and few under the AF, reducing social capital benefits of managing the subproject cycle; 29 MCs and CAs are evolving but many still fragile in their management capacity; and STU subproject processing efficiency improved markedly over time. 29 Partly accounted for by the large number of electricity subprojects implemented of necessity by specialized concession firms. 21

36 4. Assessment of Risk to Development Outcome Rating: Moderate Sustainability: Factors taken into account include the following: As noted in 3.6.3, physical assessment of subprojects and surveys of beneficiaries showed high likelihood of sustainability, but too early to definitively assess longerterm sustainability of AF-financed subprojects; Generally good O&M for basic infrastructure but weaknesses in specific types of investments indicative of the need for education and consolidation of O&M practices; Economic analysis indicates generally positive financial sustainability of subprojects; Proven wellbeing impacts of basic socio-economic services provide strong incentive to maintain them; Evidence of solid social capital formation and sustainability through the creation of CAs and MCs, but need for consolidation is a consistent message; and MCs operational in all eligible municipalities, improving local governance and burgeoning tendency to discuss and/or decide the allocation of available programs Environmental management: The environmental integrity of subprojects was supported by the following: Environmental screening processes, established for all proposals, with access to environmental expertise in specific cases; Standard subproject designs for all commonly-demanded subprojects, with design features for potential environmental issues, e.g., waste disposal for manioc mills; In-house environmental professionals in the STU; Close attention to environmental compliance during Bank supervision missions; and While environmental awareness was included in training programs for associations and Councils, surveys of stakeholders cast doubt on the quality and coverage of such training. 5. Assessment of Bank and Borrower Performance 5.1 Bank Performance (a) Bank Performance in Ensuring Quality at Entry Rating: Moderately Satisfactory This rating is based on: close collaboration between Bank and Borrower during the preparation of each phase; calibration of project design to the lessons from previous, similar operations, and reasonable targets in relation to institutional capacity and experience, relevance to Borrower s and Bank s rural strategy and priorities; incorporation of innovative elements designed to strengthen impact, improve targeting, and promote integration of programs and resources; and use of the Additional Financing mechanism only recently introduced in the Bank and with few models - to scale up project coverage and impact in a very poor state. 22

37 5.1.2 These features need to be balanced by the following: inconsistent design of project Key Performance Indicators and no amendment under AF; inclusion of graduation objectives responding mostly to internal Bank concerns, not actual project/state need, and not well-prepared with the Borrower; no clear strategy under the AF for improving deficiencies or implementing recommendations of first-phase analyses; incomplete analysis of the risks facing such a project in Piauí, especially political; and unrealistic targeting of the entire state given relatively modest loan and likely demand. (b) Quality of Supervision Rating: Moderately Satisfactory This rating is based on: supervision missions based on a framework and objectives agreed with the project team and client in advance; strong Mid-term Review supported by quality analytical products; consistent attention to safeguards performance including randomized field inspection; skilled handling of critical issues affecting final stages of AF implementation, in collaboration with the CMU, Bank fiduciary specialists and the Borrower/STU; generally satisfactory outcomes for PDO and key indicators, with some caveats; proactive attention to M&E in a difficult context, pushing for Piauí s inclusion in the Binswanger multi-state study, ensuring rigorous physical performance studies and facilitating overhaul of the project MIS; and promoting excellent audit performance through fiduciary collaboration with the Borrower This assessment is balanced by the following: some 13 supervision missions (not including fiduciary) and an aggregate 60 staff weeks of supervision for an eight-year operation; missions of short duration, the majority single-staffed, and needing more specialists to cover specific issues on a repeat basis; need for more forceful Bank intervention/supervision of the AF when performance failings stemmed directly from unsatisfactory STU leadership; lack of formal fiduciary oversight by Bank specialists until 2006 (offset by consistently excellent audit results); 30 The Bank project team maintains, in the case of the AF, that there was little to supervise due to extremely delayed implementation following effectiveness. The ICR suggests that more intensive supervision may have been advisable given specific reasons for delay, as well as provisions in the Loan Agreement obligating the Borrower to ensure conditions for project implementation. 31 It should be noted that the Bank s system does not account for the time of STCs, who comprised a major portion of project supervision and the MTR. The short actual implementation period of the AF was another factor limiting supervision staff weeks. 23

38 more effort to achieve STU s compliance with provisions of the Operational Manual designed to support the Councils, i.e., indicative annual budgets and operating resources. (c) Justification of Rating for Overall Bank Performance Rating: Moderately Satisfactory The rating of Moderately Satisfactory is based on a balanced assessment of positive and weaker elements of both stages. In particular, a state like Piauí needed a more intensive approach to supervision including periodic, multi-staffed missions for greater Bank presence and specialization, and perhaps, Bank management intervention in resolving the leadership impasse causing acute AF under-performance. Further, the strength of the project's operational strategy and mechanisms was not matched by technical capacity on the ground and this needed more attention during the preparation and supervision of both phases. 5.2 Borrower Performance (a) Government Performance Rating: Moderately Satisfactory The Borrower's performance during preparation and implementation is rated Moderately Satisfactory based on the following: two State Governments supported the project as an integral part of their respective Multi-Year Development Plans and rural strategies; State Government actively pursued preparation of the AF but could not resolve the stand-off with the Federal Government over pendencias affecting its status under the Fiscal Responsibility Law of 2002 and delaying negotiations; Government's burgeoning political support for decentralization - albeit evolving slowly - was consistent with the project's objectives and operational strategy; strong efforts to resolve counterpart funding shortages through innovative partnership arrangements, but undue dependence on very poor municipalities/mayors for a large chunk of the State's counterpart funding obligation; inadequate efforts to shield the project from political influences including appointment of an unsuitable project Coordinator; and insufficient rigor regarding the creation, roles and responsibilities of the Regional Offices. (b) Implementing Agency or Agencies Performance Rating: Moderately Satisfactory This rating reflects the following: despite numerous difficulties, worked to ensure full disbursement of both loans, achieved most elements of the PDO, and achieved/exceeded most aggregate project targets; 24

39 generally satisfactory performance under the OP, followed by a sharp decline in performance under the AF, rescued by pragmatic/effective measures in the final stages; aggressive pursuit of partnerships resulting in leveraging of an additional US$5.7 m.; successful efforts to ensure blanket coverage of MCs state-wide; consistently good audit results; and overcame bureaucratic delays and sharply improved subproject processing times The following factors balance the rating decision: poor STU leadership, undermining both AF performance and an experienced STU staff; limited "vision" of technical assistance and training and inadequate attention to its quality, coverage and continuity on the ground; and Inadequate operational support to the MCs and CAs. (c) Justification of Rating for Overall Borrower Performance Rating: Moderately Satisfactory The overall Borrower performance is rated Moderately Satisfactory on balance due to a strong quantitative performance under the OP being undermined to a degree by a series of underlying "structural" issues affecting the potential sustainability of many achievements, and a very weak performance under the AF being redeemed by a more confident, pragmatic finish once leadership issues were overcome. 25

40 6. Lessons Learned The following lessons are among the more important: 6.1 CDD promotes social capital formation. The Project s demand-driven feature improves local governance by giving poor rural communities a unique set of experiences involving: collective action; priority-setting; decision-making; and subproject financial management, operation and maintenance. Social capital under this and similar projects is both a benefit in its own right and an element in the success of community-driven rural poverty reduction. Tangible benefits from subproject investments strengthen collective capacity, generate a sense of citizenship, and raise the bar for local government accountability. 6.2 The depth of poverty in Piauí, massive demand for community and municipal investment, and modest size of the Loan required a more pragmatic, coordinated approach to benefit targeting. It was unrealistic to stimulate a level of demand in 221 municipalities which the project could not satisfy, exemplified by municipalities which either did not participate, submitted numerous subproject proposals which not financed, or received only one subproject in the eight-year period. Such a project needs a more strategic use of resources, in specific regions and/or productive segments, and attention to identifying sources of non-bank resources, particularly during preparation but also during supervision. 6.3 Allocating a portion of subproject cost for technical assistance misses the fundamental problem of the scarcity and quality of such services in rural areas. Technical assistance requires intensive analysis, a well-supervised strategy and an action plan for the duration of the project, as it represents a major constraint on many types of subprojects, especially productive. The sourcing of technical assistance is as important as its financing and needs focused attention during preparation and after. 6.4 CDD projects need to invest in continuous training of participatory institutions - in this case the Councils and associations - to build sustainable local capacity and to emancipate communities from traditional relationships which perpetuate dependence. Systematic, ongoing training is needed, including specific guidance for entities advising communities on subproject preparation. As with TA, this is not just an operational issue but an investment in long-term excellence and sustainability. Training needs to extend well into the post-implementation phase when communities typically do not receive guidance on managing their venture effectively. Advance planning, progress monitoring and technical support are needed. Partnerships and "out-sourcing" can inject dynamism and leverage diverse forms of support but need quality control. 6.5 The Municipal Councils require a regular, transparent budgeting process both for capital and expenditure - for minimum operational functionality and sustainability, and annual indicative budgets for realistic planning and resource allocation. Such provisions are already included in the Operational Manuals of all participating NE States but are inconsistently applied. Councils are entitled to 2% of the 26

41 total cost of each subproject approved, to alleviate dependence on local political/other interests. Further, an annual indicative capital budget would permit a more informed analysis of proposals presented for deliberation and avoid approval of an unrealistic slate of community investments with little possibility of financing. 6.6 The preparation and analysis of small-scale, grant-financed productive ventures require rigorous standards including for technical assistance, O&M and financial needs during the operational phase. This means a business plan, technical and financial analysis, up-front marketing arrangements, and some determination of the sources of technical support. Appropriate skills for the analytical phase are critical. Successful productive experiences should be studied carefully to determine pivotal features, replicated where feasible and appropriate, and monitored over time to assess evolution and lessons. 6.7 Water and electricity investments are essential for social wellbeing and poverty reduction but they are not sufficient for productive activity, which must address the supply side, not just the demand side. Newer CDD operations such as Piauí are operating at the frontier of this mechanism, finding that the larger scale of productive investments needed to have a quantum impact on rural poverty - and the diverse inputs/arrangements needed to maximize their feasibility - may not be possible within the existing CDD framework, i.e., the public goods financed by CDD should not be expected to push more ambitious productive goals. 6.8 This project demonstrated the potentially positive impact of a well-planned and monitored extension of the closing date. Under pressing conditions, a series of pragmatic interventions advocated by the TTL and supported by senior Bank management, State leadership, and Bank fiduciary specialists, combined with a rejuvenated STU, re-activated a failing project and brought it to a generally satisfactory conclusion including full disbursement of the loan. 7. Comments on Issues Raised by Borrower/Implementing Agencies/Partners (a) Borrower/implementing agencies The Borrower s letter of June 30, 2010 (Annex 7), expressed broad consensus with the Bank s Project assessment detailed in the ICR, which also takes into account, along with other documents, the findings of the Borrower Completion Report prepared by the STU. The Borrower noted its satisfaction in making full use of the two loans (4624-BR and BR), achieving 100% disbursement and having met (and in some cases, surpassed) the PDO and intermediate targets. The Borrower concurs with the overall rating of Moderately Satisfactory, while noting the ICR s confirmation of the relevance of the project design, evidence of quality and sustainable community subprojects, breadth of Municipal Council coverage and strong benefits targeting. 27

42 The Borrower shares the Bank s observation that the scale of rural poverty in Piauí remains substantial, requiring a depth of investment that the combined US$60 m. under the OP and AF only begin to address. Finally, the Borrower reminds the Bank of the long and fruitful partnership it has maintained with the State of the Piauí, both in terms of projects finance and technical assistance via Bank supervision. In this vein, the Borrower expresses its desire to continue this partnership under a new proposed operation, with the objective of promoting economic inclusion and broader rural development statewide. For this purpose, the State has fiscal space of US$130 m. already set aside and expects to submit its proposal (Carta Consulta) to the Federal Government shortly for approval. (b) Cofinanciers, Other partners and stakeholders N/A 28

