Document of The World Bank FOR OFFICIAL USE ONLY Report No: ICR00004225 IMPLEMENTATION COMPLETION AND RESULTS REPORT ON A GRANT (IDA-H3370, IDA-6720, IDA-H8570 and TF 99274) IN THE AMOUNT OF SDR 166.41 MILLION AND FOR IFAD CO-FINANCING ON A GRANT (C1200 and C1380) IN THE AMOUNT OF SDR 5.77 MILLION TO NEPAL FOR THE Poverty Alleviation Fund II (PAFII) June 27, 2019 Agriculture Global Practice South Asia Region This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization. CURRENCY EQUIVALENTS (Exchange Rate Effective {December 31, 2018}) Currency Unit = NPR NPR 112.23= US$ 1 US$ 1.41 = SDR 1 FISCAL YEAR July 1 - June 30 Regional Vice President: Hartwig Schafer Country Director: Idah Z. Pswarayi-Riddihough Country Manager: Faris H. Hadad-Zervos Senior Global Practice Director: Juergen Voegele Practice Manager: Loraine Ronchi Task Team Leader(s): Mio Takada, Kamran Akbar ICR Main Contributor: Javier Bronfman Horovitz ABBREVIATIONS AND ACRONYMS AF Additional Financing AM Aide Memoire CDD community-driven development CIAA Commission for Investigation of Abuse and Authority CIE Comprehensive impact evaluation CO community organization CPF Country Partnership Framework CPS Country Partnership Strategy DDC District Development Committee DPR detailed project report EFA economic and financial efficiency analysis EMF Environmental Management Framework ERR economic rate of return EIRR economic internal rate of return EMSF Environmental and Social Management Framework ENPV expected net present value ESMG Environmental and Social Management Guideline FIRR financial internal rate of return FNPV financial net present value GDP gross domestic product GESI Gender Equality and Social Inclusion GoN Government of Nepal GRM grievance redressal mechanism HDI Human Development Index ICR Implementation Completion Results report IDA International Development Agency IFAD International Fund for Agricultural Development ISN interim strategy note ISR implementation and supervision report M&E monitoring and evaluation MIS Management Information System MoLMCPA Ministry of Land Management, Cooperatives and Poverty Alleviation NPR Nepali Rupee NPV net present value O&M operations and maintenance OP/BP Operations Manual/Bank Procedures OPMCM Office of Prime Minister and Council of Ministers PAD Project Appraisal Document PAF Poverty Alleviation Fund PDO Project Development Objective PO Partner Organization PRSP Poverty Reduction Strategy Paper PSU Primary Sample Unit SDR Special Drawing Rights SoE statement of expenditures US$ United States Dollar VAT value added tax VCDP Vulnerable Community Development Plan VDC Village Development Committee TABLE OF CONTENTS DATA SHEET ...............................................................................................................................1 I. PROJECT CONTEXT AND DEVELOPMENT OBJECTIVES ....................................................... 6 A. CONTEXT AT APPRAISAL .........................................................................................................6 B. SIGNIFICANT CHANGES DURING IMPLEMENTATION (IF APPLICABLE) .......................................9 II. OUTCOME .................................................................................................................... 13 A. RELEVANCE OF PDOs ............................................................................................................ 13 B. ACHIEVEMENT OF PDOs (EFFICACY) ...................................................................................... 13 C. EFFICIENCY ........................................................................................................................... 17 D. JUSTIFICATION OF OVERALL OUTCOME RATING .................................................................... 19 E. OTHER OUTCOMES AND IMPACTS ......................................................................................... 19 III. KEY FACTORS THAT AFFECTED IMPLEMENTATION AND OUTCOME ................................ 21 A. KEY FACTORS DURING PREPARATION ................................................................................... 21 B. KEY FACTORS DURING IMPLEMENTATION ............................................................................. 21 IV. BANK PERFORMANCE, COMPLIANCE ISSUES, AND RISK TO DEVELOPMENT OUTCOME .. 23 A. QUALITY OF MONITORING AND EVALUATION (M&E) ............................................................ 23 B. ENVIRONMENTAL, SOCIAL, AND FIDUCIARY COMPLIANCE ..................................................... 24 C. BANK PERFORMANCE ........................................................................................................... 27 D. RISK TO DEVELOPMENT OUTCOME ....................................................................................... 29 V. LESSONS AND RECOMMENDATIONS ............................................................................. 30 ANNEX 1. RESULTS FRAMEWORK AND KEY OUTPUTS ........................................................... 34 ANNEX 2. BANK LENDING AND IMPLEMENTATION SUPPORT/SUPERVISION ......................... 46 ANNEX 3. PROJECT COST BY COMPONENT ........................................................................... 48 ANNEX 4. EFFICIENCY ANALYSIS ........................................................................................... 49 ANNEX 5. BORROWER, CO-FINANCIER AND OTHER PARTNER/STAKEHOLDER COMMENTS ... 55 ANNEX 6. SUPPORTING DOCUMENTS .................................................................................. 59 ANNEX 7. COMPREHENSIVE IMPACT EVALUATION ............................................................... 60 The World Bank Poverty Alleviation Fund II (P105860) DATA SHEET BASIC INFORMATION Product Information Project ID Project Name P105860 Poverty Alleviation Fund II Country Financing Instrument Nepal Investment Project Financing Original EA Category Revised EA Category Partial Assessment (B) Partial Assessment (B) Organizations Borrower Implementing Agency Ministry of Finance Poverty Alleviation Fund Project Development Objective (PDO) Original PDO The project development objective is to improve living conditions, livelihoods and empowerment among the rural poor, with particular attention to groups that have traditionally been excluded by reasons of gender, ethnicity, caste and location. Page 1 of 64 The World Bank Poverty Alleviation Fund II (P105860) FINANCING Original Amount (US$) Revised Amount (US$) Actual Disbursed (US$) World Bank Financing 100,000,000 99,969,363 99,460,999 IDA-H3370 27,500,000 27,500,000 0 IDA-49220 65,000,000 64,998,733 63,431,353 IDA-H6720 10,000,000 10,000,000 10,000,000 TF-99274 80,000,000 80,000,000 73,371,033 IDA-H8570 Total 282,500,000 282,468,096 246,263,385 Non-World Bank Financing 0 0 0 Borrower/Recipient 21,000,000 21,000,000 6,500,000 International Fund for 9,000,000 9,000,000 6,822,484 Agriculture Development Total 30,000,000 30,000,000 13,322,484 Total Project Cost 312,500,000 312,468,095 259,585,869 KEY DATES FIN_TABLE_DAT Approval Effectiveness MTR Review Original Closing Actual Closing 06-Dec-2007 19-Mar-2008 13-May-2010 30-Sep-2012 31-Dec-2018 Page 2 of 64 The World Bank Poverty Alleviation Fund II (P105860) RESTRUCTURING AND/OR ADDITIONAL FINANCING Date(s) Amount Disbursed (US$M) Key Revisions 02-Jun-2011 98.49 Additional Financing Change in Results Framework 05-Jun-2013 138.80 Additional Financing Change in Results Framework Change in Components and Cost 25-Jun-2014 162.80 Change in Loan Closing Date(s) Reallocation between Disbursement Categories 12-Feb-2016 208.91 Change in Results Framework Change in Components and Cost Change in Loan Closing Date(s) Reallocation between Disbursement Categories Change in Implementation Schedule Other Change(s) 30-Nov-2017 236.26 Change in Components and Cost Reallocation between Disbursement Categories KEY RATINGS Outcome Bank Performance M&E Quality Moderately Satisfactory Moderately Satisfactory Modest RATINGS OF PROJECT PERFORMANCE IN ISRs Actual No. Date ISR Archived DO Rating IP Rating Disbursements (US$M) 01 28-Jun-2008 Satisfactory Satisfactory 0 02 26-Dec-2008 Satisfactory Satisfactory 0 03 21-May-2009 Satisfactory Satisfactory 32.83 04 28-Nov-2009 Satisfactory Moderately Satisfactory 33.58 05 28-May-2010 Satisfactory Moderately Satisfactory 48.16 06 04-Jan-2011 Highly Satisfactory Satisfactory 75.22 07 29-Oct-2011 Highly Satisfactory Satisfactory 102.20 Page 3 of 64 The World Bank Poverty Alleviation Fund II (P105860) 08 14-Apr-2012 Highly Satisfactory Satisfactory 121.42 09 18-Oct-2012 Highly Satisfactory Satisfactory 128.57 10 28-Apr-2013 Satisfactory Satisfactory 142.52 11 29-Aug-2013 Satisfactory Satisfactory 142.52 12 29-Jan-2014 Satisfactory Moderately Satisfactory 152.18 13 11-Apr-2014 Satisfactory Moderately Satisfactory 166.61 14 26-Aug-2014 Satisfactory Moderately Satisfactory 166.61 15 27-Jan-2015 Satisfactory Moderately Satisfactory 166.61 16 22-Jul-2015 Moderately Satisfactory Moderately Satisfactory 191.40 17 25-Jan-2016 Moderately Satisfactory Moderately Satisfactory 191.40 18 21-Mar-2016 Moderately Satisfactory Moderately Satisfactory 212.72 19 19-Sep-2016 Moderately Satisfactory Moderately Satisfactory 212.72 20 20-Mar-2017 Moderately Satisfactory Moderately Satisfactory 230.82 21 29-Oct-2017 Satisfactory Moderately Satisfactory 230.82 22 27-Apr-2018 Satisfactory Moderately Satisfactory 242.32 23 26-Jul-2018 Satisfactory Satisfactory 242.32 24 01-Feb-2019 Satisfactory Satisfactory 242.43 SECTORS AND THEMES Sectors Major Sector/Sector (%) Agriculture, Fishing and Forestry 51 Fisheries 13 Irrigation and Drainage 25 Livestock 13 Social Protection 25 Social Protection 25 Page 4 of 64 The World Bank Poverty Alleviation Fund II (P105860) Water, Sanitation and Waste Management 24 Water Supply 24 Themes Major Theme/ Theme (Level 2)/ Theme (Level 3) (%) Private Sector Development 100 Jobs 100 Social Development and Protection 41 Social Inclusion 41 Indigenous People and Ethnic Minorities 7 Other Excluded Groups 7 Participation and Civic Engagement 27 Urban and Rural Development 60 Rural Development 60 Rural Non-farm Income Generation 40 Rural Infrastructure and service delivery 20 ADM STAFF Role At Approval At ICR Regional Vice President: Praful C. Patel Hartwig Schafer Country Director: Susan G. Goldmark Idah Z. Pswarayi-Riddihough Senior Global Practice Director: Constance A. Bernard Juergen Voegele Practice Manager: Adolfo Brizzi Loraine Ronchi Task Team Leader(s): Geeta Sethi Mio Takada, Kamran Akbar ICR Contributing Author: Javier Bronfman Horovitz Page 5 of 64 The World Bank Poverty Alleviation Fund II (P105860) I. PROJECT CONTEXT AND DEVELOPMENT OBJECTIVES A. CONTEXT AT APPRAISAL Context 1. At the time of the appraisal (2007), political turbulence and civil conflict due to the decade-long (1996– 2006) conflict in Nepal had imposed high costs in terms of foregone growth. With average gross domestic product (GDP) per capita of US$ 270 in 2005, Nepal was the poorest country in South Asia and the twelfth poorest in the world. Nevertheless, Nepal made significant progress in decreasing overall poverty rate from 42 percent to 31 percent during 1996–2006, although with greater gains in urban areas than rural areas. The poverty gap was also related to longstanding patterns of gender-, ethnic- and caste-based exclusion, as well as the above-mentioned conflict. Economic and social outcomes were particularly poor in remote areas, especially in the mountain districts of Mid-Western and Far-Western development regions. The Maoist insurgency that started in the Mid-Western hills in 1996 led to lack of economic opportunity and extreme inequality which fueled grievances against the Government of Nepal (GoN). Social discrimination based on these identity markers reflected economically in poverty rates, with the highest among Janajati indigenous peoples and Dalit lowest caste group. Reaching these marginalized groups and improving their access to resources, assets and income, in addition to their voice and participation in mainstream development and livelihoods opportunities, was therefore critical to a more stable, inclusive and sustainable development path for Nepal. 2. During the conflict period, the GoN (in 2003) sought the Bank’s financial and technical assistance for developing and establishing the Poverty Alleviation Fund (PAF) as a semi-autonomous agency to address related problems of rural poverty and social exclusion, using a community-driven development (CDD) instrument. With a US$ 15 million grant, the first PAF project (PAF I) was designed as a pilot, operating in six districts that were chosen on the basis of the Human Development Index (HDI) and incidence of conflict. The PAF I project achieved notable results on the ground, and the implementation pace of PAF accelerated rapidly in response to faster than anticipated and constantly growing demand from rural communities. By mid-2006, 80 percent of project resources had been committed, and consequently the International Development Agency (IDA) approved an Additional Financing (AF) grant of US$ 25 million in November 2006 to enable PAF to extend its intervention area from 6 to 25 districts. 3. Given the lessons and successes under PAF I and the associated AF, increased pressure for CDD activities outside these 25 districts and the commitment to the peace process, the GoN sought the Bank’s financial and technical assistance to scale up the PAF approach to the national level through a second PAF (PAF II) project that would cover all 75 districts, and be accessible to 30 percent of the rural population. At the time of appraisal, PAF II was envisaged to use the same targeting mechanisms and beneficiary profiles as PAF I, but was planned to extend its scope nationally in a phased manner, increasing by 15 districts each year. Theory of Change (Results Chain) 4. The theory of change behind the project design — as described in the Project Appraisal Document (PAD) — is illustrated in Figure 1. The figure conceptualizes the manner in which intervention leads to achieving the project development objective(s) (PDOs) and contributes to the overall development goal of poverty reduction. Page 6 of 64 The World Bank Poverty Alleviation Fund II (P105860) Figure 1: PAF II theory of change Project activities Outputs Intermediate PDO outcomes Long-term outcomes impact Asset creation and Community Improved household Rural support to livelihoods infrastructure created Improved living conditions (to be poverty - and improved access and use measured by access to alleviation mall-scale village and of services infrastructure) , and community Income-generating sustainabl infrastructure activities supported Improved Improved household e and market livelihoods (to be inclusive - linkages measured by income economic ncome generating Local business increase) growth sub-projects development Product development supported and linked Improved levels of and market linkages with markets participation and Community level Local participation in Community empowerment institution building decision making and individual empowerment Marginalized population Effective targeting to Participation of reached reach out to the marginalized marginalized population population Capacity building and Local prioritization of Enhanced participatory decision demands through institutional making participatory decision capacity for making at the integrated community level resources management Critical Assumptions:  Political economy of public space allows for direct funding to community level institutions.  Marginalized population is able and willing to take advantage of the opportunities created by the project.  Local government budget allocation and staffing allows for improving rural infrastructure services to compliment project activities.  Private companies and service providers recognize economic opportunities aimed at market linkages for income generation activities promoted by the project. Project Development Objectives (PDOs) 5. PAF II aimed at assisting the GoN in scaling up PAF I (approved in 2004 and closed in 2009). At appraisal, the PDO was “to improve living conditions, livelihoods and empowerment among the rural poor, with particular attention to groups that have traditionally been excluded by reasons of gender, ethnicity, caste and location” (World Bank 2007b, p.5). The Financial Agreement states the same project objective. The PDO remained unchanged during the life of the project. Page 7 of 64 The World Bank Poverty Alleviation Fund II (P105860) Key Expected Outcomes and Outcome Indicators 6. As of the theory of change illustrated above, and the project´s results framework, the following key expected outcomes and indicators were monitored.  Outcome 1: Improved household living conditions. At least 164,000 households benefiting from access to community infrastructure by the end of the project. o Indicator: Number of households benefitting from increased access to community infrastructure.  Outcome 2: Improved household livelihoods. At least 70 percent of beneficiaries increased their income by 15 percent or more against base year by the end of the project. o Indicator: Percentage of beneficiary households that have increased their incomes by at least 15 percent against base year (2007) by the end of the project.  Outcome 3: Improved level of participation and empowerment. At least 60 percent of key positions in project community organizations (COs) are from targeted households. o Indicator: Percentage of key positions held in project COs from targeted households.  Outcome 4: Marginalized population reached. At least 70 percent of CO members are women. o Indicator: Numbers and percentages of direct and indirect beneficiaries by gender and caste. Components 7. Given the success of PAF I, PAF II maintained its features and worked under the five components described below. Nonetheless, throughout the life of the project and after the first and second AFs these components evolved to maintain relevance and adjust to country needs to the extent possible. These changes are discussed in the section on “Revised Components”. 8. Component A. Small-scale village and community infrastructure (US$ 32 million at original appraisal, US$ 28.5 million at first AF, US$ 15.7 million at second AF, US$ 0.8 million of International Fund for Agricultural Development (IFAD) co-financing; total actual US$ 71.25 million at closure): The objective of this component was to finance small investments in infrastructure and services demanded by COs formed by PAF under Component D. Consideration was given to sub-projects already included in Village Development Committee (VDC) plans and/or where the District Development Committee (DDC) and/or VDC were prepared to provide matching grants. Key selection criteria included number of target households that would benefit and the expected employment impacts. Communities were expected to contribute 10 percent of the total cost on average, as well as prepare and commit to an operations and maintenance (O&M) plan for each sub-project built. To make community infrastructure sub-projects more sustainable, PAF prepared O&M guidelines to show which projects made provisions for: establishing O&M funds, creating O&M committee and imparting capacity building training for O&M personnel. 9. Component B. Income generating sub-projects (US$ 37 million at original appraisal, US$ 21.66 million and US$ 10 million of Trust Fund from Global Food Crisis Response Program (GFRP) at first AF, US$ 29.1 million at second AF, US$ 1.8 million of IFAD co-financing; total actual US$ 100.31 million at closure): The objective of this component was to provide matching grants to self-selected groups of poor and excluded people for income-generation activities, based on objective criteria including ethnicity, caste, gender and poverty levels. Beneficiaries were expected to contribute about 10 percent of the sub-project cost in cash. Particular attention was paid to the commercial/economic viability of income-generating sub-projects to avoid the risk of supply driven proposals that were not sufficiently well linked with markets. 10. Component C. Innovation and special programs (US$ 7 million at appraisal, US$ 3.07 million at first AF, US$ 9.5 million at second AF; total actual US$ 19.78 million at closure): The objective of this component was to provide an innovations window to support proposals meriting special consideration owing to exceptional Page 8 of 64 The World Bank Poverty Alleviation Fund II (P105860) needs in a given context, or demonstrating innovative ways to improve livelihoods development and reach of the targeted population. These proposals can be national or regional in coverage and may be sponsored by organizations representing or working with the target groups, target groups themselves and/or local bodies. 11. Component D. Capacity building and implementation support (US$ 19 million at appraisal, US$ 8.7 million at first AF, US$ 14.5 million at second AF, US$ 5.2 million of IFAD co-financing; total actual US$ 42.04 million at closure): The objective of this component was to improve capacity to implement PAF projects through partner organizations (POs), and to support implementing capacity of PAF. This component included the following six sub-components: (i) social mobilization of community groups; (ii) capacity building for local bodies; (iii) capacity building for target groups engaged in income generating activities; (iv) support to rural and community finance; (v) information, monitoring and evaluation (M&E); and (vi) PAF´s management information system (MIS). 12. Component E. Administration of PAF II (US$ 5 million at appraisal, US$ 3.07 million at first AF, US$ 11.2 million at second AF, total actual US$ 19.70 million at closure): This component aimed at co-financing operating costs of the PAF (staff costs and operating expenses). Recognizing PAF as an autonomous agency with the mandate to reduce poverty, the GoN committed to fund 20 percent of PAF´s operating costs. PAF was responsible for coordination of project execution, and for ensuring appropriate activity planning and timely implementation. B. SIGNIFICANT CHANGES DURING IMPLEMENTATION (IF APPLICABLE) Revised PDOs and Outcome Targets  PDO was not revised during project implementation.  PAF II had two AFs (2011 and 2013), and three “restructurings”, and targets were adjusted.  PDO indicators were unchanged throughout the project. Indicator 3 (percentage of beneficiary group members coming from targeted groups) was dropped at the first AF and was integrated with the indicator for the representation of targeted groups for decision making (see PDO Indicator 4 in the Table 1). Table 1: Revised indicators and targets At appraisal First Second Project 2007 Additional AF restructuring PDO indicators Financing (AF) 2013 2016 2011 Number of households benefitting from 1 increased access to community 15,000 100,000 164,000 No change infrastructure Percent of beneficiary households that have increased their incomes by at least 15 2 25 60 70 No change percent against base year, by the end of the project Percent of beneficiary group members 3 30 Dropped coming from targeted groups Page 9 of 64 The World Bank Poverty Alleviation Fund II (P105860) Percent of decision making positions in 4 community groups occupied by target 30 groups At first AF: changed to “Percent of key positions in project COs from targeted 50 60 No change households” Project beneficiaries 5 n/a 470,000 575,000 589,000  CO members/households  Non-members/households n/a 180,000 55,000 No change Percent of female CO members (of total CO 6 n/a 75 75 No change members) Page 10 of 64 The World Bank Poverty Alleviation Fund II (P105860) Revised Components 13. At the time of the second AF (2013), the name of Component C was changed from “Innovation and special programs” to “Product development, market linkages and pilots”, to better describe the innovative activities proposed under the AF. The objective of this component under the revised name was to support COs that were more advanced through the pocket area approach and to pilot-test the PAF model in a peri-urban setting. The revised component aimed to help COs establish production and marketing cooperatives in the form of pocket area development (which is a spatial development approach for clustered COs). The name of Component E “Administration of PAF II” was changed to “Project management, planning and M&E”. While the project management modality remained the same, a few changes were incorporated to strengthen MIS to increase coordination with local bodies, and intensify M&E activities. 14. At the time of the 2016 restructuring in response to the earthquake, while the components remained unchanged, funds were reallocated among the components. Under Component A and B, a sub-component was added to each, to specifically target earthquake-hit areas. Under Component D, two sub-components were added for social mobilization of affected COs and skills development for earthquake recovery needs. These additional sub-components were co-financed by IDA (75 percent) and IFAD (25 percent). Under Component E, a sub-component was added to strengthen technical support for after-earthquake related needs. A sub-component on knowledge management financed by IFAD was also added. 15. During the 2017 restructuring, while the components remained unchanged, funds were reallocated among components to adjust the outstanding needs for the remaining period. Similarly, the proportion of co-financing for the earthquake related sub-components were adjusted to address the financing needs of the remaining period (IFAD 75 percent and IDA 25 percent). Page 11 of 64 The World Bank Poverty Alleviation Fund II (P105860) Other Changes 16. The project had three level-2 restructurings and two AFs in about eleven years of project implementation. Financing at appraisal (H3370) was US$ 100 million (SDR 65.24 million). Details of financing from different sources are summarized in Annex 3. Cumulative total amount of IDA grant approved was US$ 245 million, total amount of Trust Fund approved was US$ 10 million, and cumulative total amount from IFAD was US$ 9 million. Key changes are summarized in Table 2. Table 2: Major changes to PAF II Year Type Objective Key changes 2008 IFAD co- Scale up the activities under Component D Additional US$ 4 million (SDR 2.5 million) of IFAD financing grant (C1200) 2011 First AF Expand the project area Additional US$ 65 million (SDR 65.24 million) of IDA Scale up targets grant (H3370) Scale up income generation activities Additional US$ 10 million (SDR 6.37 million) from under Component B to address Food Price Food Price Crisis Response Trust Fund (TF 99274) Crisis Additional US$ 27.5 million of IDA credit (49220) was proposed at appraisal, but canceled 2013 Second AF Expand the project area Additional US$ 80 million (SDR 53.4 million) of IDA Scale up targets grant (H8570) Change the name of Component B to pilot pocket area and peri-urban activities 2014 Restructuring Fully utilize the balance of the first AF Change in closing date of first AF Reallocation between disbursement categories 2016 Restructuring Add activities under A, B, D and E to Change in results framework address the needs of earthquake related Change in components and cost activities Change in closing date of second AF Reallocation between disbursement categories 2016 IFAD co- Add activities under A, B, D and E to Additional US$ 5 million (SDR 3.27 million) of IFAD financing address the needs of earthquake related grant (C1380) activities Add knowledge management activities under Component E 2017 Restructuring Fully utilize the balance of second AF Change in components and cost Reallocation between disbursement categories Rationale for Changes and Their Implication on the Original Theory of Change 17. Given the initial success of the project and its proper disbursement, the changes rationalized in the AF and the restructuring process contributed directly to achieving initially defined intermediary outcomes, and thus had no adverse implications on the project´s theory of change. Additionally, components were changed in order to respond to emergency needs due to the 2015 earthquake. Page 12 of 64 The World Bank Poverty Alleviation Fund II (P105860) II. OUTCOME A. RELEVANCE OF PDOs Assessment of Relevance of PDOs and Rating 18. PAF II was consistent with the Country Assistance Strategy (CAS) for Nepal (2003) and the Interim Strategy Note (ISN) (2007). CAS identified PAF as a promising instrument for targeted support to achieve rural poverty reduction and social inclusion objectives. ISN highlighted the benefits of community-driven development programs and the success of PAF I in delivering poverty reduction results. The expansion of community-based programs and their success in reaching previously marginalized people during conflict times, helped set this as a priority for IDA support. Likewise, the two ISNs (2010–2011 and 2012–2013) reflected continuity with the previous ISN and the 2004–2006 CAS focused on inclusive growth, fostering of gender equality and promotion of social inclusion. 19. Towards the end of the project, however, the PDOs became less relevant to the GoN. The GoN development plan identified the need for investments in economic, physical and social infrastructures as a more relevant approach to promote economic progress and help Nepal become a middle-income country by 2030. The Bank’s Nepal Country Partnership Framework (CPF) (2019–2023) highlighted three focus areas for engagement: (i) public institutions; (ii) private-sector led jobs and growth; and (iii) inclusion and resilience. PAF and CDD projects were removed from the latest CPF. As a result of Nepal’s shift to federalism, PAF as an autonomous agency was criticized for bypassing local governments, thus both the objectives and the mode of operation lost their significance at the end of the project. 20. Overall, the project remained consistent with the Bank’s strategy throughout implementation. However, there were modest shortcomings with regard to the relevance of the development objectives to the Bank CPF and the GoN’s priorities, as well as PAF’s institutional relevance as a semi-autonomous agency during the process of federalism, towards the end of the project. Therefore, the relevance of PDO is rated as SUBSTANTIAL. B. ACHIEVEMENT OF PDOs (EFFICACY) Assessment of Achievement of Each Objective/Outcome 21. The PAF II PDO was “to improve living conditions, livelihoods and empowerment among the rural poor, with particular attention to groups that have traditionally been excluded by reasons of gender, ethnicity, caste and location”. The PDO was divided into four sub-objectives: (i) improve living conditions; (ii) improve livelihoods; (iii) improve levels of empowerment; and (iv) reach the marginalized population. 22. The assessment of the project’s efficacy drew mainly from its MIS, the analysis of three panel household surveys covering 16 districts with a sample of 4,855 households, and the results from a comprehensive impact evaluation (CIE) conducted by an independent evaluator under the supervision of the National Planning Commission of the GoN (2018). The CIE was based on a mixed methods approach – assessing PAF II results and impacts using both quantitative and qualitative approach. For this, Difference in Difference estimations were calculated and contrasted with the qualitative analysis gathered in the field through 36 focus group discussions, consultative meetings and key interviews (see Annex 7). The analysis also benefited from a set of thematic studies such as CO maturity assessment, infrastructure sustainability assessment as well as the Borrower’s Project Completion Report. Sub-objective 1: Improve living conditions (At least 164,000 households benefiting from access to community infrastructure by the end of the project). This has been fully achieved. Page 13 of 64 The World Bank Poverty Alleviation Fund II (P105860) 23. The project’s MIS data showed that 317,404 households benefited from increased access to community infrastructures – almost double of what was expected. This was because allocation of block grants was reduced from the original design to encourage CO members to revolve funds among members, thus allowing more funds for investments in infrastructure. Further, many COs clustered themselves to form a CO federation to receive larger amount of funds so that they could invest in larger infrastructures, thus allowing more households to benefit from access to infrastructure. PAF II completed 4,449 sub-projects (against a target of 4,000 infrastructure sub-projects). Sub-projects mostly consisted of water supply and sanitation, irrigation, rural access, rural energy and community building. A sustainability study of community infrastructure conducted by PAF in 2014 found that about 69 percent of the infrastructure was sustainable with adequate O&M funds established by COs with contribution, while 30 percent was fairly sustainable with nearly adequate O&M funds. 