43 Annex 1. Project Costs and Financing A. Original Project (Loan 4624-BR) (a) Project Cost by Component (in USD Million equivalent) Components Appraisal Estimate (USD millions) Actual/Latest Estimate (USD millions) Percentage of Appraisal COMMUNITY SUBPROJECTS INSTITUTIONAL DEVELOPMENT PROJECT ADMIN, MONITORING AND EVALUATION Total Baseline Cost Physical Contingencies Price Contingencies Total Project Costs Front-end fee PPF Front-end fee IBRD Total Financing Required (b) Financing Source of Funds Type of Cofinancing Appraisal Estimate (USD millions) Actual/Lates t Estimate (USD millions) Percentage of Appraisal Borrower Community Associations International Bank for Reconstruction and Development

44 B. Additional Financing (Loan 7399-BR) (a) Project Cost by Component (in USD Million equivalent) Components Appraisal Estimate (USD millions) Actual/Latest Estimate (USD millions) Percentage of Appraisal COMMUNITY SUBPROJECTS INSTITUTIONAL DEVELOPMENT PROJECT ADMINISTRATION MONITORING AND EVALUATION Total Baseline Cost Physical Contingencies Price Contingencies Total Project Costs Front-end fee PPF Front-end fee IBRD Total Financing Required (b) Financing Source of Funds Type of Cofinancing Appraisal Estimate (USD millions) Actual/Lates t Estimate (USD millions) Percentage of Appraisal Borrower Community Associations International Bank for Reconstruction and Development

45 Annex 2. Outputs by Component 2.1 The following discusses the experience by component, including data on project achievements. Annex 1 shows planned versus actual total cost of each component. 2.2 Component 1: Community Subprojects The OP would provide matching grants to rural CAs to finance priority, smallscale investments of up to US$50,000. About 1,200 such investments were expected under both the OP and the AF (total 2,400), of diverse types categorized broadly as infrastructure (rural electrification, local access roads, water supply), social investments (housing improvements, school and/or health post renovation, day-care centers), and productive (community agro-processing, communal tractors, minor irrigation schemes). Subproject selection was to be community demand-driven, based on previous experience; about 70% of financing was expected to be for infrastructure, 20% for social and about 10% for productive Under the OP, priority-setting and decision-making regarding investments was conducted by three different delivery mechanisms of varying degrees of decentralization and autonomy, PAC, FUMAC and FUMAC-P. 32 PAC was the least decentralized and has been phased down/out in most NE states, while FUMAC-P was the most advanced in giving each Council its own annual budget to manage. The AF was no longer designed along these lines. The new state government elected in 2006 saw that the MCs blanketed the state and thus the strategic emphasis shifted to resource targeting based on relative HDI-M The project financed a total of 1,902 subprojects under both phases (80% of combined original target of 2,400 and 112% of the revised aggregate, taking into account the reduction of the AF target to 500 subprojects) benefiting a total 144,655 families 32 Under FUMAC, community associations participate in representative Municipal Councils with 80% majority membership of beneficiaries and civil society, and 20% of municipal authorities and other entities (percentages vary by state but beneficiaries/civil society are always the majority). The Councils meet regularly to debate community subproject proposals and set priorities, based on an indicative annual budget determined by the STU. Approved subprojects are sent to the STU for final technical and environmental analysis and approval. Funds are transferred directly to the beneficiary association. In Piauí, many pre-existing Municipal Sustainable Development Councils (CMDS) were adapted to project rules for representation (80/20 etc.) and became FUMAC Councils. FUMAC-P Councils were established as a pilot to determine the utility and feasibility of allocating a number of more advanced Councils an annual budget to manage - based on an Annual Operating Plan (POA). As with the other NE states, FUMAC-P did not function as expected and all were eventually devolved to the original FUMAC design. The main reasons were: (a) stakeholders generally agreed that the FUMAC Council served the project well; (b) NE State Governments never felt comfortable with the decentralization of actual funds management to these Councils; and (c) some FUMAC-P tended to finance very small subprojects designed to reach the maximum number of communities, resulting in low and localized impact. Under PAC, community associations submit subproject proposals directly to the State Technical Unit which transfers funds directly to the associations for approved subproject. There is no representative Municipal Council. This mechanism was a holdover from an earlier period, and utilized by communities in municipalities which had difficulty - for political/other reasons - in forming a Municipal Council, or where for whatever reason the existing FUMAC Councils was not functioning well. It is the least decentralized and studies have shown has the least impact on social capital formation. PAC is only rarely used today, in any NE state. 31

46 (123% of the aggregate estimate) - or 124,000 without repetition and some 650,000 people (about 125%). Total cost of this component was about US$28.3 m. under the OP (112% of appraisal estimate) and US$27.7 m. under the AF (115% of appraisal) including the state counterpart contribution and mayors /beneficiaries' cost share. Utilization of the Loan for Component 1 was US$20.5 m. (112% of the OP appraisal allocation) and US$20.8 m. under the AF, (115% of appraisal). Average subproject cost more than doubled under the AF, largely a function of the appreciation of the Brazilian Real and generalized cost inflation at that time. Cost also reduced the overall number of subprojects but individual subprojects were larger in many cases and benefited more families. See Annex Some 17 municipalities did not receive any subproject under the AF. STU technicians confirmed that their leadership showed little interest in presenting proposals to their respective MCs even though in some of these municipalities the MCs were very active. Communication may have been the core reason. Under the AF, the project was essentially paralyzed for many months for diverse reasons. Only in 2009 was the release of funds for approved subprojects accelerated. Press reports erroneously noted that the project was closed and would not be extended. When the extension actually proceeded, proposals already approved and awaiting financing soaked up a substantial portion of remaining loan funds, while flood emergency investments absorbed the rest Demand vs. Supply: Subproject demand greatly exceeded supply under both phases. A total 3,170 registered proposals from 2,081 associations translated into 1,385 financed subprojects under the OP and 547 subprojects from an existing 2,418 associations were financed under the AF from 1,913 proposals. Certain municipalities were highly effective in presenting proposals, a result of better organization, access to technical assistance, better quality proposals and in some cases support from political/other entities. Municipalities which demanded more subprojects generally received more. Under the OP, some 20 municipalities presented over 26 proposals each, representing 9% of total municipalities at the time and 27% of total demand The requirement that the project operate in 221 municipalities was unreasonable given the limited amount of resources and limited structural capacity of the STU. Statewide promotion created expectations which could not be fulfilled, generating frustration and eroding motivation in some areas. The PPS 1 and Borrower Completion Report for the OP (2008) also note the poor quality of many proposals despite having received initial approval and registration by the Executive Directorate - and time expended by STU technicians to process, analyze and in many cases, completely re-design them. The registration of subproject proposals with structural/other problems or which were poorlyprepared caused tension and operational bottlenecks in the project. Furthermore, some MCs reported that they were required to send all approved subprojects to the STU to avoid political difficulties with the CAs. After sending approved subprojects to the STU for final assessment, they tended to lose contact with the STU, only hearing again about a particular subproject when it was delivered to a community by a local political figure/intermediary. Supply was also affected by the need to get the mayor s counterpart contribution, a criterion for subproject approval. The fact that this commitment was not 32

47 funded under mayor s municipal budgets, subproject processing bottlenecks occurred, especially in poorer municipalities, while the mayors tried to comply and find the funds. Subproject Categories: Infrastructure: Some 1,277 infrastructure investments represented 67% of total subprojects financed under both phases. Community demand followed a well-established pattern by emphasizing rural electrification, rural water supply and housing rehabilitation which together comprised 63% of the total financed under both phases. This was consistent with other Northeast states where communities satisfy urgent needs for basic services before seeking productive investments. Average subproject cost for infrastructure was only US$23,700 under the OP but rose to US$67,400 under the AF, some 35% higher than the project limit of US$50,000 per subproject. On the other hand, average cost per family under the OP was US$407 and US$272 under the AF, indicating significantly higher number of beneficiaries per subproject. See Tables and below Productive: This category, with some 403 investments in aggregate, was 21% of total investments. Communal tractors and equipment comprised 142 or 7.5% of total productive investments under both phases and were by far the largest investment element of all types under the AF. As noted elsewhere, there was a massive procurement for tractors in the final months of the AF, handled by the STU in three block purchases, representing major cost savings over individual purchases by CAs. The sustainability of productive subprojects is considered low in many cases due to spotty training for the operational phase, and poor quality proposals and feasibility analysis. The average cost of productive investments was US$16,240 under the OP and almost trebled to US$43,610 under the AF. Average cost per family was US$305 under the OP and US$909 under the AF due to the relatively small number of beneficiaries of this category of investment Social: Demand for social investments was modest. Investments in household sanitation (septic tanks) predominated under the OP while investments in distance universities/learning surged under the AF due to the preferences of STU leadership at that time for new types of investment, although it should be noted that health and education were new types of investments contemplated at appraisal of the AF. Basic health centers classified as infrastructure but more appropriately a social investment surged under the AF. Average subproject cost of social subprojects was about US$21,000 under the OP and more than doubled to US$48,100 under the AF. Average cost per family was US$236 under the OP and US$326 under the AF. Cost Comparison by Type of Investment: Project versus Other Programs/Agencies: Consistent with conclusions reached over a decade of observation, project costs per subproject were generally consistently lower than other programs/agencies for investments of similar type and quality, as shown below. Table 2.2.1: Subproject Cost Project versus other Programs/Agencies (in R$) 33

48 SINAPI 33 INCRA Item Unit Project Market FUNASA Sec. Health Water Supply Unit 8, , System Housing M Civil construction: M Open university M , Basic health unit M , Honey prod. unit Septic Tank Unit 2, , , Electricity: -High tension triphasic Km 14,760 15,240 -Low tension triphasic Km 20,860 21,950 -High tension Unit 9,250 9,550 monophasic -Low tension Unit 11,720 13,980 monophasic -Biphasic Unit 3,107 3,567 substation -Triphasic Unit 4,903 5,182 substation Tubewell: -10 in sediment -06 in sediment -08 in rock -06 in rock Meter Cashew Unit seedlings 34 Langstroth Honey hives Targeting: Unit The project revealed the complexity of designing an effective targeting strategy. The targeting framework was established based on three priority groups categorized by relative HDI-M, and differentiated indicative allocation of project resources. At the time of the Mid-term Review (2004) MCs in the poorest municipalities were virtually moribund for lack of sufficiently organized communities to present proposals, most commonly the poorest communities. The distribution of subprojects within municipalities was also problematic, showing that poverty targeting mechanisms could be substantially affected by local inequalities. In practice, targeting was at the time based more on the spatial distribution of demand than the pre-established poverty criteria. Approval and financing patterns showed that relative IDH-M was not a major factor in such decisions. 33 National Civil Construction Index representing average of multiple programs, State of Piauí 34 For quantities above 20,000 seedlings 34

49 Further, pressure to maintain pace and implement more subprojects led to resources flowing to the better-organized/more receptive communities. Municipal concentration of subprojects was apparent under both phases (see Tables and 2.2.2) Finally, an important constraint on the access of the poorest communities/municipalities was the requirement that the municipal mayors share counterpart funding obligations. Many lacked this capacity or were delayed in producing their share which slowed project execution and further skewed targeting to better-off, fiscally more robust municipalities. Much greater focus on mobilization and organization was recommended to improve the access of the poorest communities, the cost-sharing obligation on the mayors was lifted under the AF and intensive efforts were made to establish partnerships to leverage counterpart/additional funds. Targeting overall was acceptable evaluation showed that families entering the project more recently were poorer than their predecessors - but by end-op and AF, 11 and 25 municipalities respectively, had not received an investment. 35

50 Table 2.2.2: Subproject Concentration, Municipalities, Phase I ( ) % Municipalities # Subprojects # Municipalities (Cumulative) One 23 11% >1 and % >3 and % >6 and 43 77% > 9 and º % > 12 and % > % Total 209 Note: Total municipalities with subprojects = 209, indicating that 11 municipalities in the project area had no subprojects. Table 2.2.3: Subproject Concentration, Municipalities, Phase II ( ) % # Subprojects # Municipalities Municipalities (Cumulative) One 63 34% >1 and º % >3 and º % >6 and º % > 9 and º % > 12 and º % > % TOTAL 186 Note: 25 municipalities in the project area had no subprojects financed. Graduation: As noted in the main text, graduation was a new design element arising from internal Bank concerns about using grant financing for productive activities. The Bank and STU consulted during appraisal with the Bank of Brazil and Bank of the Northeast to explore linking graduating communities to financial institutions through information sharing and coordination related to group credit candidates, participation of banks representatives at Municipal Council meetings and feedback to the STU by these banks on successful initial experiences with such borrowers. It was planned that the MIS would monitor the share of rural poor that had received grant financing for productive investments and subsequently graduated to commercial credit The reality of the project s graduation vision vs. the experience on the ground suggests that its inclusion was largely Bank-driven. Most CAs in Piauí, a desperately poor state, simply do not qualify for graduation as envisaged. A much denser level of investment is needed to make the planned transition and the project MCs have only a 36

51 limited annual budget envelope. This plan was not activated and as it turned out, few productive subprojects had the potential to move to more intensive, commercial operations. While all CAs with a productive investment were "graduated" in the sense of being ineligible to receive another such investment, this practice was essentially a normal part of the regular targeting process due to the need to ration project resources. Secondary investments designed to complement/enhance the performance of an initial productive subproject, were permitted The framework and "bridging" activities intended to transition communities to formal credit markets including MIS-based tracking of graduating communities - were not launched, being subsumed by other project concerns including those associated with government and/or administrative turnover. While there was some concentration of lending, the vast majority of associations received one subproject within the overall project period - high demand met limited resources. By end-op and AF, 11 and 25 municipalities respectively, had not received an investment. 2.3 Component 2: Institutional Development This component was designed to finance technical assistance and training sufficient for each main implementing group (MCs, CAs and STU) to acquire the capacity to carry out its responsibilities. This component also financed consultants to assist the STU in its monitoring, supervision and evaluation roles, as well as technical assistance for state modernization. The component was also intended to expand ongoing piloting of the use of information technology by CAs and MCs to increase transparency by providing real time information about the program, as well as using the internet to connect communities to markets both nationally and internationally Under this component, the OP financed 558 mobilization and training events including seminars and workshops for about 19,680 technicians, leaders, and CA representatives. Of these, 154 mobilization events brought together about 7,500 leaders to explain the OP; and 154 events were financed to train and renovate the MCs to conform to project rules due to political and institutional changes occurring every four years following state and municipal elections. Another 30 short courses benefited the 1,386 STU technicians, partner entities, community leaders and Council representatives. Finally, 37 events in the state capital, Teresina, provided guidance on the preparation and analysis of productive subprojects, and community organization and management. The State also hosted/organized a multi-state seminar on integration, partnerships and perspectives on market access for beneficiary producers, public and private agri-business entities, with participation of the Bank, the Inter-American Institute for Cooperation in Agriculture (IICA), Bank of Brazil, Caixa Econômica Federal, Bank of the Northeast, Service for Support to Small and Micro Business (SEBRAE-PI), Carrefour, and several large national supermarket chains. The project s ambitious transparency and internet marketing goals were not realized under the OP or AF This component used the same TA and training model for both phases. The STU recruited a large number of private and public entities to deliver these services but there 37

52 was no underlying strategy and quality control tended to be uneven. TA and training were seen more as operational inputs than long-term capacity-building and sustainability factors. While MIS data show that the project exceeded its aggregate training targets, closer assessment of the impact of this training on the ground in the organizational capacity of associations and Councils, in O&M quality and the quality and sustainability of productive investments reveals weaknesses in the approach to TA/training and the challenge it posed to the STU. Training and technical assistance (TA) represented an important constraint on capacity-building. Training appears to have been a one-off event for many stakeholders, more informational than capacity-building, and treated as secondary to the physical/financial aspects For the AF, some 87 training courses for all stakeholders including CAs, MCs and STU staff were planned, and achievement was 100%. However, given the numbers of CAs, the over 500 subprojects financed by the AF, and the 221 MCs in existence by then, these targets seem remarkably modest. Similarly with technical assistance, no targets were established for the Councils under the AF and technology transfer to the associations for subproject implementation fell far short of targets: 471 contracts to support the 517 subprojects actually financed by the AF As noted in the main text, the STU established four, strategically-located regional offices: Picos with eight staff was established in 2003, and Floriano, Sao Raimundo Nonato and Piripiri with one staff each, in Although pressed by the Bank for many years to establish a regional presence, the STU never felt the need. Cost, political reasons, resistance to decentralization of control to field offices, along with headquarters staff eager to travel long distances to increase their low wages (travel per diems were equally modest) probably explain the reluctance to open such offices. Responsibilities of ROs were never well-defined and their staffing may have been mostly political FUMAC Municipal Councils: The Councils represented the interests of the project and its target clientele in the municipalities and were the main focal point for project coordination and operations at the local level. From 2003, the newly-elected state government defined a new strategy for the Councils for their composition and election whereby CAs and civil society bodies began to choose their own representatives. The profile of the Councils changed markedly with a large number of leaders of unions, associations, and civil society interests elevated to leadership positions, roles formerly occupied by the municipal Mayor and his cabinet/authorities. This democratized the Councils and supported social capital formation and empowerment Renovation of the Councils permitted changes in selection criteria which had been one representative for each community association, opening space for representation by municipal regions which involved several associations. Under the old system, many communities were not represented, particularly when a municipality had many communities. Under the new arrangements, each municipal region elected its representative in a mini-plenary thereby conferring greater legitimacy by permitting all regions of the municipality to have a seat in the Council and on the other hand 38

53 diminishing the excessive interference of local managers in selecting council members representing the CAs This also had a negative collateral effect whereby many municipal mayors lost "control" of the Councils and as a result ceased to provide the institutional and logistical support which guaranteed the Councils functions/operations, rendering many of them virtually non-viable, at least for a time. However, this prompted Council members to become very resourceful in accomplishing their obligations in creative ways. It is important to note that the project Operational Manual permitted a percentage of the cost of each approved subproject to be used by the Councils to cover operating costs but most Councils under the OP were unaware of this tool and the State Technical Unit did not follow through to ensure they received it. Lack of operating funds was a major source of fragility in many Councils and limited their ability to fulfill their full slate of responsibilities. It is understood that the Additional Financing rectified this situation but whether it was universally applied across all 221 Councils is unclear FUMAC-P Councils: This mechanism had a limited life, resulting in 21 subprojects in three municipalities, based on municipal development plans. To become a FUMAC-P Council, a FUMAC Council was graduated to a stage where it could decide on the type of subprojects to be financed and the allocation of resources based on these plans and participatory work practices. FUMAC-P Councils received an actual annual budget allocation based on these plans. While the results were satisfactory physically and financially, FUMAC-P showed problems in exercising control over procurement and the submission of statements of expenditure (SOEs), as well as social capital formation. By end-project, the three FUMAC-P Councils had devolved to FUMAC PAC: There were 25 PAC municipalities at OP outset and 35 by end-op, through which some 180 subprojects were financed. PAC was not used under the Additional Financing. Growth in their numbers was not due to their merits but to the need to address problems causing the de-activation of certain FUMAC and FUMAC-P Councils, e.g., in the case of FUMAC, discord with new municipal governments, and for FUMAC-P, lack of appropriate institutional arrangements. In this manner, the FUMAC mechanism was consolidated as the best instrument In more recent years, the State Government introduced a strategy of strengthening local political life via administrative reform and introduction of standards for municipal management. While many MCs kept abreast of these developments, others fell behind due to lack of opportunity and motivation to meet, for example to discuss/approve new subprojects, and lost the initiative to engage with regionalization proposals and to participate in the State s regional development councils Modernization of the State: The OP included funds for consultant services, studies and training for state modernization, e.g., administrative efficiency, competitiveness and other elements. About US$400,000 was invested in training public sector managers, preparation of the Multi-year Development Plan (PPA), technical publications, data surveys and acquisition of computer equipment for the Department of 39

54 Socio-economic Studies of the State Secretariat of Planning. These funds were administered by IICA under its Technical Cooperation Program. Opinion verifies that albeit a small investment, its results were valuable Total cost of Component 2 was about US$1.46 m under the OP (91% of appraisal estimate) and US$1.0 m under the AF (31% of appraisal). The Loan financed 100% of this component. 2.4 Component 3: Administration, Supervision, Monitoring and Evaluation Administration and Supervision: The STU is linked to the Secretariat of Planning and until 2002 operated under the structure held over from the previous project. Electoral change in 2003 provoked an administrative reform under which the STU was brought into line with the state s new administrative model, while still maintaining relative financial and administrative autonomy. Another electoral turnover in 2007 created three managerial units: operations, administration and a separate unit within the STU for the Bank-supported land reform project (Crédito Fundiário). The STU in its final stages had a staff of 84 of which 67 professionals/technicians and 17 administrative staff, up from earlier levels due to the expected demands of the AF. The STU functioned fairly dynamically, although bureaucratically, in part due to the system of financing subprojects in installments which required agile procedures The STU was able to reduce the duration of the subproject cycle very significantly over the course of the aggregate period and from the OP to the AF. The elapsed times from initial proposal to technical review, to approval and to execution saw a marked reduction in the bureaucratic lags that had hindered implementation: down from an average 69 days (OP) to an average 15 days (AF) from the registration (entry) date to date of analysis, and an average 107 days (OP) to an average 79 days (AF) from analysis to release of financing. See Annex 1 for costs Monitoring and Evaluation: Project monitoring and evaluation (M&E) comprised first, data collection and analysis for a Management Information System (MIS), upgraded for both the original and AF projects. The system comprised subproject information, financial management, and project management modules, integrated to permit improved monitoring of the entire subproject cycle, and with real-time data entry and monitoring from the field. Inclusion of community profiles was intended to link the MIS to the evaluation cycle. The field supervision team travelled long distances to assess progress and monitor standards. Second, evaluation studies were to include annual physical performance studies, a Mid-term Review (MTR) evaluation with field surveys of beneficiary opinion and control groups, and general impact evaluation program, including a baseline study to be conducted six months after effectiveness and repeater surveys at two-year intervals. Performance was as follows MIS: Performance of the MIS was mixed over the period in part due to limitations on what was actually monitored, and inconsistencies in the data collected. For example, the system did not pick up the distribution of demand or of category, the latter 40