24. According to CIE (2018), improved access to infrastructure helped free up time and relieve people from drudgery. Women and children alike experienced a decrease in hours spent on household chores, such as water fetching. While adults had time to attend to income generating and community level activities, children, particularly girls, had the opportunity to attend school more regularly. Better connectivity (such as roads and bridges) allowed community members timely access to medical attention and improved their access to the market. With irrigation infrastructure, CO members were able to increase their income and food sufficiency through agriculture. Similarly, benefits from community buildings (which included multi-purpose halls, schools, vegetable collection centers, chilling centers and rural roads), improved the living conditions of CO members as they had better access to these services. 25. In response to the 2015 earthquake, PAF supported the rehabilitation of community infrastructures in earthquake-hit areas. Initially, a rapid assessment showed that 233 infrastructures were partially damaged and 57 were completely damaged. Later, during a detailed assessment, rehabilitation of only 256 projects were identified, and hence the target was revised to 256 at the second restructuring. The project substantially achieved the revised target. Sub-objective 2: Improve household livelihoods (At least 70 percent of beneficiaries increased their income by 15 percent or more against base year by the end of the project). This has been fully achieved. 26. According to the household survey analysis, 71.4 percent of households increased their income by at least 15 percent against the base year by the end of the project. This result exceeded the target by 1.4 percentage points. 27. Further, to supplement the PDO indicator on income, the CIE also analyzed: (i) food security (one of intermediate indicators), (ii) asset creation, and (iii) expenditure, to ensure the findings on household livelihoods. In terms of food security, 41.5 percent beneficiaries improved their food security in terms of months of food sufficiency, which is less than the target of 68 percent but still a very positive result. In terms of asset creation and accumulation, CIE results showed overall positive and statistically significant impacts as well. PAF intervention helped CO members to accumulate assets in various forms. Targeted households significantly increased the probability of possessing a radio and a phone (a significant increase of 8 percentage points in the probability of owning a radio compared to the control and 3 percentage points increase in phone possession for PAF II beneficiaries was calculated). In terms of livestock assets, beneficiaries from PAF II increased the total value of livestock more than the control population. CIE results showed positive and statistically significant effects on household expenditure. Households that accessed the PAF revolving fund increased their total expenditure by 22 percent compared to control households. Treated households significantly increased their expenditure on human development investments. On average, beneficiary households spent 43–53 percent more on health and education compared to the control group. Page 14 of 64 The World Bank Poverty Alleviation Fund II (P105860) 28. The pocket area development pilot was added under Component C at the second AF to carry out activity-based development programs. The pilot was originally expected to establish 60 pocket areas, however, at the time of the 2016 restructuring, the target was reduced to 30 pocket areas to allow adequate time to monitor the results of the first 30 pocket area pilots and to nurture the pocket areas to establish a sustainable market and finance linkages. 29. The pilot successfully established 30 pocket areas as per the revised target. It further supported the 38 selected cooperatives (comprising mainly PAF CO members) through a series of training programs on entrepreneurship and enterprise development. The pilot benefitted 363 COs and 8,759 CO members. PAF MIS data (baseline and endline of the pilot) indicated that the activity was successful in generating income worth NPR 393 million among the participating pocket areas with average household income increase of 35 percent (in real term) over the baseline, and generated employment for 5,708 individuals (of which 65 percent were women). 30. The peri-urban pilot was also added under Component C of the second AF to support the poor population in peri-urban areas through a set of vocational training and income generation activities. Initially, the pilot was expected to establish 90 peri-urban COs. However, after a thorough assessment, peri-urban costs were discovered to be higher, and hence the number of COs were reduced to 40 in two peri-urban areas to provide adequate funds. 31. The 40 COs (20 in Kathmandu and 20 in Butwal) supported coverage of 1,014 households involved in 174 sub- projects (111 for income generation activities and 63 for skill enhancement training). PAF MIS data showed that there was a capital accumulation of NPR 75 million with an increased income of 27.3 percent (in real term) among the participating beneficiaries. 32. At the time of the 2016 restructuring, a sub-component was added under Component B to support the earthquake-hit COs through additional block grants to revitalize their revolving fund to support income generation activities. This sub-component supported 11,578 CO members in rebuilding their livelihoods. The initial assessment indicated that about 4,900 COs (around 40,000 CO members) who were affected by the earthquake would need asset/cash transfer support for revitalizing the revolving fund. Later, a detailed analysis indicated that 1,500 COs (around 14,000 households) were actually affected by the earthquake and eligible for additional support. Hence the target was reduced to 14,000 households. The project substantially achieved the revised target. Sub-objective 3: Improve level of participation and empowerment (At least 60 percent of key positions in project COs are from targeted households). This has been fully achieved. 33. About 63 percent individuals from the targeted populations (Dalit, Janajati and Muslim) filled key positions in COs. This represented an overachievement with respect to the target of 60 percent. Furthermore, women represented 78 percent of key positions. PAF was successful in increasing the participation of females and minorities in COs and community decision making. This was one of the most frequent benefits reported by beneficiaries during the field visits. Page 15 of 64 The World Bank Poverty Alleviation Fund II (P105860) Figure 2: Key position distribution by cast/ethnicity Source: PAF MIS 34. Formation of COs, social mobilization and capacity building were key contributors to improvements in socio- economic empowerment of women. The results from the CIE report and Borrowers Completion Report showed that in all COs visited, members expressed that PAF involvement had changed the community’s views on minorities and women. PAF was able to integrate the marginalized population and give them a space to voice their ideas and concerns as well as provide them resources (knowledge, monetary and social) to improve their own lives. Focus group discussions highlighted benefits related to active participation in CO meetings, increased mobility among women, increased respect for Dalits and increased level of self-dignity. Given that PAF CO members frequently participated in several groups and development activities it is hard to attribute the empowerment effect solely to PAF. However, the analysis of controlling the membership to other groups indicated that the estimated probability of women being asked when a property is sold increased between 5.7 and 22.3 percentage points when compared with non-CO member households (World Bank, 2018). 1 Sub-objective 4: Reach the marginalized population. (At least 70 percent of CO members are women). This has been fully achieved. 35. PAF MIS revealed that of the direct beneficiaries, 80 percent were women and 66 percent were ultra-poor (i.e. those who lack sufficient food for more than three months in a year). The PAF strategy was effective in reaching its target groups (Dalit, Janajati, Muslim, women and the poor) and successful in achieving the sub-objective. Nearly, 32 percent of PAF CO members were Janajati, 22 percent were Dalit and 3 percent were Muslim. The remaining 43 percent of beneficiaries were from other castes/ethnic backgrounds including Brahmin, Chhetri and Thakuri (considered to be higher castes but supported by PAF due to poverty criteria). Fieldwork did not find any evidence of elite capture, and overall the project can be considered very successful in reaching the historically marginalized population with resources and opportunities to better their lives. 1Results differed depending on the household survey panel used for the estimations resulting in results variance. For details see “Nepal Poverty Alleviation Fund Quantitative Impact evaluation” World Bank 2018. Page 16 of 64 The World Bank Poverty Alleviation Fund II (P105860) Figure 3: Beneficiaries distribution under PAF II Source: PAF MIS Justification of Overall Efficacy Rating 36. The overall efficacy of the project is rated as SUBSTANTIAL, since the project almost fully met its intended outcomes as defined by the PDO. There were modest shortcomings at the level of intermediate outcomes (i.e. food sufficiency improvements, CO members benefitting from cash transfer under earthquake activities, number of higher level institutions established). C. EFFICIENCY Assessment of Efficiency and Rating 37. The efficiency of the project is rated as SUBSTANTIAL. An ex-post economic and financial efficiency analysis (EFA) was undertaken to assess the financial and economic impact of PAF II. As the project was demand-driven a standard ex-ante EFA was not carried out at appraisal. The ICR analysis assessed the economic and financial viability of the project based on economic and financial data from a sample of 400 randomly drawn beneficiary households. The EFA was carried out on the basis of individual analysis of selected income generating activities and infrastructure schemes for which reliable and sufficient data were available to measure project benefits. 38. Returns to investments in selected activities were estimated using the respective production models. Main project benefits ensued from the promotion of income generating activities (Component B) and the construction/rehabilitation of key infrastructures (Component A) with support from the project. The returns to investments from capacity development for earthquake rehabilitation and program management, which were added under the first restructuring of the second AF, were excluded from the analysis due to lack of reliable data. Thus, the financial and economic returns estimated in the ICR efficiency analysis are a lower bound estimation of the benefits generated by project investment. 39. Based on the range of income generating activities and micro-infrastructure projects, 10 varied indicative activities were assessed in financial and economic terms.2 The EFA at project completion covered activities that accounted for 68 percent of the total project cost. The results showed considerable variations in returns across investments. The returns estimated showed that the investment in all income generating sub-projects and small-scale infrastructure sub-projects were reasonable and economically viable in terms of both financial and economic returns. While financial returns from income generating activities and small infrastructure 2 The ten activities assessed were: (i) poultry, (ii) cow/buffalo, (iii) goat, (iv) piggery, (v) retail business, (vi) vegetable farming, (vii) irrigation, (viii) micro hydro, (ix) water supply and (x) rice mill – six from income generating activities and four from small infrastructure programs. Page 17 of 64 The World Bank Poverty Alleviation Fund II (P105860) investments ranged from 6 percent to 26 percent, economic returns from these investments ranged from 13 percent to 47 percent. 40. The ex-post EFA at project completion found that the financial and economic returns to project investments in income generating activities and small infrastructures were 18 percent and 17 percent, respectively. The financial net present value (FNPV) and the expected net present value (ENPV) were about NPR 14,837 million (US$ 145 million) and NPR 11,978 million (US$ 117 million), respectively. Both economic and financial returns are net of returns from employment as the input wage costs were deducted for each investment. Actual benefits would be even higher if the returns on employment were considered. 41. If the conservative estimates of overall project costs were evaluated against the benefits of income generating and infrastructure activities only, the economic return was 12 percent with NPV of NPR 4,715 million (US$ 46 million). As the project was implemented over a longer period than originally planned, the analysis also tested the sensitivity to this prolonged implementation and found that both economic and financial returns declined but remained economically viable (see Annex 4). Therefore, the overall cost-effectiveness of the project is rated as SUBSTANTIAL. Table 3: Economic and financial returns from the ex-post efficiency analysis Investments Actual ENPV ERR (%) FNPV FIRR (%) investment ('000 NPR) ('000 NPR) ('000 NPR) A. Income generating activities Poultry 57 70 14% 130 17% Livestock-cow- 60 81 17% 149 21% buffalo Livestock-pig 45 66 21% 100 25% Livestock-goat 78 36 16% 54 18% Retail business 70 13 13% 35 17% Vegetable farming 20 8 19% 15 26% Total Income 9,455,616 7,851,027 16% 13,839,583 20% generating activities B. Infrastructure Irrigation 1,000 1,116 22% 572 16% Micro hydro 5,050 19,048 47% 8,244 22% Rice mill 645 285 15% 207 14% Water supply 955 916 21% (382) 6% Total infrastructure 7,540,241 9,763,552 24% 1,355,862 13% Total investment in 16,995,857 11,978,133 17% 14,837,188 18% income generating activities and infrastructure components Overall project 24,985,000 4,715,275 12% 7,574,330 13% Page 18 of 64 The World Bank Poverty Alleviation Fund II (P105860) D. JUSTIFICATION OF OVERALL OUTCOME RATING 42. The overall outcome rating is MODERATELY SATISFACTORY. This rating takes into account the SUBSTANTIAL rating for relevance of the PDO, and the SUBSTANTIAL rating for both efficacy and efficiency. There were moderate shortcomings with regard to the relevance of PDO to the Bank CPF and the GoN’s priorities towards the end of the project. There were modest shortcomings at the level of intermediate outcomes. E. OTHER OUTCOMES AND IMPACTS Gender and Social Inclusion 43. Since inception, PAF has been successful in implementing gender and social inclusion aspects. Women have been the project’s most benefited group. Although no specific quota was set for women, their participation in COs was required by design, which ensured their involvement. According to PAF MIS 80 percent of overall beneficiaries were female, around 60 percent of PAF II beneficiaries belonged to lower castes and only 1 percent was considered non-poor. 44. PAF intervention contributed to empowering women economically, socially and politically. According to CIE (2018), active participation of women in CO meetings resulted in increased confidence and self-respect. This in turn promoted women’s control over resources as well as improved their participation in decision-making processes in their households (such as those related to property buying and selling). Based on field visits and qualitative assessment carried out by the 2018 CIE and ICR teams, the formation of COs, social mobilization and capacity building activities were main contributors to women’s empowerment. 45. Women in COs had their voices heard, and strengthened leadership and visibility in their local communities. The 2017 election results showed that 1,109 women CO members ran for office in the local election. Of these, 472 were elected to represent their constituencies in various posts. Although it is difficult to attribute these accomplishments solely to PAF, field visits and focus group interviews with women in COs indicated that their leadership was built by their exposure and involvement in PAF. 46. The project was also successful in reaching out to those who were socially excluded. By design, PAF promoted the involvement and participation of ethnic and caste minorities. The focus group discussions carried out under CIE (2018) and the Borrower’s Project Completion Report (2019) showed that PAF engagement helped to increase self-respect and self-dignity especially among the Dalit. Institutional Strengthening 47. The project impacted on institutional development of three types of institutions: (i) COs and POs; (ii) higher level institutions, including CO network and cooperatives; and (iii) PAF as a GoN’s semi-autonomous agency. 48. COs and POs. PAF support was significantly successful in strengthening COs, but some of these were likely to require continued support from the government. Under PAF II, over 27,000 COs were created and supported, and over 400 POs were mobilized to implement PAF activities at the community level. Both COs and POs benefited from capacity building activities aimed at strengthening their capacity to implement development activities. These institutions received capacity on social mobilization, project planning, financial management, accounting, product development and marketing, among many others. 49. PAF conducted a comprehensive CO maturity assessment in 2016 to evaluate the level of sustainability. The results showed that among the 16,334 COs studied, 42 percent were identified as sustainable or close to becoming sustainable with a little help from the PAF. The remaining COs would need continuous support and further capacity building. 50. Higher level institutions. Recognizing that a significant number of COs were likely to continue to need support from PAF, especially in book-keeping of revolving funds, PAF developed a “graduation strategy of COs” in 2016, Page 19 of 64 The World Bank Poverty Alleviation Fund II (P105860) and started to promote CO networks, which are clusters of COs at the district level, and then to strengthen institutional capacity of CO networks through a series of trainings of community resource persons in financial management. A total of 554 CO networks, which support around 8,000 COs, were formed and equipped with their own local resource persons who were expected to provide technical support to CO members especially in book-keeping and financial management and also to link them with local level governments for financial support beyond PAF closure. 51. Further, as per “the graduation strategy”, PAF promoted and encouraged mature COs to form cooperatives and link them with the local cooperative department. PAF promoted 383 cooperatives consisting of 1,105 COs benefiting 19,570 households. They were engaged in various agriculture related productive activities, building on income generation activities and revolving fund established by COs. Selected cooperatives were provided with capacity to nurture their entrepreneurial and management skills, as well as promote business linkages to local and central markets, and financial services. 52. PAF as semi-autonomous agency. PAF was created by a Parliament Act in 2004 as a semi-autonomous agency to address rural poverty and social exclusion using a CDD instrument. PAF has been successful in maintaining its administrative cost at 4 percent of the project cost throughout its implementation. It has also contributed to the GoN when necessary, such as to address the food price crisis in 2008 and earthquake emergency needs in 2015. However, it was obvious that the institutional strength of the PAF declined gradually as the PAF expanded in size and scope without institutional reform. It also faced difficulty in managing more than 400 POs across the country. Further, after the CIAA case, an institutional review conducted by the Bank team in 2016 revealed weaknesses in staff motivation, organizational culture and management style. It also identified the dependence on donor resources as a weakness of the institution. 53. In November 2018, just a month before project closure, the Bank was informed that the Cabinet had endorsed a Plan of Action for closure of the project and for integrating the poverty alleviation agenda within the GoN’s program. The core function of PAF was transferred to the Ministry of Land Management, Cooperatives and Poverty Alleviation (MoLMCPA) and the respective decentralized governments. As of writing this ICR, the PAF Secretariat still functions with minimal number of staff funded by the GoN’s budget, and it is expected that the core functions will be gradually transferred to MoLMCPA and local governments. Mobilizing Private Sector Financing 54. PAF raised private sector funding, although limited in size, especially under the second AF, by linking mature COs with financial institutions and livestock insurance schemes. After the 2010 Mid-Term Review and under the second AF, PAF started strengthening institutional development of COs, helping them establish production and marketing cooperatives in the form of pocket area development, and supporting higher level institutions and peri-urban pilots. PAF was successful in establishing and strengthening 632 economic linkages to the market and financial institutions. 55. According to PAF MIS data, 11 PAF cooperatives were able to mobilize NPR 18 million from micro-finance institutions, local banks and commercial banks, and link CO cooperatives with financial institutions. 56. For COs operating livestock activities, PAF supported the linkage with insurance schemes. Through its social mobilizers, PAF was able to raise awareness on the benefits of insurance. According to PAF MIS, promotion by social mobilizers resulted in several COs having their livestock insured – 511 animals were insured, i.e. 238 households insured from risk related to the selected income generating activity. Poverty Reduction and Shared Prosperity 57. The main purpose of the project was poverty reduction and it was successful in improving the living standards of poor and marginalized populations. 58. As detailed in the Outcome section, according to quantitative CIE (Annex 7), 71.4 percent households Page 20 of 64 The World Bank Poverty Alleviation Fund II (P105860) increased their incomes by at least 15 percent against the base year by the end of the project. The project was successful in poverty reduction when measured by “income” of the targeted group. More than 60 percent of PAF beneficiaries increased their food sufficiency. Other Unintended Outcomes and Impacts 59. The institutional building and empowerment process resulted in higher negotiation power. Beyond PAF activities, COs and their members perceived themselves as more capable to negotiate with non-government organizations, donors and local level government institutions. Similarly, by having access to loans at a reasonable rate, poor households started to use their bargaining power to negotiate better conditions with informal moneylenders. Through this economic empowerment process, several beneficiaries gained access to formal banking services that offered regulated interest rates and thus avoided predatory lending. As a result, according to the CIE report, the annual interest rate that the beneficiaries faced had declined significantly compared to that of the baseline and control groups. III. KEY FACTORS THAT AFFECTED IMPLEMENTATION AND OUTCOME A. KEY FACTORS DURING PREPARATION 60. Based on the success of PAF I, PAF II maintained the CDD approach as the core design and a lean central office partnering with local organizations. At the time of appraisal, PAF II was envisaged to use the same targeting mechanisms and beneficiary profiles as PAF I, but extending its scope nationally in a phased manner, increasing by 15 districts each year. 61. The signing of a peace agreement in November 2006, the establishment of a multi-party government and the country’s plans to transition to a representative democracy, generated positive expectations for a successful second phase. At preparation time the GoN was very supportive of the CDD model and the overall development achievements that PAF was able to deliver on the ground. 62. Both PAF I, and other CDD projects in South Asia have provided valuable lessons that guided the design of the project. Given the success of these projects at times of conflict and the role they played in reducing social tensions, the expansion of PAF was identified as a high priority for IDA support under the government development plans. B. KEY FACTORS DURING IMPLEMENTATION Factors Subject to Government/or Implementing Agency Control 63. Throughout implementation, the project progressed positively towards the PDOs. The project gradually expanded with effective targeting, strong community participation and institutions, as evidenced by various data and assessments described in the Outcome section. CO-level fiduciary mechanism was adequate as observed by field visits and Audit Report. 64. While the governance and institutions at CO level were strong and growing, the management of PAF as semi- autonomous agency suffered from its weakening governance over time. Building on the success of PAF I, there was a positive working relationship between PAF and GoN reflected by the high profile of the board members, who were able to persuade the Prime Minister and bureaucrats to make favorable decisions regarding PAF. The autonomous status of PAF was well perceived during that period, especially right after the end of the conflict. Financial management was adequate with some minor issues detected during supervision. Thus, the ratings in implementation and supervision reports (ISRs) between 2010 and 2012 were “Highly Satisfactory” and “Satisfactory”. However, considering that fiduciary case in 2015 (para 66) was charged against these initial years (2010-2012), there is a likelihood that the governance of PAF started to become weakened from initial Page 21 of 64 The World Bank Poverty Alleviation Fund II (P105860) years. As a result, the positive perception slowly eroded during the second half of the project´s life. Major factors considered to have weakened PAF governance are discussed below. 65. Composition of the PAF Board and reporting framework. PAF Board consisted of the Prime Minister as Chair Person, with a Vice Chair and four members appointed by the GoN and one member from the National Planning Commission. It reported to the Office of Prime Minister and Council of Ministers (OPMCM). While it may have suited the conditions at the time of inception during the conflict, PAF suffered from this governance arrangement because of the frequent change in leadership. With the situation stabilizing it is natural that the Prime Minister of the country would not have time to chair Board meetings, which according to the PAF Act needs to be held every two months. As a result, some key decisions of PAF were delayed. Moreover, throughout the second half of PAF II the Board started to lose its strength and convening power and was not as strong as in its earlier years. 66. Overlap of administrative power between the Vice Chair and Executive Director that was endorsed by PAF Act/By-law. As a result, in some cases, the Vice Chair tended to interfere in the day-to-day business, which caused implementation delays. The Bank team was informed that this was an ad-hoc arrangement at the time of inception, when an Executive Director was not appointed and only the Vice Chair was in place. The PAF Act was never amended since inception and this overlap of responsibilities became a serious constraint to efficient implementation and decision-making. 67. Weak fiduciary capacity of PAF and development of Action Plan to strengthen fiduciary system under the new management. One of the major issues that affected the implementation of PAF II was the weak fiduciary capacity, which triggered the investigation of the Commission for Investigation of Abuse and Authority (CIAA – an apex constitutional body of the GoN to investigate and probe cases against persons holding any public office). The initial investigation that started in late August 2014 focused on field visit advances for PAF portfolio managers during the initial financial years (2010, 2011 and 2012). CIAA alleged that some of the field expense receipts were forged, while others did not comply with government standards. The explanation given by PAF regarding these receipts was that in remote areas official receipts were non-existent and thus in the field several receipts had to be produced in the locally available format. The Bank team and ICR team do not have access to the files and information of CIAA investigation. 68. Immediately after the CIAA case, a new Executive Director was appointed in August 2015 and new staff came on board to replace staff who were suspended by the CIAA case. With the new management of PAF and continuous support from the Bank, PAF developed an “Action Plan to improve the fiduciary system, and took preventive measures such as: review and amendment of field travel policy, including direct payments to hotels and a permanent vehicle hire contract rather than an ad hoc one; ceiling on the amount of advance payment as well as strict follow up from the finance department as regards to the advance resettlement. 69. In March 2015, the CIAA charged 14 PAF staff, including the former Executive Director, the former Chief of Finance and Administrative Division, account officers and several portfolio managers with misuse of field travel funds in 2010, 2011 and 2012. In May 2016, an additional 19 PAF staff were charged. In July 2016, the CIAA investigation concluded, and all reviewed documentation was returned to the PAF office. The CIAA forwarded the charges to a special court, where individual court hearings are taking place. In 2018, the special court convicted nine former PAF staff including the former Executive Director. Each former staff was charged for irregularities in field trips ranging from US$ 700 to US$ 3500.3 The Bank was informed that many of invoices and receipts associated with the field trips had been verified during the court process and the original charged amount was significantly reduced. At the time of writing of this ICR, the case has been brought to the Supreme Court. It was noted during the ICR review that PAF held around 80 staff and disbursed on average around US$ 3 Kathmandu Post, September 2018. Page 22 of 64 The World Bank Poverty Alleviation Fund II (P105860) 20 million per year (out of which 94 percent was directed to the investment to COs in the form of block grants and capacity building) during PAF II. 70. Weakened relevance of semi-autonomous agency during transition to federal structure. The relevance of an autonomous agency was seen as an obsolete while transitioning to federalism and often criticized for bypassing the established local governments. 71. Transfer of PAF core function to line ministry and local level governments under the federal structure. Towards the end of the project, PAF started to intensify the closure of the project. The uncertainty of PAF as a semi-autonomous agency of the GoN remained until a month before closure. At the time of writing of this ICR, the core functions of PAF were being transferred to the MoLMCPA and new local governments, while the PAF Secretariat continues with a small number of core staff funded by the GoN. This is a positive note that the GoN has mainstreamed and integrated the poverty agenda within its program in accordance with the new framework of the federal structure. Factors Subject to Bank Control 72. Overall, the task team leadership remained stable (on average 3–4 years), which helped the team to establish confidence and identify critical needs of the project at the appropriate time. Missions were regularly conducted, and AFs and restructurings were carried out to meet the demands of the GoN to scale up the project. 