55 particularly prone to ambiguity. The MIS was renovated under the AF including by: standardizing the nomenclature used for field-based (real time) data input; data formatting adjustments including the grouping of subprojects into categories with homogeneous characteristics; and upgraded capacity to permit automatic emission of IFRs as per Bank requirements. (b) Evaluation: Performance was generally good although not quite as planned. Top quality studies supported the Mid-term Review (survey-based field study with controls, a Physical Performance Study and case studies) and two follow-up Physical Performance Studies conducted for end-op and the AF. The Bank team was also able to insert Piauí with the States of Ceará and Rio Grande do Norte into the quasi-experimental impact evaluation conducted by Binswanger et al. in 2005, results of which are summarized in Annex However, the State withdrew from impact studies contracted during the OP phase with the Inter-American Institute for Cooperation in Agriculture (IICA) and the Federal University of Campinas (FECAMP). Renewed negotiations on a baseline study for end-2005 delivery also fell through when the State objected to their cost and technical parameters. Under the AF phase, the surge in project activity did not occur until the final seven months thus eliminating any chance of reviving a full evaluation. However, the STU did contract a final physical performance study to gauge the quality and sustainability of subprojects financed in that short, intense period. Key Data on Project Execution: The following tables quantify key aspects of the project under both phases: 35 Rural Poverty Reduction in Northeast Brazil An Evaluation of Community Driven Development: Binswanger, Amazonas, Barbosa, Costa, Menezes, Pazello and Romano, World Bank

56 A. PHASE 1: Original Project ( ) Table 2.4.1: Subprojects Financed, by Category, Phase I ( ) Subproj. Cost Category # % Families (US$) Aver. Subproj. Cost (US$) Cost per Benef. Fam. (US$) Rural Electrification % 34,316 11,775,588 20, Rural Water Supply % 24,522 9,469,479 22, House Rehab % 1,472 5,580,448 45,004 3,791 Sanitation 80 6% 6,717 1,780,014 22, Bee-Keeping 57 4% 1, ,526 17, Sub-Total % 68,881 29,596,056 23, All Others 124 9% 13,994 1,993,562 16, TOTAL % 82,875 31,589,617 22, Table (a): Subproject Investments by Subprogram, Phase I ( ) Subprogram # subprojects % subprojects Total Invested (US$) FUMAC 1, % 24,914,420 FUMAC-P 1 0.1% 520,571 PAC % 2,840,801 TOTAL 1, % 28,275,792 Table 2.4.2(b): Municipalities and Subprojects by Subprogram, Phase I Subprogram Municipalities No. Subprojects Est. Actual Est. Actual PAC n/a 181 FUMAC n/a 1,203 FUMAC-P 7 3 n/a 1 Total: n/a 1,385 42

57 Table 2.4.3: Subprojects by Category and Type, Phase 1 ( ) Avg. SP Cost (US$) Cost per Ben. Fam. (US$) Aver. Fam. per SP Infrastructure # SP % SP # Fam. Cost (US$) Rural Electrification % 34,316 11,775,588 20, Rural Water Supply % 24,522 9,469,479 22, Housing Rehab % 1,472 5,580,448 45,004 3, Sub-Total: 1,124 97% 60,310 26,825,516 23, All Others 31 3% 6, ,322 18, Total Infrastructure: 1, % 67,287 27,412,837 23, Productive Bee-keeping 57 42% 1, ,526 17, Agro-Industry 34 25% 2, ,787 9, Small Animal 18 13% ,137 20, Tractor 13 10% 1, ,809 19, Sub-Total: % 6,369 1,927,258 15, All Others 13 10% ,433 20, Total Productive: % 7,185 2,192,692 16, Social Sanitation 80 84% 6,717 1,780,014 22, Crèche 8 8% ,333 16, Health Center 5 5% ,595 12, Sub-Total: 93 98% 8,035 1,971,942 21, All Others 2 2% ,146 6, Total Social: % 8,403 1,984,089 20, Total Subprojects: 1, % 82,875 31,589,617 22, Table 2.4.4: Subprojects, by Category and Subprogram, Phase I ( ) FUMAC- Category FUMAC P PAC Total Rural Electricity Rural Water Supply Housing Rehabilitation Bee-Keeping Sanitation Systems Sub-Total: 1, ,261 All Others Total: 1, ,385 43

58 Note: Under Phase 2, all subprojects were financed under the FUMAC Municipal Council mechanism. The three FUMAC-P Councils devolved to FUMAC, and PAC was eliminated. B. PHASE II: Additional Financing, Table 2.4.5: Subprojects, by Category, Type and Cost, PCPR-PI ( ) Infrastructure # % SP # Fam. SP Cost (US$) Avg. SP Cost (US$) Cost per Ben. Fam. (US$) Avg. Fam. per SP Basic Health Centers 49 40% 26,966 4,659,052 95, Rural Electrification 43 35% 2,023 2,375,919 55,254 1, Rural Water Supply 30 25% 1,195 1,188,085 39, Total Infrastructure: % 30,184 8,223,056 67, Productive Tractor % 8,018 5,758,741 44, Small Animal Husband % 1,312 2,294,945 45,899 1, Cashew Prod. Process % 1,868 1,733,993 43, Beekeeping 22 8% 609 1,017,700 46,259 1, Agro-processing Centers 20 7% ,247 33, Horticulture 3 1% 35 68,179 22,726 1, Extraction 2 1% 46 61,814 30,907 1, Seed Production 2 1% 74 91,611 45,806 1, Total Productive: % 12,859 11,687,231 43, Social Distance University 86 68% 16,859 4,299,408 49, Household Sanitation 41 32% 1,878 1,810,427 44, Total Social: % 18,737 6,109,835 48, Total Subprojects: % 61,780 26,020,122 50, Table 2.4.6: Combined - Phase I and Phase II, by Subproject Type and Cost Phase I ( ) Phase II ( ) Avg. Cost per Avg. SP Cost per SP Ben. Cost Ben. Fam. Avg. Fam. Cost Fam. # SP (US$) (US$) per SP # SP (US$) (US$) Avg. Fam. per SP Type Infrastructure 1,155 23, , Productive , , Social 95 20, , TOTAL: 1,385 22, , Note: Average subproject cost more than doubled in Phase II, largely a function of appreciation of the Brazilian Real and cost inflation, in turn reducing the expected number of subprojects. 44

59 Average families per subproject also nearly doubled under Phase II, due to the high number of families benefiting directly under the Distance University and Basic Health Center subprojects. 45

60 Table 2.4.7: Combined Phase I and Phase II, Selected Subproject Categories and Costs Phase I ( ) Phase II ( ) Avg. SP Cost (US$) Cost per Ben. Fam. (US$) Avg. Fam. per SP Avg. SP Cost (US$) Cost per Ben. Fam. (US$) Avg. Fam. per SP Category # SP # SP Rural Elect , ,254 1, Rural Water Supp , , Beekeeping 57 17, ,259 1, Household San , , Tractor 13 19, , Table 2.4.8: Piauí Rural Poverty Reduction Project (OP and AF): Analysis of Community Associations and Families Benefited (without repetition) Community Associations Community Associations with: OP AF # Families Total 5 subprojects subprojects subprojects , subprojects , subproject ,485 1,285 Total 1, ,856 1,584 Note: MIS data confirm there was no repetition of beneficiary families between the OP and AF. Data was taken from the MIS for the OP and AF, and CAs from each were stacked. Since the name of each community association is unique in the MIS, sorting by name yields the number of repeat associations. Matching the number of beneficiary families for each subproject (and for each community association) permits the netting out of any repetition. The analysis is conservative (i.e., biased towards an under-estimation) since the repetition factor ignores the possibility that some associations are repeating, but with different families within these associations having benefited. 46

61 Table 2.4.9: Aggregate Project Performance Targets and Results, OP and AF ( ) Project Components and Activities Unit Planned OP Actual OP Institution Responsibl e Planned AF Actual AF Community Subprojects Subprojects implemented Ben. Assns # 1,200 1,385 1, Infrastructure # 834 1,155 na Productive # na Social # na 127 Beneficiaries STU, MCs - total families # 82,000 82,875 36,000 61,780 - total people # 360, , , ,000 Community assns benefited STU, MCs # na 501 Municipalities with subprojects STU # Municipal Councils created - FUMAC STU # na 18 - FUMAC-P STU # 7 3 na na Institutional Development Prep. of annual TA/training program STU, MCs - STU # Municipal Councils # na na Community mobilization STU, MCs - Beneficiaries # seminars Local leaders # particip No. training courses offered to: - Beneficiary associations STU # All Municipal Councils FUMAC, # 154 na FUMAC-P - FUMAC Municipal Councils # na - STU staff # na Reduced to 500 subprojects coincident with the Level Two Restructuring in January

62 Technical Assistance provided: - Subproject Implementation STU, MCs Subprojects 1, , attended - to Municipal Councils STU contracts na na Project Administration, Supervision, Monitoring and Evaluation Supervision - Subprojects STU, MCs # visits 2,400 3,803 2,400 1,188 - Municipal Councils STU # visits na 212 Annual Operating Plans (POA) - Prep. POA for FUMACs MCs # 308 na na na - Consolidation, preparation of project POA STU # Standard designs - Preparation STU # 5 10 na 10 - Review and updating STU # 20 2 na 4 Special Account establishment STU Yr 1 done Information Campaign STU - presentation to Bank Yr 0 Yr 0 Yr 1 Yr 1 Yr 1 - Campaign launched Yr 1 Yr 1 Yr 2 Yr 2 Yr 2 Operational Manual STU - Preparation Yr 0 Yr 1 Yr 1 Yr 1 - Review and adjustment Yr 1 Yr 1 Yr 2 Yr 2 Monitoring Reports - Monthly disbursement summaries STU # Annual, semi-annual reports STU # External audits STU # Evaluation Studies - Phys. Performance Reviews/Mid-term STU # 1 1 Yr 2 Yr 3 and 4 - Impact Evaluation STU -- baseline # 1 na na na -- final # 1 na 37 Yr 3 na 37 The project was included in the quasi-experimental, joint evaluation of three states including Ceará and Rio Grande do Norte (Binswanger et al). 48

63 49

64 Annex 3. Economic and Financial Analysis Results and Impact of the Rural Poverty Reduction Project - State of Piauí 3.1 The Project, with an original phase and an Additional Financing phase implemented in the period , contributed decisively to improving the level of community wellbeing of the poorest rural communities, by providing access to investments in basic social and economic infrastructure, and increasingly, opening up productive opportunities capable of generating income. 3.2 As seen elsewhere in this report, the Project financed 1,902 community subprojects, benefiting about 124,000 families (without repetition), distributed in 221 municipalities, and organized in some 1,600 CAs represented in 221 MCs. Of the 1,902 community subprojects financed and implemented, two-thirds comprised the following: water supply (442), rural electrification (631), farm tractor (142) and honey production and processing (79). These four types of subproject s accounted for 68 % of the total subprojects financed, and 51% of all families benefited. 3.3 The Project was subject to a rigorous quasi-experimental study as well as a number of physical performance studies and implementation reviews whose main findings are summarized in Annex 5. As noted in this report, a final impact evaluation study was not conducted. Therefore, to complement the known results from studies summarized in Annex 5, a financial/economic analysis based on the principal infrastructure and productive subprojects financed (i.e., water supply, rural electrification, farm tractors and honey production and processing) was performed and is the main object of this annex. Methodology 3.4 A case study approach was adopted, based on the four main types of subprojects and the communities and associations demanding, executing and operating the investments financed. A small sample of subprojects (two for each type of subproject) was selected. Statistical extrapolations for the entire universe of subprojects financed were not intended, but there was a certain comfort level in exploring the results obtained for three basic reasons: (i) the apparent homogeneity (representativeness) of the units comprising the universe by type of subproject; (ii) the intentional decision to include subprojects considered by knowledgeable people to be moderately successful, that is, the sampled units were not limited exclusively to very successful subprojects; (iii) the STU took care to select subprojects considered typical ; and (iv) the communities comprising the sample were geographicallydispersed. 3.5 An STU technical team conducted the field research, interviewing beneficiaries and community association leaders, those knowledgeable about the subprojects and their histories. 50