73. Even though the Bank team experienced several challenges, due to the evolving political environment and the changes in leadership, teams continued to engage in the discussions for proposing new ideas about the project and on how to sustain its positive achievements on the ground. A detailed assessment of the Bank’s performance during implementation is made Section C below. IV. BANK PERFORMANCE, COMPLIANCE ISSUES, AND RISK TO DEVELOPMENT OUTCOME A. QUALITY OF MONITORING AND EVALUATION (M&E) M&E Design 74. The initial design of the M&E and the set of key performance indicators focused mostly on outputs and processes, which resulted in a weak results framework. At appraisal, the results framework was developed without an explicit theory of change. The initial M&E design centered on the number of sub-projects and beneficiaries. While little attention was given to the outcomes that the project activities generated, the indicators ensured that the targeted population was effectively reached. 75. As part of the M&E design and since PAF I, the project conducted multiple rounds of household surveys and focus group discussions to evaluate the outcomes and impacts of PAF activities. However, due to the rapid expansion of PAF, some of the survey samples suffered from attrition while the control groups got lost over time mostly due to their inclusion as beneficiaries (Annex 7). 76. As mentioned earlier in this document, throughout the life of the project the results framework did experience changes in terms of indicators and targets to better reflect progress made towards the PDO. M&E Implementation 77. PAF maintained a MIS covering all project activities. However, there were several problems related to lack of capacity, especially due to expansion of activities, which involved almost 400 POs to report to PAF MIS. Inadequate human resources hindered the proper implementation of the M&E system, particularly after the CIAA investigation caused a high level of M&E staff turnover. The lack of data consistency and accuracy across time was a major issue of project M&E implementation. Page 23 of 64 The World Bank Poverty Alleviation Fund II (P105860) 78. Monitoring of the revolving fund, and the capacity to keep account books at the local level also posed challenges. This task was undertaken mainly by POs, which were not necessarily capable of calculating critical indicators, such as repayment rate. At the later stage of the project, efforts were made to strengthen book- keeping capacity of COs by training community resource persons. M&E Utilization 79. The data on relevant indicators were collected systematically and the results framework was populated and reported to the GoN and the World Bank team. The initial positive results in terms of outputs, targeting and coverage contributed to a decision at Mid-Term Review to expand the project to more districts. Data were also used effectively for ICR and CIE. However, with the high level of human resources turnover and the lack of capacity, PAF could not fully utilize the massive amount of data collected to engage with the GoN or for decision-making on policy. Knowledge Management 80. One activity financed by IFAD was added to Component E at the time of the 2016 restructuring. The objective of this activity was to enhance PAF capacity through knowledge management and information dissemination, including lessons learnt, capturing best practices and successful innovations, and sharing of knowledge among PAF stakeholders. 81. The activity sufficiently achieved the targets set in the results framework and significantly contributed to strengthening the knowledge base, and information and communication materials. It also promoted PAF’s visibility in the media, sharing knowledge and experiences, as well as pioneering actions of COs by offering CO innovation awards for their groundbreaking ideas for poverty alleviation. Justification of Overall Rating of Quality of M&E 82. The overall rating of the M&E is MODEST. Even though the design of the M&E was adequate, there were significant shortcomings in both implementation and utilization of the M&E system. B. ENVIRONMENTAL, SOCIAL, AND FIDUCIARY COMPLIANCE Environmental and Social Compliance 83. Compliance with Environmental and Social Safeguards was rated as “Moderately Satisfactory” in most implementation and supervision reports (ISRs). Towards the end of the project, the safeguards compliance improved significantly with allocation of dedicated staff and continuous training on environment and social safeguard issues at the community level. However, field observations continuously noted room for improvement, especially for safeguard related documentation at the CO level. 84. PAF II was classified as a Category “B” project under OP/BP 4.01 on Environmental Assessment and triggered policy, OP/BP 4.04 on Natural Habitats and OP/BP 4.10 on Indigenous Peoples. The environmental and social impacts sub-projects were limited by the small size of investments that were deemed low risk. An Environmental Management Framework (EMF) developed for managing environmental risks consisted of environmental screening procedures, sectoral environmental guidelines, and a list of sub-projects that the PAF did not support. 85. During implementation, the environmental screening procedure and guidelines were revised to better institutionalize environmental and social concerns in the 30 most common types of sub-projects. The Vulnerable Community Development Plan (VCDP) and Operational Manual were developed to better target vulnerable groups, increase their participation in decision-making processes, as well as identify and mitigate potential risks caused by project activities. The project did not entail land acquisition and resettlement, but the VCDP provided procedures for acquiring voluntary land donations. Page 24 of 64 The World Bank Poverty Alleviation Fund II (P105860) 86. The second AF triggered OP/BP 4.36 on Forests because there was a possibility of some infrastructure sub- projects being in or near community forests; similarly some livestock sub-projects would increase pressure on community or government forests. In 2017, PAF revised the EMF further and named it Environmental and Social Management Guideline (ESMG). Thirty-one sector-specific sub-project ESMGs were designed to help sub-projects implement community groups for identification of environmental and social risks, and link mitigation measures during the selection and preparation of plans, and construction of sub-projects. 87. After the second AF, PAF engaged an environmental and social/gender specialist to oversight safeguard management and mainstream gender. With these initiatives, the safeguard compliance improved. PAF organized a series of trainings for portfolio managers and POs to: help internalize existing safeguard procedures, put forward the requirement of completing environmental screening and prepare any needed mitigation measures for the sub-project. Capacity building within PAF and POs on environmental and social safeguards was significant but had scope for improvement. 88. While a database for land acquisition records was established by the project, proper documentation of land donation and public land use was observed to be weak. The role of POs/portfolio managers was insufficient in disseminating environmental safeguard concerns to COs and linking safeguard concerns with mitigation measures during the selection, monitoring, preparation of detailed project reports (DPRs) and construction of sub-projects. 89. Under the second AF, the grievance redressal mechanism (GRM) was institutionalized. The GRM guideline was developed and GRM committees were established in 43 of the 55 project districts. These committees are in the process of being handed over to the local government during the federal transition with a revised guideline. 90. In preparation for project closure, and transfer of core responsibilities to new local level governments, the Bank team reviewed the environmental and social safeguard status and reported that 700 copies of ESM (earlier ESMF – Environmental and Social Management Framework) guidelines had been printed and will be distributed to 553 local level governments where PAF sub-projects were implemented. This guideline has been disclosed on the PAF website and will be disseminated to local level governments for its operational use in addressing environmental and social issues in the construction of small sub-projects. However, given the capacity of the newly installed local level governments some concerns remain in terms of safeguard compliance for the outstanding infrastructure sub-projects financed under PAF II. Fiduciary Compliance 91. Financial management. PAF’s financial management performance was rated as “Moderately Satisfactory” in most ISRs, except for the time immediately after the CIAA investigation when the rating was “Unsatisfactory” or “Moderately Unsatisfactory”. With intensive support from the Bank, the fiduciary performance improved towards the end of the project. 92. During implementation, there were significant delays and quality issues related to submission of periodic trimester implementation progress reports, including interim unaudited financial reports, due to the weak capacity of finance department staff and their frequent turnover. The project also faced a delay in the appointment of an internal auditor, which is considered to have contributed to systematic internal control weaknesses and lack of timely corrective measures. In general, the Bank’s supervision assessments and internal audit reports highlighted internal control weaknesses, delays in settlement of advances, and a lack of proper registration of expenses. 93. Other weakness observed included : (i) main and subsidiary grant registers not updated; (ii) source wise expenditure information not maintained; (iii) several unreconciled items noted in the bank reconciliation statement without identification and adjustment; (iv) mistakes in main and subsidiary grant registers; (v) Page 25 of 64 The World Bank Poverty Alleviation Fund II (P105860) weaknesses in monitoring financial and physical progress for payment certification; (vi) outstanding advances in non-operating accounts; (vii) internal audit reports not received on time; (viii) new building being constructed without obtaining necessary clearances; (ix) delay in reimbursement to the Government treasury; (x) checks not issued on time; (xi) lack of grants reconciliation with IDA; (xii) delay in accounting of advance received from IDA; (xii) weakness in asset management; (xiv) payments made to some POs long after their contracts expired; (xv) no physical verification of fixed assets; and (xvi) lack of compliance with past agreed actions within agreed timelines. 94. As mentioned above, immediately after the CIAA case, PAF made a series of efforts to take preventive measures and strengthen its fiduciary system. With intensive support from the Bank and regular follow ups, these fiduciary and control weaknesses gradually abated over time. These measures included: (i) updating of the grant register; (ii) booking of source wise expenditure as per the grant agreement; (iii) reconciling most of the unreconciled items; (iv) obtaining and scrutinizing appropriate documents during payment procedure; (v) ensuring proper record keeping of assets; and (vi) verification of physical assets. 95. And as a result, the audit report observed positive improvements in 2016 and 2017, and the financial management rating was upgraded to “Moderately Satisfactory” in January 2017. 96. In general, the GoN approved budgets for the project in a timely manner, allocated adequate funds (20 percent of the PAF operating costs) from the GoN resources. 97. CO level fiduciary mechanism: During implementation, the Bank team made regular visits to COs and observed CO-level fiduciary performance, and in general found the COs keeping their own bank accounts in a transparent manner. The Audit Report also recognized over years the strength of fiduciary mechanism owned and managed by COs (Audit Report 2015 and 2016). According to the independent assessment as regards to the performance of the CO’s revolving fund (MCRIL 2018), the book keeping of the revolving fund has been done adequately with peer-monitoring mechanism. However, the assessment pointed out that the COs, especially in remote areas, had tendency to depend on the technical support from POs for book keeping. This finding is also consistent with CO maturity assessment (para. 49). Towards the end of the project, as described in para 49 and 50, PAF with support from the Bank made a series of efforts to strengthen fiduciary system at CO level. This included formation of higher level institutions including CO network or to Cooperatives. Through these higher level institutions, around 9,000 COs will continue to be supported for book-keeping beyond the project closure. The Bank team was informed that the rest of COs might be further supported by local governments to form higher level institutons (para 98. Procurement. PAF performed satisfactorily in the implementation of the procurement plan agreed between the Borrower and the Bank. A Project Operation Manual covering procurement management was developed, and agreed procedures were followed in project procurement management. The overall procurement management was effective as the project had a dedicated procurement officer. Procurement activities mainly involved hiring of a large number of individual consultants by the PAF for a short duration. Frequent complaints on irregularities in consultant selection were lodged to the oversight agencies of the GoN. Reviews of the selection process for such complaints could not establish irregularities in the selection process. 99. PAF COs also carried out minor procurements during implementation of the agreed programs for COs. Non- government organization partners assisted in the formation of COs and guided in the execution of CO programs. The COs followed “production and operations management” procedures for small value procurements. During filed verification most COs were observed to lack capacity to properly maintain procurement registers and files. 100. In the initial phase of PAF II, some contracts were declared mis-procured following ex-post review of sample contracts. All these contracts were of small value. The Borrower reimbursed the amount that was disbursed by the Bank against the mis-procured contracts. Page 26 of 64 The World Bank Poverty Alleviation Fund II (P105860) C. BANK PERFORMANCE Quality at Entry 101. Given the positive results of PAF I and based on the growing demand from rural communities to reach out to the rest of the country, the design of PAF II maintained its PAF I modality and components. The Bank successfully ensured that PAF II was well aligned with both the World Bank and Nepal’s development agenda at entry (as reflected in the 2003 CAS, Nepal’s 2002–2007 Poverty Reduction Strategy Papers and the ISN 2007). 102. At entry, different models were also considered, however PAF’s CDD modality was selected due to its effectiveness in targeting and results on the ground (PAD, p.4). The Bank also incorporated knowledge from past and ongoing development programs using the CDD model globally and in the region. 103. At entry, the Bank should have had a clear consolidation/graduation strategy for COs. The project was designed with an expectation that it would continue to expand. Quality of Supervision 104. Throughout the life of the project, the Bank provided, to the extent possible, support needed to address factors that affected project implementation. 