65 Questionnaires were used, tailored to the nature of each subproject type studied, with questions permitting both open and closed responses. 3.6 Net present values (NPV), benefit/costs ratios (B/C) and Internal Rates of Return for each subproject type sampled were estimated. Cost and benefits streams were built on the basis of actual field data collected. IRRs were estimated using a 10-year period (in the case of water supply, a 15-year horizon was applied) and the NPV was calculated using a 10% discount rate. The IRR-1 was calculated using actual average values of the variables defined (see below) to measure costs and benefits. IRR-2 reflected a sensitivity analysis, by assuming that cost would be 20% higher and simultaneously, revenues/benefits would be 20% lower. All monetary values were expressed in Brazilian Reais. 3.7 Costs and benefits identified for each type of subproject were the following: Honey production and processing: Total investment (initial investment); annual operating costs; revenues derived from services provided by the processing unit controlled by the association; value of incremental production and improved quality of honey; value of incremental employment generated by honey production and processing and revenues from association membership fees. Farm tractor: Total investment (initial investment); annual operating costs; revenues derived from services provided by the tractor and associated implements and value of incremental production of corn and beans. Rural Electrification: total investment (initial investment); value of electricity consumed and paid by beneficiaries (a proxy for willingness to pay for not measured benefits such as comfort, security etc); revenues from association membership fees; revenues from taxes paid on appliances bought by beneficiaries; value of incremental employment generated by availability of electricity and value of incremental investments on houses generated by the availability of electricity. Water supply with household connection: total investment (initial investment); annual operating costs; revenue for water consumption and membership fee (proxy for willingness to pay for un-measured benefits); value of time of family members freed from having to fetch water; value of incremental employment generated by availability of water; value of reduced disease and value of incremental investments. 3.8 The results of the exercise are shown in Table 3.1. Given the rather strong limitation of sample size, although compensated in part by the manner in which the units were selected and interviewed (see 3.4 above) the results serve mainly as indicators of the magnitude of possible real value. Table 3.1: Piauí Financial Analysis of most Commonly-Financed Subprojects Type of Subproject PV of Benefits PV of Costs B/C ratio IRR-1 IRR-2 Honey production and processing , ,80 4,65 63% 43% Farm tractor , ,52 2,18 35% 17% Rural Electrification , % 22% 51

66 Water supply with household connection , ,45 0,83 3% na 3.9 The major finding is that except for water supply, evidence suggests that financial returns are robust for honey production, tractor and electrification even under very conservative assumptions (shown by the sensitivity analysis IRR-2). Honey production and processing in Piauí has proven, over the years to be viable both in economic and financial terms, with significant and increasing exports to many countries. Nevertheless, the absolute values of benefits and costs tend to be relatively small in view of the small number of families per subproject, contrary to what happens with investments in farm tractors and rural electrification. As per rural electrification, although monophasic networks dominate, the types of additional employment generated are associated with activities induced by simple electricity availability, e.g., small local food stores, bars, manioc processing units, and additional investments in housing In regard to water supply, as shown in other studies (such as Binswanger 2005), its positive impact on beneficiaries wellbeing is widely recognized, despite the inability to fully quantify this impact in this exercise. The approach here was to take payments made by beneficiaries for water consumed and member fees paid to the community association as proxies for the willingness to pay for the identified (but not measured) benefits. This was a conservative approach (i.e., bias towards under-estimation of benefits). The real matter is that a relatively high percentage of beneficiaries in Piauí do not pay membership fees or for water consumed, thereby contributing to a lower level of total benefits. Additionally, a relatively low level of incremental investments was induced by the availability of water. Looking at the Project as a whole 3.11 The next stage in this exercise was to examine the possible indicative financial and economic performance of the Project, covering both phases (OP and AF). For the purpose of this aggregate analysis, the universe considered comprised 1,733 financed subprojects, of which 1,216 in the first phase (OP) and 517 in the second phase (AF) [see Table 3.2]. Data were taken from the Project MIS. Values originally denominated in USD were converted to Reais at the exchange rate of R$1.80 per USD. Two additional assumptions were made: (i) all subprojects were financed in year zero; and (ii) a period of 10 years was considered for the estimation of the IRR. The first assumption, although strong, is not believed to invalidate the exercise in its indicative feature; to avoid its use would have implied a much longer period of analysis (2002 to 2010 plus 10 years of implementation for each subproject), separating investments (and benefits) for each year For the sake of uniformity, the same 10-year period for estimation of the IRRs was applied to all types of subprojects, including water supply, for which a 15-year period was originally used, yielding an IRR of 3% (see Table 3.1). With the reduction to 10 years, the estimated IRR became negative (-3%). The main results are presented in Table The main results are presented in Table 3.2 below: Table 3.2 Aggregate estimates of Project Economic and Financial Performance 52

67 Type of Subproject OP AF Total IRR Honey production and processing % Farm tractor % Rural Electrification % Water supply with household connection % Total % All other types Total % % of total (total -1/total -2) Given the variety of categories of subprojects, the selection of the universe was determined by the subproject categories for which field data were collected. For instance: only 565 water supply subprojects, with household connection and cisterns, were considered. The selected subprojects represent from 42.2% to 67.1% of the universe of the AF and OP subprojects, respectively (about 60% in total) 3.15 Table 3.2 shows that the estimated, weighted IRR for the four types of subprojects would be around 39% (Total -1), a very robust and comforting result. A sensitivity analysis was carried out by charging the cost of all 1,733 subprojects to the four selected types, i.e., the net benefit stream was calculated on the basis of cost of all subprojects and benefits for the selected four types only. This is a rather conservative assumption. Even so, the resulting weighted IRR is around 10%, a similar value to the discount rate. This exercise, despite its limitations, appears to suggest that the Project was economically and financially viable. 53

68 Annex 4. Bank Lending and Implementation Support/Supervision Processes (a) Task Team members Names Title Unit Lending Andrea Ryan Engineer TWUW S Andrew Parker Sr. Rural Development Economist PRMVP Anna Roumani Consultant LCSAR Claudio Mittelstadt Financial Management Specialist LCSFM Edward Bresnyan Sr. Rural Development SpecialistLCSAR Enzo de Laurentiis Lead Procurement Specialist LCES Isabella Micali-Drossos Senior Counsel LEGLA Joachim von Amsberg Country Economist LCC5C João Barbosa de Lucena Consultant LCSRE Jorge Muñoz Lead Rural Development Specialist LCSAR Klaus Deininger Land Administration Specialist DECRG Luciano Wuerzius Procurement Specialist LCSPT Luis Coirolo Team Leader LCSER Maria de Fátima Amazonas Sr. Rural Development SpecialistLCSAR Mariana Montiel Senior Counsel LEGLA Marta Molares-Halberg Senior Counsel LEGOP Morag van Praag Financial Officer LOAG3 Raimundo Caminha Consultant LCSAR Susana Amaral Financial Management Specialist LCSFM Túlio Barbosa Consultant LCSAR Supervision/ICR Anna Roumani Consultant LCSAR Claudio Mittelstaedt Consultant LCSFM Daniella Arruda Program Assistant LCC5C Edward William Bresnyan Senior Rural Development Spec. LCSAR Fabson Vogel Financial Management Specialist LCSFM Isabella Micali-Drossos Senior Counsel LEGLA Joao Barbosa-De Lucena Consultant LCSAR João Vicente Campos Financial Management Specialist LCSFM Jorge Muñoz Lead Rural Development Specialist LCSAR Jose Janeiro Financial Management Specialist LCSFM 54

69 Luis O. Coirolo Consultant SASDA Maria de Fatima Amazonas Senior Rural Development Spec. LCSAR Raimundo N. Caminha Consultant LCSAR Tulio Barbosa Consultant LCSAR 55

70 (b) Staff Time and Cost Staff Time and Cost (Bank Budget Only) Stage of Project Cycle USD Thousands No. of staff weeks (including travel and consultant costs) Lending FY FY FY FY FY FY FY FY FY FY Total: Supervision/ICR FY FY FY FY FY FY FY FY FY FY FY FY FY Total:

71 Annex 5. Beneficiary Survey Results A. Quasi-experimental Impact Study (Binswanger et al, 2009) 38 Methodology 5.1 Research covered 864 households in 108 communities and 90 municipalities, with half comprising project beneficiaries and half a group of control communities, in the Northeast States of Piauí, Ceará and Rio Grande do Norte. The sample in each state was 18 treatment and 18 control communities. The study was based on data on the current situation in 2005 (when field research was done) compared to 2002, the latter using recall, i.e., questions asked in 2005 on subjects' previous situation. 39 The control group was selected from communities which had requested and been awarded subprojects but where the subprojects had either not yet been implemented or had been implemented only a short time before the field research. The treatment group consisted of communities with subprojects approved in the first year of the original project, i.e., Various methods were used to reduce selection bias Analysis showed that the treatment groups were a little better off than the control group in the period before project execution in Level of education of household heads, access to electricity, home ownership, and total value of electrical appliances, agricultural tools and animals, were higher for the treatment group. Since the control group was selected from among then-recent project beneficiaries, it was concluded that the program, already reaching poor communities in 2002, was constantly improving its focus on poor communities over time. Communities receiving subprojects later were initially worse off in terms of infrastructure, education, access to health clinics, and certain kinds of household assets. Household assets 5.3 The study assessed the accumulation of per capita household assets and found positive effects for all models estimated, although none of the coefficients associated with the treatment group were statistically significant. Several reasons were suggested: a rise in income of families living in poverty is probably consumed rather than invested, so measuring only the impact on assets would not reflect the project's real effect on income; the sample size was relatively small for measuring small increases in capital; and, 38 Rural Poverty Reduction in Northeast Brazil - An Evaluation of Community-driven Development. Binswanger, Amazonas, Barbosa, Costa, Menezes, Pazello and Romano, World Bank, The field work for this evaluation was conducted in 2004 and The option of returning to the FECAMP (2005) treatment and control groups was rejected because FECAMP included only subprojects implemented between the end of 2003 and 2004 so that the period between implementation and impact assessment was too short to provide results on medium- and long-term impact. 57

72 measuring recall data can generate errors and wide variations. Further study was recommended. Family wellbeing 5.4 Families who obtained access to electricity (dominant in the State of Piauí) and water supply (important in State of Ceará) via the project would not have had access to these services without the project, and a greater portion of beneficiary communities had access to electricity and water than the control group Electricity access increased an average 9% in the period, and household assets rose substantially. Access to water supply rose an average 12% from among surveyed cohorts, and housing conditions improved. In communities with access to the project, a greater proportion of families had access to electricity and water compared to the control group. There was significant improvement in health: infant mortality fell by an average 38% and the incidence of diarrhea fell by 70%. Other diseases Chagas disease, asthma, hepatitis and dengue - all declined sharply among beneficiaries in the period. Social capital 5.6 Binswanger et al. interviewed municipal, community and household subjects to assess the project's social capital formation capacity, with important results: Associations: Results ranging from 83% to as high as 98% were obtained for indicators such as the growth in institutional capacity of CAs since 2002, sustainability of community subprojects, and better results than purely "supply-driven" development, at lower cost. Stakeholders interviewed pointed to the sustainability of most community subprojects and strong social impact on strengthening CAs, improving quality of life, and increasing rural family income. Levels of community participation in associations, collective decision-making and collective activities had grown. CAs (and MCs) were viewed as inclusive, participatory and democratic institutions where decisions were made by the majority. Associations' institutional capacity to represent the community and resolve its problems was rated highly, and the associations were seen as influential in community life, and selected democratically (91% of respondents). 40 At the time field work was conducted, numbers of water supply subprojects in Piauí were still relatively small but became the second largest category by end-project. It is assumed that water supply benefits reported by Binswanger can be extrapolated to Piauí. 58