105. Even though the Bank team experienced several challenges, due to the evolving political environment and changes in leadership, it continued to engage in discussions proposing new ideas about the project and on how to sustain its positive achievements on the ground. It is evident that the team was aware of the institutional capacity limitations and worked closely with the PAF to close those gaps. Throughout the project, the Bank provided dedicated technical support to enhance M&E capacity, especially in updating the MIS and strengthening impact evaluation analysis. 106. There were four critical issues which the team faced during implementation: (i) opportunities to modernize project design and to strengthen intuitional and governance framework at the time of the two AFs and restructuring; (ii) development of consolidation/graduation strategy for COs towards the end of project; (iii) PAF II project closure process, future options and transfer of core functions to local governments; (iv) action plan to strengthen the fiduciary system. The following section describes how the team tackled these challenges and what the team could have done more. 107. Opportunities to modernize the project and strengthen PAF governance. The Bank might have missed opportunities at entry of two AFs to align project design so that PAF could remain relevant in the post-conflict context. The two AFs could have been used to reformulate the project and adapt its components and modality to post-conflict country needs and priorities. Similarly, the Bank could have used the opportunity of AFs to engage the GoN to amend the PAF Act/By-law so that PAF could update its institutional and management framework. 108. Throughout the project implementation, the Bank teams made an effort to modernize the project, including the review of PAF Act/By-law based on the results of the Mid-Term Review. Several times the team raised the need to reform the PAF By-law (which regulates the PAF institutional and management framework and could have been amended with fewer processes) to improve its institutional efficiency. However, considering the changing political leadership and the uncertainty around the project, as well as the GoN’s demand for expansion, it was difficult for the Bank team to redefine the project direction or to make siginificant changes. The Bank team therefore applied the same approach and direction as exercised during PAF I 109. Action plan and capacity building support to strengthen fiduciary system. Another challenge that the Bank faced was the CIAA investigation and the need for constant support to strengthen the fiduciary system. The ICR team held extensive discussions with former and current task teams and analyzed whether preventive Page 27 of 64 The World Bank Poverty Alleviation Fund II (P105860) measures were possible prior to the CIAA case and whether the team provided adequate support to strengthen the fiduciary system after the CIAA case.  Prior and during the CIAA case: The CIAA cases were against the misuse of travel funds during 2010, 2011, 2012, when the project was progressing well on the ground and the performance was rated “Highly Satisfactory” and “Satisfactory”. Interviews with former task teams indicated that, prior to the CIAA investigation, adequate fiduciary oversight had been provided, including supervision missions, audit report review and subsequent follow up. The Bank team carried out fiduciary reviews based on samples where anomalies detected later by CIAA were not found. In addition, the team periodically carried out detailed fiduciary reviews and based on samples that were reviewed, control weaknesses were noted through an action plan to rectify such issues. During 2010–2012, there were some minor concerns (such as delay in accounting for travel expenses), which were documented in the aide memoire (AM) and audit reports but these deficiencies were not reported. The World Bank task team/ICR team could not access any CIAA reports, and hence it was not possible to further understand the case. It is to be understood that it is not possible to review all invoices of hotels and travels, but a sample review did not spot such deficiencies.  Post-CIAA case: Immediately after the CIAA case, the Bank escalated the case to the appropriate level within the Bank. The Bank conducted several fiduciary reviews that did not identify any major irregularities. The Bank nonetheless continued to closely monitor project implementation through to its completion. Moreover, the Bank team downgraded the fiduciary rating to “Unsatisfactory” and the project fell into “problem” status. The Bank team intensified its support to fiduciary capacity building and provided dedicated technical support to strengthen the fiduciary system. It took almost 1.5 years to bring the project back on track through intensive support to strengthen the PAF fiduciary system in accordance with an “action plan”. As a result of preventive measures and the team’s close monitoring and support, the Office of the Auditor General (OAG) reported positive improvements in 2016 and 2017, and the financial management rating was upgraded to “Moderately Satisfactory” in January 2017, and the project came out of “problem” status.  Apart from the CIAA case, during implementation, the Bank received a few corruption complaints, and conducted several fiduciary reviews of various transactions. Even though these efforts did not substantiate the complaints or other information received, the Bank continued to closely monitor project implementation through to its completion. 110. Development of a graduation/consolidation strategy towards the end of the project. The original design of the PAF II did not envisage a “consolidation/graduation” of COs. The project kept expanding and creating new COs even towards the end without adequate resources for consolidation of results. In the absence of an established consolidation strategy and as per second AF design, PAF was fully engaged in establishing new COs until the end of the project. 111. The Bank could have developed a “graduation/consolidation” strategy of COs at early stage, rather than expansion to new districts so that their sustainability could be enhanced. Instead of expanding to new districts, the Bank could have engaged the GoN and PAF at the time of AFs to focus on the consolidation. At the time of second AF (2013), the Bank team added activities under Component C to support mature COs to form cooperatives. However, time and funding were not adequate to cluster all COs under cooperatives. Further, at the time of the first restructuring of second AF (2016), an attempt was made to negotiate with PAF and GoN to cancel new CO formation and divert funds towards consolidation related activities. However, the PAF management did not accept the suggestion as the funds were already committed to the new districts. 112. Nevertheless, towards the end of the project (since early 2016), the Bank team helped PAF to develop a “graduation strategy” for COs. In accordance to the graduation strategy, PAF allocated limited remaining Page 28 of 64 The World Bank Poverty Alleviation Fund II (P105860) resources for consolidation, including formation and capacity building of higher-level institutions such as PAF cooperatives and CO network (see paragraphs 49 and 50) and nurturing community resource persons to work with CO network so that they can continue to support COs beyond the project. 113. Engagement of PAF and GoN in the project closure, future options and the transfer of poverty agenda to local governments. Since three years before the project closure, the Bank team communicated clear message to PAF that IDA would not be able to support PAF as per the existing modality beyond December 2018 because the Bank’s CPF could no longer support the CDD approach. Towards the end of the project (since early 2016), the Bank team reiterated the importance of sufficient preparation of the project closure to ensure all on-going activities should be completed in a timely manner. The Bank team assumed that PAF as an institution of the GoN would continue beyond the project closure. In October 2017, the PAF Board formed a high-level committee as per the instructions of the Prime Minister and Chief Secretary to discuss and propose the future strategy/options of the PAF beyond December 2018. The committee comprised the Vice Chair of the National Planning Commission, Secretary of Ministry of Finance, OPMCM and Ministry of Law. The Bank team supported the PAF in preparing a preliminary future strategy and options, program costs and business plan for the next 10 years, which proposed a few options for more efficient delivery mechanism and a revised focus on rural enterprise development, job creation, entrepreneurship support. The institutional options included (i) an apex micro-finance organization, so that it can cover at least its operational cost; and (ii) the transfer of PAF functions to a line ministry. 114. While supporting PAF in exploring future options, the Bank engaged the PAF to prepare for “clean closure” of the project assuming that local governments will take over the core mandates of poverty alleviation as per the Constitution. PAF developed inventories for community infrastructure and for revolving funds to be handed over to local governments. Nevertheless, the Bank team towards the end of the project faced difficulties in clean closure of two outstanding activities – community infrastructure and communication to COs on project closure – because there was a strong expectation from PAF that it would continue. 115. The Bank team had frequent missions towards the end of the project and discussed the “clean closure” processes with Office of Prime Minister and Council of Ministers (OPMCM) to ensure smooth transition. In November 2018, just a month before the project closure, the Bank was informed that the Cabinet had endorsed a Plan of Action for closure of the project and for integrating the poverty alleviation agenda within the GoN’s program. The core function of PAF was transferred to the MoLMCPA and the respective local governments after July 16, 2019. PAF activities would be monitored, supervised and coordinated by a division of MoLMCPA after shifting the responsibility for the sustainability of COs and small infrastructures to the division. Justification of Overall Rating of Bank Performance 116. Based on the moderate shortcomings in ensuring the quality at entry and quality of supervision, the overall World Bank performance is rated as MODERATELY SATISFACTORY. D. RISK TO DEVELOPMENT OUTCOME 117. The key risks to the development outcome of PAF II are related to concerns over the sustainability of COs established by the project, and transfer to new local level governments under the coordination of the MoLMCPA. 118. At the end of the PAF II project, there was significant uncertainty regarding the future of PAF operations and functions. A month before the project end, the Bank was informed that the core function of the PAF Secretariat would be transferred to MoLMCPA in accordance with the Cabinet decision. As per the Constitution, the implementation of poverty related programs falls under the mandate and jurisdiction of the provincial and Page 29 of 64 The World Bank Poverty Alleviation Fund II (P105860) local level governments, while the MoLMCPA is expected to coordinate a poverty-related agenda at the central level. 119. Consequently, all staff of the PAF Secretariat funded by the PAF II project were formally informed of the termination of the contract. In December 2018, a new Executive Director (seconded from the GoN) and a new Vice Chairperson came on board, and the PAF Secretariat remained with a small number of core staff. Discussions on the amendment and ultimate repeal of the PAF Act are ongoing in accordance with the Plan of Action. Until the PAF Board makes a decision, the PAF Secretariat will continue as a semi-autonomous entity legally, and the PAF Act will stay on. 120. At the time of writing this ICR, all documents and records on COs were transferred to new local level governments who were expected to continue to support the selected COs under the overall coordination of the MoLMCPA. Accordingly, an official communication from the GoN was sent to respective local level governments on the revised implementation modality and their responsibility. The Bank was also informed that discussions would continue to determine the institutional modality for COs, which may be federated into cooperatives. 121. While it is encouraging to note the GoN’s strong ownership of PAF activities and continued commitment to poverty alleviation, there is still uncertainty and risk that both changes in administration and legislation have the potential to negatively affect development outcomes at the community level due to low capacity of new local level governments. There is a particular concern over newly created COs under the second AF, which may not be sustainable due to the relatively short-term support from PAF (two–three years). V. LESSONS AND RECOMMENDATIONS 122. As a long-standing CDD project, aimed at reaching poor and marginalized populations, PAF provides several lessons and recommendations for similar projects in the future. 123. Lesson 1: A demand driven approach and direct funding are effective in keeping target communities at the forefront and reaching out to the intended beneficiaries . PAF encouraged communities to take initiatives to improve their livelihoods, particularly in organizing themselves into COs. PAF provided resources directly to the poor, thereby bringing in efficiency and transparency to the activities and ensuring community ownership. PAF’s targeting strategy was particularly effective as evidenced by data. PAF has effectively used local knowledge and expertise by working with local NGOs. This was important considering Nepal’s diverse ethnicity and languages. 124. Recommendation 1: Future projects should continue to adopt the demand driven approach that allows intended beneficiaries to willingly participate in the project, thereby increasing their sense of ownership and enhancing institutional sustainability. In terms of targeting, it is critical to utilize local knowledge and expertise by working closely with local resource persons and NGOs. 125. Lesson 2: Saving mobilization is effective for initiating revolving funds and has the potential to leverage funds from financial institutions. At the inception stage, direct funding under Component 2 was given as a “grant” for emergency aid. However, COs themselves gradually felt that the grant should be owned by the communities and revolved so that their income generating activities could continue to grow. Community organization members gradually started the practice of monthly savings. Almost all newer COs were accumulating savings with the amount being determined by the members. Some COs that experienced Page 30 of 64 The World Bank Poverty Alleviation Fund II (P105860) financial transactions through the revolving fund successfully formed cooperatives and leveraged funds from financial institutions to invest in large income generating activities. 126. Recommendation 2: Future projects with a similar approach should ensure to integrate regular savings for revolving funds. After the initial process of mobilization of community level institutions, projects should be encouraged to link them with financial institutions for larger loans and investments. 127. Lesson 3: The institutional design of a semi-autonomous agency, which was established with a specific mandate (including, project implementation), should be critically reviewed, revisited and reformed in accordance with the changing environment and needs. PAF as a semi-autonomous agency proved to be appropriate during the time of conflict under PAF I, however during PAF II several shortcomings were exposed. PAF governance structure and policies became less relevant to the GoN’s plans especially during transition to federalism. Implementation delays and management difficulties resulted due to the Board being chaired by the Prime Minister and overlap of administrative powers between the Executive Director and Vice Chair. The lack of a strong human resources system and institutional development plan made it difficult to recruit qualified technical staff, especially after the CIAA investigations. This institutional weakness proved to be especially hindering for the fiduciary and M&E systems. 