73 Importantly, CAs growing strength and role in the MCs had changed relations between rural communities and State authorities as measured by the rise in municipal authorities responsiveness to community demands. Households surveyed expressed respect for their association which was linked among most subjects interviewed with a growth in trust, solidarity, cooperation and participation in collective activities. At the household level, levels of formal and genuine participation of beneficiaries in identifying local priorities and selecting the appropriate community action increased. Consistent with this, rates of satisfaction with the actual demand of respective communities and the actual investment implemented were equally high. Ratings above 80% were recorded for items such as associations capacity to bring subprojects that benefit the community, represent community s interests, promote collaboration between communities, strengthen ties of friendship between residents, and participate in the meeting to select the investment. 5.7 In summary, Binswanger et al. (2006) view the scale of measured changes in social capital as signaling a process of intensive growth in the stock of social capital and that the Northeast projects are having an important influence on this. After implementation of subprojects, social capital continues to grow but at a slower pace. The effects of setting the process in motion are durable. The projects influence is through the use of new, specific forms of institutional arrangements to help communities become more active, to engage them in inter-community collaboration networks, and to transform their relations with the State from dependency to partnership. Municipalities: 5.8 Results are as follows: In 79% of municipalities, based on interviews with members of MCs, it was felt that organization of communities had had positive effects on municipal administration, and in 81% it was believed that establishing the Councils had had the same effect. Even so, the use of the Councils as channels for providing information on the actions of the municipal administration had not grown and remained around 12%. While a significant portion of Councils have become forums for discussing public policy and other issues of community interest, most deliberation is limited to the project itself and to PRONAF. In 35%, 36% and 45% of municipalities, Councils discuss other federal, state and municipal programs, respectively. But in only 28%, 20% and 28% respectively, are Councils forums for deliberation on those programs and their resources. 59

74 The conclusion is that the Councils are still evolving as instruments for implementing public policy and public accounting. However, the Councils have a critical role in representing poor rural communities traditionally excluded from established channels for political representation. 5.9 All results indicate, with statistical significance and a high degree of confidence, that the Northeast projects have a strong impact on social capital formation, derived from the creation of associations and execution of subprojects. B. Physical Performance Studies (a) Mid-term Review Study (Instituto Civitas, 2004) 5.10 The Mid-term Review Study combined an evaluation and physical performance study (PPS) covering a randomly-selected statistical sample of 40 subprojects in 38 communities and two municipal centers, in 37 municipalities, implemented in the period of April 22, 2002 to June 30, The sample comprised 27 infrastructure, 8 social and 5 productive investments. Some 159 questionnaires were applied to 120 beneficiary families and 39 association leaders. Results Sources of project information: Responses varied widely, but only 10% of respondents had learned about the project from its information dissemination campaign, due to the lack of an appropriate institutional communications strategy. Local political leaders stepped in to fill the gap. with the result that the project s benefits were attributed to them. Subproject preparation: Few of those interviewed knew who had prepared their subproject proposal. Subproject processing time: Cumulatively, 60% of subprojects received approval within two months from presentation and 80% received approval within four months. Cumulatively, 62.5% got through the period from approval to receipt off their first installment in around two months. In terms of overall time from presentation of the proposal to receipt of the first installment, only 12.5% saw their financing about two months later and 48% saw it in around four months. Finally, the time from receipt of the first installment and subproject finalization averaged over six months. Subproject cost: About 12% of subprojects cost up to R$20,000 and 47.5% up to R$50,000. The unit cost never exceeded R$70,000. Average cost was R$45,000, well below the permitted limit of R$75,000. Average cost per family was R$912. About 65% of the sample had had subproject cost reduced at approval, mainly because the actual costs charged by implementing agencies - in the case of electricity 60

75 and water for example - were almost invariably lower than the proposal. The result was that project unit costs tended to be below market. Implementation responsibility: In 43.4% of cases, subprojects were implemented by third parties and in 32.7% of cases by the beneficiary association itself. Subproject status: Some 82.5% declared their subproject concluded and operating normally; 10% said it was concluded but had deficiencies; and 5% said it was concluded but non-operational and 2.5% said it was abandoned. In cases of deficient operation, design/analytical problems appeared to be relevant, e.g., toilets installed in a community with no water supply; for electricity, interrupted service was an issue; and for the tractor, certain components were not being used as envisaged. Subproject adequacy: Some 70% of respondents said their subproject fullyattended their needs and project objectives. Under-dimensioned and overdimensioned subprojects were often the result of initial design problems and events which occurred following implementation. Subprojects benefited on average, 67 families. About 93% said their subproject was correctly-located. Again, those which were not were attributed to poor design. Around 79% of subprojects attended association members and non-members. Physical quality of materials used rated good or excellent in 82% of cases. Less than 1% rated quality poor. Community monitoring committees: Two thirds of respondents confirmed that the required subproject monitoring committee was established and the same said such groups performed well. Community counterpart contribution: About 87% of respondents had contributed through their labor. Method of execution: About 28% said implementation/acquisition was handled by contractors/third parties; another 20% procured through three quotes from suppliers; high levels of misinformation were evident on this subject (47%). Supervision visits by the STU: A high 92% of subprojects received three supervision visits (the standard) from STU technicians. Operation and maintenance (O&M): Concession firms represented a high percentage, not surprising given that 45% of the sample were electricity subprojects. In general, a certain organizational fragility was evident in the associations, especially those where collective O&M was required. Individual O&M performed better. Apart from electricity, most of the remainder functioned without operating rules. About 80% of subprojects were either functioning without interruption or rarely interrupted, considered notable given that technical assistance delivery was considered of debatable quality. Non-functionality was almost purely technical - lack of fuel etc. 61

76 Impact of subprojects on beneficiary families: Some 91% said their subproject had improved their lives. No-one claimed their life had worsened as a result. Main reasons for positive responses were that the subproject: improved community organization; stopped families from migrating out; reduced the incidence of disease; promoted school attendance and literacy classes; and, improved family income and agricultural production. Employment and income generation: The study showed that about two-thirds of subprojects did not generate jobs but it is known that not all subprojects can do this. Some types of subprojects generate other kinds of benefits which can be more important, e.g., water supply, cisterns, health posts, septic systems and childcare centers. Electricity subprojects demonstrate the extreme limitations of poor communities in responding economically to their subprojects due to lack of financial resources - especially so in the case of electricity - or the fact that the subproject itself was inadequate to generate employment. In about 27% of cases variable levels of job creation were noted; in 10% of cases, one job was created, and in 12.6% of cases, more than three jobs were created. The average overall was 1.34 jobs per subproject. Conclusions: Indisputable evidence of successful physical implementation of subprojects and quality; Need for STU to reduce bureaucracy in subproject processing; Evidence of party-political interventions at local level; Incipient level of training for local personnel to support operation and maintenance/repairs of subprojects; Modest capacity of subprojects to generate employment and income; Some fragility in associations in part due to sub-contracting and weak quality of training programs; Some observed weaknesses in environmental awareness; Information campaigns need strengthening due widespread misinformation about project among local authorities indicative of fragile integration with the Councils and associations; (b) Physical Performance Study, Phase 1 (Fonseca and Melo, 2009) 5.11 This study covered the period from January 1, 2002 to July 23, 2008, sampling 50 subprojects in 10 of the most frequently-demanded/financed categories, representing 95% of total subprojects financed at the time, (1,368 from a total demand of 3,170). Overall OP demand was 3,170 and total financed 1,437 of which 73% infrastructure, 19% productive and 8% social. Some 314 questionnaires were applied in 14 municipalities. Main findings were as follows: Certain municipalities were much more effective in preparing and presenting proposals (example of one municipality presenting 87 proposals). Some 20 municipalities presented more than 26 proposals. Some 9% of total municipalities at 62

77 the time accounted for 27% of total demand. Those which requested more tended to get more. Study notes that the years 2002, 2004 and 2006 were years of most intense activity, i.e. elections in 2002, new State Government revved up the project in 2004, another election in 2006 followed by a massive slide in 2007 due to lack of interest of the new STU Coordinator. The 120 municipalities below HDI-M of 0.589, supposed to absorb 70% of subproject financing, actually received 58% of value and 49% of subprojects. Demand and approvals: Type of subproject % of Demand % of Approvals % Not Financed Infrastructure 74% 82% 65% Productive 18% 10% 27% Social 8% 8% 8% 85% of all subprojects financed were completed and operational. 10% were still under construction, and 5% abandoned or concluded by not operating. Water supply: 67% operating normally and 13% with deficiencies. Some 13% were non-operational mainly because associations lacked administrative capacity to operate the system, they were poorly-organized, and/or were not paying user fees. Submerged pumps burn out and difficult to repair. Associations do not understand why they should do O&M, indicating inadequate training and TA. Communities unable to understand why O&M was not "handled by AGESPISA as in the cities". Electricity: 100% concluded and 95% operating normally but 5% experiencing frequent interruptions to service. Beneficiaries pay R$30-30/month. Damaging oscillations in supply affect machinery but many impacts: public lighting; schools and health posts; conservation and storage of medicines and food; night school and literacy classes. Manioc mills: O&M poor but sample not big and not considered representative. Standard designs: All subprojects took into account the STU s standard designs. In 76% of cases, subprojects met their objectives, 16% were regular, i.e., mostly subsized or located in the wrong place. Both water and electricity had variable results for adequacy of design to objectives but overall, very Satisfactory. Unlike the AF, about 30% of all subprojects were implemented directly by the beneficiary associations in part due to their simplicity, e.g., honey, castor bean 63

78 crushing, goats, tractors and septic tanks. Such subprojects good for the local economy because associations tend to shop locally and give payment in labor. They also tended to have better quality materials. There was impressive resource mobilization by subprojects implemented directly by associations, even though over a period of 6 years. An estimated 26.4% of project resources (OP) were injected directly into local municipal economies. Extrapolating from the PPS sample to the total 1,437 subprojects under the OP, an estimated R$23.7 m. were mobilized. Similar situation noted in 66% of associations using firms, which also purchased a large amount of materials, labor, equipment and inputs. In about 50% of cases, the firms purchased locally. Monitoring committees, required under the Operational Manual, were invariably present but the "balance of power" was uneven. Their lesser stature compared to the executing firms and the STU limited to some extent their performing their full oversight and enforcement role. Their effectiveness was quite limited. They acted mostly as facilitators and at times, provided services and acted as agents for the contractors. Surprising data on the education level of beneficiaries: 80% of heads of household with education higher than primary school; 44% with primary level complete; 30% with secondary complete; 6% with tertiary complete; 66% with half to one minimum salary; andb20% with 1-2 minimum salaries. 84% of residents of project-eligible locales were formal members of associations, on average for 7.5 years. Survey showed that being a member or not of an association represented neither an impediment nor guaranteed access to the project. Technical assistance experience was mixed. All associations received some form of guidance or TA at various stages of a subproject, supplied by the STU teams plus an army of private suppliers, NGOs, unions and others with some experience supporting rural communities. But many interviewees had never received TA or knew nothing about it. Even so, satisfaction levels with TA were quite high. Uniformly lower unit costs/prices for the project (housing, cisterns, septic tanks, electricity, tube wells, simple water supply systems) compared to market, FUNASA, Secretariat of Health, Civil Defense and INCRA. Satisfaction levels were high: 81% said the project fully satisfied their needs and 6% said it did not. Sustainability appeared good: 84% were concluded and operational and 74% showed excellent/good design versus objectives. 74% used excellent or good quality materials. 64

79 Beneficiary cost sharing was quite high in one form or another, the highest percentage being in labor. Operation and maintenance of collective subprojects was problematic: fees were collected only for basic operations not more expensive longer-term repairs and replacements. O&M by individuals and families was more stable due in part to strong sense of ownership. Effects and impacts: Results/outcomes generally successful but in some cases there are issues. The survey did not see any new productive activity resulting from these investments. (b) Physical Performance Study (complementary), Phase 2 (Fonseca and Melo, 2010) 5.12 The PPS for the Additional Financing used a similar format, methodology and sampling to its predecessor, was intended to complement it, and covered 23 of the 517 subprojects implemented from July 30, 2008 to December 31, Main findings were as follows: Processing agility had further improved and bureaucracy reduced. The overall period from receipt and registration of the subproject proposal to release of financing was reduced to an average 92 days. Average of about 4 months from release of funds to subproject completion. Satisfaction levels were high but this was hard to assess because 30% were still under completion mostly the health posts and open-university investments. Some 70% were concluded and fully-operational. The study notes the three large procurement Shopping events and reiterates the Bank s finding of economic efficiency. Construction materials were generally of superior quality and quality of the final product Satisfactory. Cost-effectiveness: Study reiterates a longstanding finding that the CDD methodology promotes cost-effective socio-economic infrastructure, compared to similar projects executed under other agencies/programs. Unit costs universally cheaper than the market. See Annex 2. Capacity of a subproject to attend to a community s needs and to its stated objectives was rated well, bearing in mind the 30% incomplete. Among the insufficient were septic tanks where there were not enough for the whole community, and electricity which, being mostly mono-phasic, restricted its use for productive 65

80 activity. In the case of tractors about 33% were non-operational because only recently-acquired and beneficiaries had already prepared their fields. Beneficiary implementation: An important finding was that none of the AF investments were implemented by the beneficiary association compared to 30% under the OP - despite the program s long-standing characteristic throughout the NE of fostering implementation involvement to promote social capital formation. Most associations consulted three firms based on a master list prepared by the STU. Contractor purchasing however, tended to be outside the associations area whereas with the latter as executors, most purchasing was done locally with economic benefits for the immediate locale. Labor however, tended to be sources locally. The study does not fault this practice seeing it as economically rational, especially so in the case of the three successful Shopping events for tractors conducted by the STU with the blessing of the associations and the Bank procurement specialist. Monitoring committees, required under Brazilian law and the project Operational Manual, were uncommon although some form of critical oversight was present in a high proportion of cases. Operation and maintenance was excellent or good in 87% of cases of subprojects for collective use by association members but problems with the collection of inadequate user fees for more complex repairs/replacements, and associations administrative difficulties complicated the efficiency and effectiveness of O&M. For individual family subprojects, 66% were rated good and similar ratings were registered for investments operated by public concessions firms, with some caveats. Sustainability even though most AF subprojects were very recent suggests the following: no subproject was abandoned; vast majority adequate for objectives (location, serving targeted families, high levels of satisfaction, quality materials). But some absurd situations indicating poor analysis and technical assistance. Social capital: However, apparent erosion under the AF of the project principle of decentralized application of resources and gradual increase in inclusion of beneficiaries as active implementation agents. No single event counters project rules but beneficiaries lost ground/opportunities to execute even simple subprojects, notwithstanding the economic efficiency of block purchasing which was a speed issue as much as anything in a grossly-delayed project. The social capital benefits of collaborative action, financial control and learning, cooperation, left space for a re-emergence of assistencialismo and erosion of the sense of ownership of the goods acquired by not having managed the subproject cycle. Questions regarding tractors: their suitability for small farmers cultivating three ha/year, many of which with animal traction; questionable technical assistance in 66

81 regard to the agronomic wisdom of a tractor; and environmental effects on sandy, semi-arid soils and caatinga (scrubland) areas. 67

82 Annex 6. Stakeholder Workshop Report and Results N/A 68

83 Annex 7. Summary of Borrower s ICR and Comments on Draft ICR A. Summary of Borrower ICR The following is the Executive Summary from the Borrower Completion Report. EXECUTIVE SUMMARY (Informal Translation) 7.1 The picture of rural poverty in the State of Piauí, in its actual details, began to take shape in the 1970s when the growing masses of rural residents were not included in the successes of the Brazilian Miracle, thus emerging as a social problem. 7.2 From the second half of the 1970s onwards, the State Government and the World Bank established a partnership envisaging a reduction in the levels of rural poverty as an important part of efforts to change social conditions in Piauí. This started in this period with the execution of the Northeast Integrated Rural Development Program - POLONORDESTE - and continued into the 1980s, with the execution of the Vale do Parnaíba Program and the Program of Support to Small Rural Producers collectively and widely known as the PAPP. At the start of the 1990s, based on experiences acquired under these projects, the World Bank and State Government decided to create a new rural poverty reduction concept notable for adopting more participatory methodologies and instruments, and decentralized activities in municipalities and communities. Thus the generation of the PCPRs Rural Poverty Reduction Projects was launched, the first of which was executed from 1996 to 2000 with Bank support. 7.3 With minor variations and successive improvements, these projects financed basic infrastructure for rural communities providing stimuli and the means of production for small-scale family agriculture and community activities. 7.4 In addition, associative bodies were established and consolidated, and encouraged to adopt progressively, mechanisms and participatory opportunities designed to include rural residents in the full cycle of subproject implementation. 7.5 Along with other projects/programs provided by various governmental agencies, the PCPRs contributed positively to improving the living conditions of the population of Piauí, as the State Municipal Human Development Index (HDI-M) rose from in 1991 to in Despite these advances, in 2001 World Bank studies showed that in a population of 2.7 million inhabitants over 63% remained in conditions of absolute poverty, in 69

84 virtually all municipalities of Piauí 41. Also, it was noted that 7.3 of every 10 families residing in rural areas survived below the poverty line Further, given that resources under previous PCPRs were insufficient to meet persistent demand, the State Government proposed a new loan to the World Bank. 7.8 The proposal was approved in 2001 under the name PCPR II with a budget of around US$60.0 m., US$45.0 m. from a Loan Agreement with the World Bank, US$9.0 m. of State Government counterpart, and US$6.0 m. in contribution from the beneficiary communities in the form of labor and local materials. 7.9 The Project was executed in two phases covering the period from April 2002 to January 31, 2010, with World Bank resources totaling US$45.0 m. - US$22.5 m. for each phase. The first, called PCPR II - Phase 1 (Loan 4624-BR), executed from April 22, 2002 to January ; the second, an Additional Financing (Loan 7399-BR), continued on with activities under execution in phase 1, and closed on January 31, 2010 having fully expended the Loan The structure of PCPR II comprised three components to which were allocated funds and activities consistent with project objectives, as follows: Community Subprojects 43, Institutional Development and Administration, Supervision, Monitoring and Evaluation. Community subprojects absorbed about 90% of project resources to attend directly to Community demands The Project s target population were the poorest rural communities in all 222 State Municipalities (except the Capital municipality) with population of up to 7,500 inhabitants, including municipal centers, and especially those with the lowest Municipal Human Development Index (HDI-M) Beneficiaries were small farmers, salaried rural workers, settlers, renters, female heads of household and some Quilombola communities, residents of organized communities represented by a community association. The idea, an innovative characteristic, was that the PCPR II provided associations with the opportunity to participate in identifying their needs and formulating demands, and implementing and managing their investment subprojects, using a Municipal Council as their approval mechanism. 44. The role of the State Government, through its executive Technical Unit tasked with implementing the PCPR II, 45 concentrated on processing, 41 - Defined as the population with average income below US$1 per day/per capita Expression defined as the population with income below US$2 per day/per capita. 43 Initially, the design of this component contained three divisions (PAC, FUMAC and FUMAC-P), with progressive levels of decentralization. In 2003 it was decided to discontinue PAC and FUMAC-P keeping only FUMAC with the denomination of "community subprojects" because it was the best mechanism for decentralization. 44 Created under PCPR-1, they were responsible for the analysis and prioritization of community demand; under PCPR-II they were trained to promote co-participation with local public bodies in the sense of complementary resources, as well as to participate actively in supervising the execution of investment subprojects. Many Councils received guidance and small training about Project rules and all received financial help to cover their operational costs (transport, food for council members etc), corresponding to 1% of the value of each subproject financed by the respective municipality. 45 Linked to the State Secretariat of Planning with a central office in Teresina and five Regional Offices (Picos, Piripiri, Floriano, S. Raimundo Nonato and Bom Jesus). 70

85 granting of financing and enforcing the rules of the game, leaving to the beneficiaries the responsibility for managing the resulting resources and goods The resources for financing community subprojects were provided in the form of matching grants up to a limit of US$50,000, leaving to the beneficiary association the obligation of contributing at least 10% of total subproject cost, which could be supplied in the form of labor and/or materials, facilitating community participation. In this way, the project provided financing for small-scale investments in basic economic, social and productive infrastructure designed to improve health, education and wellbeing of poor population and their opportunities for employment and income To ensure the means and conditions for Project implementation and management, resources and effort were also invested in institutional development: technical assistance, consultancies and training for CAs, MCs and Project management, as well as the modernization and operational capacity of the Technical Unit It is worth noting the 330 (579% of target) in direct training events for CAs, 247 training events for MCs, 54 event for STU staff and 4,971 supervision visits made by the STU s technical teams to the subprojects to verify progress and quality Also notable were the linkages to public and private entities through cooperative mechanisms and partnerships, from the involvement sought from municipal mayors - giving logistical support to associations and the Council members - to the Technical Cooperation Agreement with IICA - through which studies, training events and specialist consultancies were arranged The operationalization of the PCPR II in general, was conducted in conformity with the rules and procedures in the Loan Agreement and Loan Agreement of the Additional Financing, as well as the contents of the Aide Memoires of Bank supervision missions, and of the conclusions of the Physical Performance Studies. In almost all cases, subproject implementation complied with the following steps in the Operational Manual, that is: 7.18 Selection of Subprojects Demanded: done by the CAs, in assemblies/meetings, through discussion of their real needs and definition of the top priorities, by vote registered in minutes of proceedings; these agreed demands were then brought to the project in the form of a Carta Consulta - a request for financing, with basic information about the intended investment, about the proposing entity and the socio-economic conditions of the demanding community Priority-setting in the Municipal Council: Before sending these Cartas Consultas of the many associations to the STU, they were sent to the Municipal Council passing through a full discussion, analysis and prioritization. Deliberative forums for the PCPR projects, the Councils had at least 80% of their members from the beneficiary associations Request for Financing: Once prioritized by the Councils, these Carta Consultas were registered by the STU Appraisal of Demand: An activity conducted by the STU s technical team, visiting the community to evaluate the relevance of the request and whether it fulfilled the requirements for eligibility (e.g., legitimacy, environmental impact, operational capacity of the association to manage the subproject) Subproject Preparation: This commenced as soon as the viability of the demand was confirmed. Standard subproject designs were provided by the STU or their services of third parties were contracted, to prepare the specific technical subproject Analysis and Decision re Subproject: Once received, subprojects are registered in the project management system and submitted to an obligatory analysis of its institutional, technical, 71