128. Recommendations 3: The institutional design of an implementing agency should be critically reviewed over time to ensure proper work flow of the decision-making process, and to establish a strong human resources system to continue to attract qualified staff, especially in a rapidly changing environment. Given the difficulty to attract qualified fiduciary staff in countries like Nepal, governments in the future could consider seconding their own fiduciary staff to semi-autonomous agencies to ensure consistency with the government’s fiduciary system. 129. Lesson 4: The institutional design of a semi-autonomous agency should allow it to mobilize funds to cover operational costs to enhance the sustainability of the agency. During PAF II implementation, Nepal graduated from an IDA grant recipient country, while the GoN was clear on IDA Credit should be directed to larger investment opportunities, such as infrastructure. The first AF attempted to combine IDA “Credit” with IDA Grant at Appraisal but at negotiation GoN declined IDA credit. The second AF was the last IDA grant for the GoN. 130. For over a decade, IDA grants from the World Bank and IFAD grants solely funded the PAF. PAF was aware of the urgency to consider future funding and diversify its funding sources, at least to cover its operational cost while seeking for programmatic investments from elsewhere, including GoN resources. The existing PAF Act had provisions to mobilize assistance, grants and other necessary financial resources from government/non- government organizations, foreign governments of institutions, international institutions or agencies or persons, to conduct programs in consonance with its objectives. Towards the end of the project, PAF developed a future strategy/options (para. 113) that included an option for it to become an apex micro- finance organization, so that it can cover at least its operational cost. While the idea of PAF’s having capacity to mobilize resource was well received by the GoN, the proposal for having a continued semi-autonomous agency was not well accepted by the GoN in the context of transition to a federal structure. 131. Recommendation 4: Future projects should be designed to enable semi-autonomous agencies to become sustainable on their own beyond project closure, and to identify other options to become, to the extent possible, a sustainable agency rather than being permanently dependent on donor funds. Such options could include a wholesale agency for micro-finance, or service providers for the government with expertise and experience in business development, or to become a flexible, effective and responsive development and emergency support platform. For instance, Pakistan Poverty Alleviation Fund (PPAF), which also started as the Page 31 of 64 The World Bank Poverty Alleviation Fund II (P105860) Bank Project, and yet is a private sector institution independent and autonomous in its operation and allocation of resources, and works alongside the public policy to bring services and products to the grass-root. PPAF has its own micro-finance bank. 132. Lessons 5: CDD approach can be integrated into the federalism and decentralization process. As the country stabilized and local level governments were in place, PAF’s independence was perceived as bypassing line ministries, local governments and other development programs and competing with them in terms of resources. During last year of the Project, PAF intensively consulted with new local and provincial governments. There were local governments who were willing to adopt PAF’s CDD approach and requested its technical assistance to develop their local development plans. Due to the shortage of manpower and lack of funds, PAF could not adequately respond to these demands from the local governments. However, this indicates that CDD approach will remain relevant in the federal context for poverty alleviation with effective targeting. PAF has a comparative advantage with its extensive network across the country and could continue to contribute to GoN’s poverty alleviation mandate by working closely new local governments through technical support, even if not implementing projects directly. 133. Recommendation 5: When countries move to a federal system, CDD approach could be incorporated in the format of technical assistance to local governments for their local development plan. CDD approach can be mainstreamed and linked to local government activities. Specific efforts must be made to facilitate this process in order to guard the sustainability of these community level institutions established by the projects. 134. Lesson 6: Clear consolidation/graduation plan should be in place at entry to enhance sustainability of community level institutions. Based on the satisfactory performance of PAF I and GoN’s initial interest it seemed natural to scale up the project. Continuing with the initial design proved to have benefits, however, under PAF II it proved to be challenging due to the changing environment in Nepal. Moreover, it should be recognized that the project funding from development partners is time-bound and would not last permanently. The lack of a consolidation strategy built into the design, and lack of vision about the future and how beneficiaries could graduate from the project all prevented a proper and clean closure. 135. Recommendations 6: Before scaling up rapidly successful interventions of a smaller pilot, a graduation/consolidation strategy should be embedded in the project design. During the design phase of the scaling up, it is necessary to ensure that proper information and human resources are in place or will be acquired to ensure development effectiveness. A clear consolidation strategy should be developed upfront and carried out to: consolidate the results at an early stage and strengthen community level institutions so that they become economically and institutionally sustainable; and to create higher-level institutions which cluster community level institutions so that they can continue to support them beyond the project closure. 136. Lesson 7: Long-standing projects working in the changing socio-political environments should require specific attention to adapt project design and implementation over time in order to maintain consistency and relevance to the evolving development challenges. 137. Recommendation 7: In this type of environment, projects need to be designed and implemented with more flexibility in mind, which will allow the project to better adapt to the evolving environment. Having more frequent high-level meetings to discuss the relevance of the design in light of socio-political changes could help projects progress in line with the new institutional demands. 138. Lesson 8: Fiduciary issues should have adequate attention at high level of the Project management and corrective measures should be taken and should be continuously monitored and followed up . PAF suffered Page 32 of 64 The World Bank Poverty Alleviation Fund II (P105860) from weak fiduciary system which triggered the CIAA investigation, which subsequently demanded intensive efforts and technical assistance to bring the project back on track. While the ICR does not intend to suggest that the CIAA case could have been prevented, it does observe that fiduciary issues have to be raised to the attention of highest level of project management and governance. 139. Recommendation 8: The project and the Bank should prioritize the fiduciary issues as a core management responsibility and should develop a workplan for consistent follow up in case there is no progress nor remedies for specific concerns and issued raised at missions. . Page 33 of 64 The World Bank Poverty Alleviation Fund II (P105860) ANNEX 1. RESULTS FRAMEWORK AND KEY OUTPUTS A. RESULTS INDICATORS A.1 PDO Indicators Objective/Outcome: Outcome 1: Improved household living conditions. Formally Revised Actual Achieved at Indicator Name Unit of Measure Baseline Original Target Target Completion Number of households Number 0.00 15000.00 164000.00 317404.00 benefitting from increased access to community 16-Jun-2008 31-Dec-2007 31-Jul-2013 31-Dec-2018 infrastructure Comments (achievements against targets): Number of households includes Non-CO members as well No of beneficiaries surpassed the original targets because of infra increased and CO federation, which made COs to opt for larger infrastructure Objective/Outcome: Outcome 2: Improved household livelihoods. Page 34 of 64 The World Bank Poverty Alleviation Fund II (P105860) Formally Revised Actual Achieved at Indicator Name Unit of Measure Baseline Original Target Target Completion Percentage of beneficiary Percentage 0.00 70.00 71.43 households have increased their incomes by at least 15% 31-Dec-2007 31-Dec-2007 31-Dec-2018 against base year (2007) by the EOP Comments (achievements against targets): Objective/Outcome: Outcome 3: Improved levels of participation and empowerment. Formally Revised Actual Achieved at Indicator Name Unit of Measure Baseline Original Target Target Completion Percentage of key positions Percentage 0.00 60.00 62.80 in Project community organization that come from 31-Dec-2007 31-Dec-2007 31-Dec-2018 targeted groups Comments (achievements against targets): Targeted households include Dalits, Janajati and Muslims Objective/Outcome: Outcome 4: Marginalized population reached. Formally Revised Actual Achieved at Indicator Name Unit of Measure Baseline Original Target Target Completion Page 35 of 64 The World Bank Poverty Alleviation Fund II (P105860) Direct project beneficiaries Number 0.00 575000.00 589000.00 776156.00 16-Jun-2008 31-Jul-2013 29-Feb-2016 01-Jun-2018 Female beneficiaries Percentage 0.00 75.00 80.00 Indirect project Number 0.00 55000.00 17916.00 beneficiaries Comments (achievements against targets): A.2 Intermediate Results Indicators Component: Component A. Small Scale Village and Community Infrastructure Formally Revised Actual Achieved at Indicator Name Unit of Measure Baseline Original Target Target Completion Number of infrastructure Number 0.00 4000.00 4449.00 sub-projects that are completed with target 31-Dec-2007 31-Jul-2013 31-Dec-2018 community participation, according to agreed design and quality standards Comments (achievements against targets): Page 36 of 64 The World Bank Poverty Alleviation Fund II (P105860) Formally Revised Actual Achieved at Indicator Name Unit of Measure Baseline Original Target Target Completion Percentage of infrastructure Percentage 0.00 70.00 69.00 sub-projects with a functional O&M system 31-Dec-2007 29-Jul-2011 31-Dec-2018 Comments (achievements against targets): Definition of O&M system 1. Establishment of O&M fund 2.O&M Mechanism Formally Revised Actual Achieved at Indicator Name Unit of Measure Baseline Original Target Target Completion PAF supported Number 0.00 288.00 256.00 219.00 Infrastructures rehabilitated 01-Feb-2016 16-Feb-2016 30-Nov-2017 31-Dec-2018 Comments (achievements against targets): Initially, Rapid assessment shows 233 Partially damaged, 57 completely damaged. Later, during detailed assessment only 256 projects identified, hence target was revised to 256 at the 2nd restructuring. Page 37 of 64 The World Bank Poverty Alleviation Fund II (P105860) Component: Component B: Income Generating Sub-projects Formally Revised Actual Achieved at Indicator Name Unit of Measure Baseline Original Target Target Completion Percentage of IGA Percentage 0.00 54.00 57.00 community organization members who belong to 31-Dec-2007 31-Jul-2013 31-Dec-2018 targeted HHs. Comments (achievements against targets): Formally Revised Actual Achieved at Indicator Name Unit of Measure Baseline Original Target Target Completion Percentage of CO members Percentage 0.00 50.00 60.00 accessing funds from the RF more than one time for IGAs 31-Dec-2007 31-Dec-2007 31-Dec-2018 during the lifetime of the project Comments (achievements against targets): CO members who take loan 2 and more time are 60% CO members who take loan At least 1 time are 90% Formally Revised Actual Achieved at Indicator Name Unit of Measure Baseline Original Target Target Completion Page 38 of 64 The World Bank Poverty Alleviation Fund II (P105860) Percentage of CO members Percentage 0.00 68.00 41.50 with improved levels of food security 31-Dec-2007 29-Jul-2011 31-Dec-2018 Comments (achievements against targets): Formally Revised Actual Achieved at Indicator Name Unit of Measure Baseline Original Target Target Completion CO members benefited from Number 0.00 40000.00 14000.00 11578.00 asset/cash transfer for RF revitalization 01-Feb-2016 09-Feb-2016 30-Nov-2017 28-Dec-2018 Comments (achievements against targets): The original estimate indicated that hhs of 4900 Cos and around 40,000 CO members need asset/cash transfer support for revitalizing the RF.Detailed analysis indicated that the hhs of 1500 Cos (around 14,000 hhs) were eligible Component: Component C: Product Development, Market Linkages and Pilots Formally Revised Actual Achieved at Indicator Name Unit of Measure Baseline Original Target Target Completion Number of higher level Number 0.00 403.00 383.00 institutions formed 31-Mar-2013 31-Jul-2013 31-Dec-2018 Comments (achievements against targets): Page 39 of 64 The World Bank Poverty Alleviation Fund II (P105860) Includes only cooperatives (multi and single purpose) which are already registered and not CO federations or CO networks Formally Revised Actual Achieved at Indicator Name Unit of Measure Baseline Original Target Target Completion Number of pocket areas Percentage 0.00 60.00 30.00 30.00 developed and supported by the project for collective 31-Mar-2013 31-Jul-2013 30-Nov-2017 29-May-2018 commercial production and marketing of commodities Comments (achievements against targets): Due to the shortage of the remaining funds and also to spend adequate time to monitor the results of the first 30 PA pilots, the target was revised to 30 at 2nd restructuring Formally Revised Actual Achieved at Indicator Name Unit of Measure Baseline Original Target Target Completion Number of COs functioning Number 0.00 90.00 40.00 40.00 in urban/peri-urban settings 31-Mar-2013 31-Jul-2013 04-Nov-2016 29-May-2018 Comments (achievements against targets): COs functioning mean Cos formed, registered and received grants from PAF Page 40 of 64 The World Bank Poverty Alleviation Fund II (P105860) The fund allocated for this fund was USD 840,000. Later it was discovered that the peri-urban’s costs are higher, and hence the number of Cos were reduced to provide adequate funds. Component: Component D: Capacity Building and Implementation Support Formally Revised Actual Achieved at Indicator Name Unit of Measure Baseline Original Target Target Completion Number of COs mobilized Number 0.00 28298.00 27722.00 and made agreements with PAF 31-Dec-2007 31-Jul-2013 31-Dec-2018 Comments (achievements against targets): Formally Revised Actual Achieved at Indicator Name Unit of Measure Baseline Original Target Target Completion Skilled persons trained in Number 0.00 14300.00 4500.00 4535.00 earthquake resistant construction techniques 01-Feb-2016 10-Feb-2016 30-Nov-2017 01-Dec-2018 Comments (achievements against targets): This indicator was added at 1st restructuring as EQ component in February 2016. 9,000 masons, 5900 carpenters are identified in the project area. Original target was 14300. Initially, the training was planned for 4 days, Page 41 of 64 The World Bank Poverty Alleviation Fund II (P105860) Due to the guideline from the Government of Nepal(Department of Urban Development and Building Construction (DUDBC), which has mandate of 7 days masons training, the training event was adjusted to 180 events (for 25 Participants/event), targeting 4500 participants (through the Level 2 restructuring approved in November 2017) Component: Component E: Project Management, Planning and M&E Formally Revised Actual Achieved at Indicator Name Unit of Measure Baseline Original Target Target Completion Grievances registered related Percentage 87.50 94.00 96.00 to delivery of project benefits addressed (%) 01-Nov-2012 25-Jun-2014 31-Dec-2018 Comments (achievements against targets): Formally Revised Actual Achieved at Indicator Name Unit of Measure Baseline Original Target Target Completion Number of innovation Number 0.00 125.00 285.00 awards to COs 01-Feb-2016 29-Feb-2016 01-Dec-2018 Comments (achievements against targets): Formally Revised Actual Achieved at Indicator Name Unit of Measure Baseline Original Target Target Completion Page 42 of 64 The World Bank Poverty Alleviation Fund II (P105860) Number of Knowledge Number 0.00 272.00 469.00 Management Products developed 01-Feb-2016 29-Feb-2016 01-Dec-2018 Comments (achievements against targets): Page 43 of 64 The World Bank Poverty Alleviation Fund II (P105860) A. KEY OUTPUTS BY COMPONENT Objective/Outcome 1: Improved household living conditions. Outcome indicators 1. Number of households benefitting from increased access to community infrastructure. 