86 financial and environmental feasibility by STU technicians, following which a decision to proceed or be re-done, is emitted Signing of the Contract: This follows standard models, following which the resources are transferred to the associations Subproject Execution: The acquisition of goods and contracting of services or works needed for the subproject were entirely under the responsibility of the beneficiary associations - following procurement and contracting procedures established in the Operational Manual - via direct administration or the contracting of specialized firms Supervision: This has the purpose of ensuring that the execution of works, procurement and use of the resources are in accordance with project rules and attend the interests of the beneficiary communities. In addition, the STU technician, Supervision Committee and Association President attest to/sign off that the agreement (re the subproject) has been complied with and certifying that the project is concluded and can be formally delivered to the community Data contained in the processes of approval and release of subprojects demonstrated that all subprojects received assistance and monitoring by Technical Unit teams and which were visited at least twice during subproject execution. On the other hand, the World Bank, throughout both phases, performed 13 supervision missions of the project including field visits and the preparation of supervision reports (Aide Memoires) which attested to the same extent on the Project s performance Given the execution period of the project, the numbers show that the results obtained exceeded, proportionally, the targets envisaged at appraisal In terms of physical performance, the PCPR II had a demand of 5,085 proposals for subprojects of which 1,909 were approved and financed; their distribution by type was 1,282 infrastructure, 405 productive and 222 social. Compared to the Project Appraisal Document - PAD these numbers correspond to 91%, 93%, 115% and 60%, respectively, demonstrating the operational capacity of the project In addition, with the achievement of these targets, the PCPR II ought to have provided and impressive supply of goods and services to enormous contingents of families in the rural zone, such as: access to energy for families; clean water supply to families; septic tanks for 9,056 families; community tractors for 5,668 families; open university for 5,674 families; birthing rooms and basic health units benefiting 2,430 families; small-scale agroindustries and equipment to benefit local agricultural products for 3,981 families; and equipment for the breeding/raising of small animals by 5,874 families; donation of seedlings, seeds and equipment for agricultural crops benefiting 2,156 families, urban infrastructure for collective use by 4,686 families and improved housing for 1,918 families Overall, 108,212 families were benefited (99% of the target) or 528,819 people (107% of target), resident in 1,095 different localities within 222 municipalities; in regard to "special groups", the PCPR II benefited about 1,412 families, or 6,408 people through the financing of 37 subprojects The results and material goods derived from the PCPR II are certainly provoking positive impacts on health, education, housing, employment, income, transport and in the exercise of citizenship The project also mobilized 3,419 CAs in the formulation and delivery of proposal letters (Carta Consulta); within these, some 1,514 associations (congregating some 59,041 associates) obtained financing - a number corresponding to 112% of the target. In regard to CAs of "special 72

87 groups" such as women and Quilombolas, project benefits arrived for 32 entities benefiting 1,410 association members The total value of investments was R$124,092,183.82, with the average of R$65, per subproject The positive outcomes attest to the full success of PCPR II which translates into lessons for future projects, as follows: 73

88 Lessons Learned: To consider as a criterion for subproject approval, the actual availability of beneficiaries to support the costs of operation and maintenance through the payment of user fees; As an economic factor the significance of electric energy is found in the reproduction of the labor force; but the mono-phasic energy provided has been shown to be insufficient as a factor of production. Public lighting of villages, supply to schools and health posts and mainly, household lighting (conditions of comfort, access to ample and varied consumer markets, refrigeration and television with parabolic antenna) have provoked promising effects; however, to combat poverty, energy has to be used as a propulsion force. Technically, the MIS has become an ideal instrument for its ends/purposes; however, there are problems with its uploading, limiting it to its use as a tool for supervision, control and planning. To be consolidated, the MIS must incorporate the question of "empowerment" of technical teams, since the updating of the data has to be done by the executors of activities which feed into the system. The establishing of inter-institutional partnerships to execute the PCPR II was a promising initiative; but the experiences suggest the need for certain adjustments, especially in regard to the internalization (by the partners) of the ideas of democracy, citizenship, autonomy and the creation of conditions for their progressive exercise within the associations and community councils. In regard to the permanence of the investments implemented, we reiterate the particular situation of those who which are operated and maintained by the CAs. These activities have been the greatest challenge in the mission assumed by the associations, communities and users of water supply systems, manioc mills and communal agricultural tractors, since they replicate the problematic subprojects. The recurring causes involve the insufficient resources available to cover operational costs and also, the uncertainty in regard to resources necessary for maintenance and repairs. In most cases, these services are costed as "caixinhas and other contributions, at the same time in which the payment of tariffs restrict one to small cases and with values well below what is needed. Presented with this complexity, which also involved the immense majority of similar rural communities, suggests the search for other forms of management which guarantees functionality of the goods and services of good quality, at fair prices and adequate to the socio-economic conditions of users, with the participation of the community guaranteeing their sustainability. 74

89 B. Comments on the ICR (a) Borrower/implementing agencies The letter below was provided by the Borrower following its review of the Bank s draft ICR: GOVERNO DO ESTADO DO PIAUÍ SECRETARIA DO PLANEJAMENTO COORDENADORIA DE COMBATE À POBREZA RURAL - CCPR PROJETO DE COMBATE À POBREZA RURAL PCPR II FASE II OF.Nº 312/2010 Teresina, 30 de junho de 2010 Ilmo. Senhor Edward Bresnyan Gerente do Projeto PCPR II FASE II - PI no BIRD Brasília DF. Prezado Senhor, Ao tempo em que cumprimentamos Vossa Senhoria, reportamo-nos à versão preliminar, que tivemos a satisfação de receber, do Relatório Final de Implementação (ICR- Implementation Completion Report) do Projeto de Redução da Pobreza Rural do Piauí, encerrado no dia 31 de janeiro de O referido projeto contou com o inestimável apoio financeiro do Banco Mundial, através da concessão de dois empréstimos e na permanente atenção dessa instituição em ações de acompanhamento e de assessoramento técnico. Cumpre-nos também manifestar a nossa satisfação em ter conseguido aplicar a totalidade dos recursos recebidos, no montante de US$ 45,0 milhões, e de ter conseguido atingir, e em alguns casos até mesmo superar, as metas pactuadas, conforme constatado no citado ICR. Em termos gerais, a opinião do Governo do Estado, através da análise procedida pela Unidade Técnica do PCPR, é a de que o conteúdo do ICR reflete com aceitável exatidão os principais eventos que marcaram a execução deste Projeto durante os quase dez anos de sua execução. As informações referentes à consecução das metas, á aplicação dos recursos, aos possíveis impactos obtidos, e até mesmo aos problemas enfrentados, corroboram perfeitamente com as informações similares apresentadas no Relatório Final do Estado, preparado pela Unidade Técnica do PCPR, cujo Resumo Executivo é parte integrante do ICR. 75

90 Deste modo, uma vez ratificado, em substância, o ICR, a análise detalhada feita pela Unidade Técnica se concentrou em extrair deste documento lições e experiências que sirvam como subsídios para se empreender novos projetos com vistas ampliar e consolidar os impactos dos novos investimentos sobre as condições de vida das populações pobres. Neste sentido, após atenta leitura do documento, tomamos a liberdade de tecer alguns comentários sobre os tópicos mais relevantes do ICR conforme a seguir: 1. Com respeito à avaliação geral do Projeto, a nota dada no ICR como sendo Moderadamente Satisfatória, (item 3.4, pag. 16) nos causou, num primeiro momento, certa surpresa, tendo em vista o fato de que as metas físicas e financeiras foram cumpridas. O próprio ICR reconhece o comprometimento mantido pelo Governo para prover os recursos de contrapartida e resolver os eventuais problemas operacionais surgidos na execução. No entanto, ao verificarmos as justificativas apresentadas no Relatório do Banco para a referida avaliação, constatamos que o conceito de Moderadamente Satisfatório tomou como base resultados considerados bastante auspiciosos para os executores, conforte descritos no ICR (pag. 23 e 24), que reproduzimos a seguir: Relevância do desenho e dos objetivos do Projeto, mantidos nos dois financiamentos, para as necessidades do setor rural no Piauí; Resultados geralmente positivos, para o alcance dos objetivos de desenvolvimento, consubstanciados pelos Indicadores de Desempenho, avaliação quase-experimental e outras análises do projeto; Evidências, baseadas em levantamentos de campo, da qualidade e sustentabilidade da maioria dos investimentos financiados, principalmente os de infraestrutura econômica básica; Experiência positiva de integração, conseguindo adicionar mais US$5.7 milhões para investimentos complementares, em regime de parcerias institucionais; Focalização satisfatória dos benefícios do projeto para os redutos mais pobres. Ampla cobertura dos Conselhos Municipais, em todo o estado; Resultados agregados atingindo ou superando metas agregadas, para a maioria dos indicadores.. Tendo em vista os conceitos que deram base às avaliações dadas para aspectos mais específicos do Projeto - todas também Moderadamente Satisfatória, a Unidade Técnica confirma também sua concordância com estas notas. 2. Sobre as lições aprendidas ou sobre as experiências relevantes do Projeto, temos a informar que o elenco dessas lições, mencionadas no item 6 do ICR, constituem pontos de reflexão e de orientação, que deverão ser devidamente considerados, não somente no contexto dos projetos financiados pelo Banco Mundial, como também em outros programas a serem implementados no Estado do Piauí. 76

91 3. Consideramos como relevantes as lições sobre a desproporção entre o pequeno valor do projeto e a dimensão do problema da pobreza no Piauí; a carência de uma assistência técnica mais massiva e mais presente junto aos produtores; a necessidade de capacitação dos diversos atores envolvidos na execução de projetos descentralizados; a questão da sustentabilidade operacional dos Conselhos Municipais, e a constatação, extremamente pertinente do ICR de que os investimentos em água e eletricidade, por si, não se mostram capazes de impulsionar as atividades produtivas, e reduzir a pobreza rural, sem uma série de outros insumos. E para justificar esta assertiva o ICR, conclui enfaticamente: As comunidades pobres da zona rural precisam ter acesso ao capital para investimento (produtivo), para poder tirar proveito da renda potencial(proporcionada) pelos serviços básicos, e de assistência técnica e treinamento para gerenciar a complexidade das oportunidades produtivas, em termos de acesso aos mercados e geração de empregos. (pag 24). Com respeito a esta última consideração (da necessidade de se direcionar as intervenções para promover a geração de renda), o Governo do Estado manifesta o seu interesse em continuar a longa e produtiva parceria que vem mantendo com o Banco Mundial, através de uma nova operação, cujo objetivo de desenvolvimento seria mais amplo do que o almejado no contexto dos PCPR. Para tanto, o Governo do Estado já alocou grande parte da sua capacidade de endividamento para uma operação de crédito no valor estimado de US$130 milhões. Ao tempo em que está em fase de preparação de uma Carta-Consulta para ser submetida a aprovação da SEAIN/COFIEX apresentando as idéias básicas pretendidas para o futuro Projeto. Por fim, em nome do Governador do Estado, do Secretário de Planejamento e da equipe da Unidade Técnica, apresentamos congratulações a Vossa Senhoria e à equipe do Banco Mundial, responsável pela preparação do ICR. O documento em apreço, pela abrangência e profundidade das análises, é de inegável importância informativa, e sem dúvida alguma, se constituirá em uma referência para a concepção de futuros programas visando reduzir os níveis de pobreza do Estado. Atenciosamente, Fernando Antonio Danda Vasconcelos Coordenador Geral da CCPR/UT-PCPR 77

92 78

93 Annex 8. Comments of Co-financiers and Other Partners/Stakeholders N/A 79

94 Annex 9. List of Supporting Documents 1. Project Appraisal Document (PAD), June 4, 2001, Report No BR 2. Project Paper (Additional Financing), June 16, 2006, Report No Rural Poverty Reduction In Northeast Brazil An Evaluation of Community Driven Development, Binswanger, Amazonas, Barbosa, Costa, Menezes, Pazello and Romano, World Bank Relatório para Revisão de Meio Termo (Mid-term Review Report), Instituto Civitas/IICA, September 2004 (including EDF) 5. Estudo de Desempenho Fisico EDF, Relatorio Final (Physical Performance Study), Fonseca and Melo/IICA, Estudo de Desempenho Fisico dos Subprojetos da Fase II EDF Complementar, Fonseca and Melo/IICA, Borrower Completion Report Loan 4624-BR (Phase 1), Lerzundi Silvera/IICA/SEPLAN, May Borrower Completion Report Loan 7399-BR (Phase 2), Fonseca/SEPLAN, June Implementation Status Reports (ISR) 10. Supervision Aide Memoires 11. Fiduciary Supervision Reports (Financial Management and Procurement) 12. Audit Reports and correspondence 13. Loan and Guarantee Agreements 14. Other preparation and supervision documents 80

95 MAP: State of Piauí 81