1. Number of infrastructure sub-projects that are completed with target community participation, according to agreed design and quality standards. Intermediate results indicators 2. Percentage of infrastructure sub-projects with functional O&M system. 3. Number of PAF supported infrastructures rehabilitated (under earthquake rehabilitation activities). 1. 317,404 households benefitted from increased access to community infrastructure 2. 4449 iinfrastructure sub-projects completed with target community participation, Key outputs by component (linked to the achievement of according to agreed design and quality standards. Objective/Outcome 1) 3. 69 percent of infrastructure sub-projects completed with functional O&M system. 4. 219 PAF supported Infrastructures rehabilitated (under EQ component). Objective/Outcome 2: Improved household livelihoods 1. Percentage of beneficiary households that have increased their incomes by at least 15 Outcome indicators percent against base year by the end of the project. 1. Percentage of Income Generation Activity CO members who belong to targeted households. 2. Percentage of CO members accessing funds from the revolving fund more than one Intermediate results indicators time for IGAs during the lifetime of the project. 3. Percentage of CO members with improved levels of food security. 4. Number of CO members benefited from asset/cash transfer for revolving fund revitalization. Key outputs by component 1. 71.43 % of beneficiary households have increased their incomes by at least 15 % (linked to the achievement of Objective/Outcome 2) against base year by EOP Page 44 of 64 The World Bank Poverty Alleviation Fund II (P105860) 2. 57 percent of IGA CO members belong to targeted households. 3. 60 percent of CO members accessing funds from the revolving fund; more than one time for IGAs during the lifetime of the project. 4. 41.5 percent of CO members with improved levels of food security. 5. 11,578 CO members benefited from asset/cash transfer for revolving fund revitalization. Outcome 3: Improved levels of participation and empowerment Outcome indicators 1. Percentage of key positions held in project COs from targeted households. Intermediate results indicators Key outputs by component 1. 63 percent of key positions in project COs are from targeted households. Outcome 4: Marginalized population reached Outcome indicators 1. At least 70 percent of CO members are women. 1. Percentage of income generating activities among CO members who belong to Intermediate results indicators targeted households. 1. 80 percent of CO members are women. Key outputs by component 2. 57 percent of IGA CO members belong to targeted households. Page 45 of 64 The World Bank Poverty Alleviation Fund II (P105860) ANNEX 2. BANK LENDING AND IMPLEMENTATION SUPPORT/SUPERVISION A. TASK TEAM MEMBERS Name Role Preparation Supervision/ICR Mio Takada, Kamran Akbar Task Team Leader(s) Frauke Jungbluth (Former) Task Team Leader Gayatri Acharya (Former)Task Team Leader Chandra Kishor Mishra, Shambhu Prasad Uprety Procurement Specialist(s) Yogesh Bom Malla Financial Management Specialist Kiran Gautam Team Member Ama Esson Team Member Jaya Sharma Social Specialist Purna Bahadur Chhetri Team Member Javier Bronfman Horovitz Team Member Ramesh Raj Bista Team Member Annu Rajbhandari Environmental Specialist Karishma Wasti Team Member Tashi Tenzin Infrastructure Specialist Md. Mansur Ahmed Agriculture Economist Yoshiko Ishihara Social Development and Institution Specialist, FAO Page 46 of 64 The World Bank Poverty Alleviation Fund II (P105860) A. STAFF TIME AND COST Staff Time and Cost Stage of Project Cycle No. of staff weeks US$ (including travel and consultant costs) Preparation FY07 1.525 6,108.54 FY08 16.650 101,094.71 Total 18.18 107,203.25 Supervision/ICR FY08 6.275 28,581.17 FY09 15.882 94,212.42 FY10 25.000 114,144.04 FY11 21.688 114,794.65 FY12 28.187 73,122.69 FY13 42.477 90,309.60 FY14 51.252 143,940.42 FY15 60.825 209,366.00 FY16 46.699 187,377.61 FY17 36.073 130,994.99 FY18 36.567 226,633.70 FY19 24.093 130,682.82 Total 395.02 1,544,160.11 Page 47 of 64 The World Bank Poverty Alleviation Fund II (P105860) ANNEX 3. PROJECT COST BY COMPONENT Cumulative Amount at Amount of IFAD Cumulative Amount at Amount at IFAD co- Actual at Original TF at first supplemental Amount of Percentage of Components first AF second AF financing Project Appraisal AF co-financing Approval Approval Closing (US$M) (US$M) (US$M) (US$M) (US$M) (US$M) (US$M) (US$M) Small-scale village and 77 92.54 community infrastructure 32 28.5 15.7 0.8 71.25 Income generating sub-projects 10 99.56 100.75 37 21.66 29.1 1.8 100.31 Product development, market 19.57 101.07 linkages and pilots 7 3.07 9.5 0 19.78 Capacity building and 4 47.4 88.69 implementation support 19 8.7 14.5 1.2 42.04 Project management, planning 20.47 96.24 and M&E 5 3.07 11.2 1.2 19.70 Total 100 65 10 80 4 5 264 253.08 95.86 Page 48 of 64 The World Bank Poverty Alleviation Fund II (P105860) ANNEX 4. EFFICIENCY ANALYSIS NEPAL: Poverty Alleviation Fund Phase II Overview The Project Development Objective (PDO) of the Nepal Poverty Alleviation Fund Project is to improve living conditions, livelihoods and empowerment among the rural poor, with particular attention to groups that have traditionally been excluded by reasons of gender, ethnicity, caste and location. The project aimed at improving living conditions and livelihoods of the rural poor by: (a) investing in community selected and managed sub-projects that create access to basic socio-economic infrastructure and services, increase assets, generate employment and expand income-generating opportunities in poor villages; (b) enhancing the capacity of local bodies, particularly the Village Development Committees (VDCs), to provide better services for poor and socially excluded groups; and (c) improving government efforts to better coordinate support targeted towards poor and excluded groups. The project had five components: (i) small scale village and community infrastructure; (ii) income-generating sub-projects; (iii) innovations and special programs; (iv) capacity building; and (v) administration of PAF II. This project was implemented between 2008 and 2018 (with two restructurings for Additional Financing (AF) in 2011 and 2013, respectively) under the International Development Association (IDA) financing with co-financing from International Fund for Agricultural Development. This ex-post economic and financial analysis covers both original financing and AFs of the project. The direct economic benefits of the project were achieved through several interventions, such as: (i) improvements in access to small-scale social and economic infrastructure and services, as prioritized by the beneficiary communities; (ii) generation of incremental employment at the village level, including both short-term participation in the implementation of sub-projects and longer-term jobs resulting from economic activities facilitated by PAF; (iii) increase in capital assets and/or incomes of beneficiary households; and (iv) increase in citizen participation and voice in community decision making. The main quantifiable benefits were from improvements in access to small-scale village and community infrastructure and income-generating sub-projects supported by PAF II. In addition, the project’s activities under other components were expected to produce non-quantifiable benefits in the form of: (i) capacity building of local bodies and rural communities; (ii) support to rural and community finance; and (iii) social mobilization of community groups. However, due to the unavailability of reliable data as well as the unpredictability of investment choices, benefits from these investments were excluded from the efficiency analysis. Strong demand-driven nature of the project implementation mechanism made it difficult to perform a traditional ex-ante economic and financial analysis (EFA) of project investments during appraisal. It was not considered relevant to provide an economic rate of return (ERR) for the project, as the exact type of income generating activities was unknown. Therefore, the results from the ex-post economic analysis are not directly comparable with the results from the ex-ante economic analysis. Data collection In addition to the regularly collected MIS data by the M&E team of the project, the ex-post economic Page 49 of 64 The World Bank Poverty Alleviation Fund II (P105860) analysis extensively used data from 400 randomly drawn samples4 from beneficiary households of six income-generating activities and four small infrastructures in two districts in 2017. Prices and wages were updated with recent secondary data. When needed, the analysis also used data from the comprehensive impact evaluation report of the project. The efficiency analysis also used information from various secondary sources that included the Central Bureau of Statistics and the Nepal Rastra Bank. The information used in the economic analysis were crosschecked and verified during field visits of the ICR review mission. Methodology An ex-post EFA is undertaken to assess the financial and economic impact of the PAF to assess its ex-post economic viability. At appraisal, a standard ex-ante EFA was not carried out given the strong demand- driven nature of project activities, and because identification of future project activities for ex-ante EFA would have been considered difficult. Given the lack of comparable financial and economic returns from an ex-ante EFA at appraisal, this efficiency analysis assessed the economic and financial viability of the project based on the economic and financial returns derived from this analysis using actual survey data from 400 randomly drawn beneficiary households. The efficiency analysis was carried out on the basis of individual analysis of selected income-generating activities and infrastructure schemes for which reliable and sufficient data were available to measure project benefits. Returns to investments in selected activities were estimated using their respective production models. The main project benefits were sourced from the promotion of income generating activities and the construction/rehabilitation of key infrastructures with support from the project. The benefits from women empowerment and other spillover effects were not considered. While these benefits are clearly important, they are difficult to quantify. Also, the economic returns from the economies of scale could not be captured due to lack of detailed input and output costs for the non-beneficiary households. Moreover, returns to investments on capacity development for earthquake rehabilitation and program management were excluded from the analysis due to lack of reliable data. Thus, the financial and economic returns estimated in this efficiency analysis were a conservative estimation of the benefits generated by project investment. Key assumptions The final financial cost of the project in 2018 prices was estimated at NPR 24,985 million (equivalent to US$ 244 million). Costs of project components were distributed in the following shares: Component A: small-scale village and community infrastructure (30 percent); Component B: income-generating sub- projects (38 percent); Component C: innovations and special programs (8 percent); Component D: capacity development (16 percent); and Component E: administration of PAF II (8 percent). The economic cost of the project was estimated by removing price contingencies and all taxes and duties from the financial cost. Main source of benefits from the project, accounted to in this analysis, was returns from income generating sub-projects and small-scale village and community infrastructure. The EFA at project completion covered activities of sub-components that accounted for 68 percent of the total project cost. The project impact was estimated by aggregating benefits (with their respective weights) to and costs of 4 This sample survey was carried out to conduct an interim economic analysis of income generating activities and small infrastructures supported by the PAF II project in 2017. The economic analysis stated that project investments for income generating activities were economically robust and viable. Page 50 of 64 The World Bank Poverty Alleviation Fund II (P105860) those selected investments for which benefit streams were estimated. The project benefits were assessed for 15 years at 2018 financial prices, using the opportunity cost of capital at 10 percent. Financial prices of locally traded outputs and inputs were converted into economic prices by deducting direct subsidies and taxes. Economic import parity prices were calculated for cereals and fertilizers using international reference prices, transport, marketing and other costs of importing through Kolkata, India. Moreover, for non-tradables a standard conversion factor of 0.9 was applied to the local price component. The opportunity cost of labor was valued at NPR 13,450 per month (around US$ 120) based on the current minimum wage for Nepal. The exchange rate used in the analysis was NPR 102.0 per US$ 1.0. Investment models were prepared for producers/farmers in selected income generating activities (poultry, livestock, retail business and vegetable farming) and small-scale infrastructure (irrigation, micro hydro, water supply and rice mill). The selection of sub-projects was based on the availability of quantifiable benefits and costs in the sources mentioned above. However, the project’s investments in the components of selected sub-projects accounted for more than 68 percent of the total project cost. Financial and economic analysis of selected subprojects Based on the range of income generating activities and micro-infrastructure projects, 10 different indicative activities (six from income generating activities and four from small infrastructure programs: (i) poultry, (ii) cow/buffalo, (iii) goat, (iv) piggery, (v) retail business, (vi) vegetable farming, (vii) irrigation, (viii) micro hydro, (ix) water supply, and (x) rice mill) were assessed in financial and economic terms. The EFA at project completion covered activities of sub-components that accounted for 68 percent of the total project cost. The results showed considerable variations in returns across investments. The returns estimated from investment models showed that the investment in all income generating sub-projects and small-scale infrastructure sub-projects were reasonable and economically viable in terms of both financial and economic returns. While financial returns from income generating activities and small infrastructure investments ranged from 6 percent to 26 percent, economic returns from these investments ranged from 13 percent to 47 percent. Poultry. The first model represents the income generating sub-projects for poultry. With an investment of NPR 57,000 (US$ 560) in poultry, the financial net present value (FNPV) was estimated around NPR 160,000 (US$ 1600) with the financial internal rate of return (FIRR) and economic internal rate of return (EIRR) of 19 percent and 16 percent, respectively. The internal rates of return of more than 10 percent implied that investment for income generating activities in poultry is economically viable. Livestock-cow/buffalo. With a total investment of NPR 60,000, the FIRR was 21 percent and FNPV was NPR 157,000 (US$ 1.5 thousand). The EIRR for the investment in cow/buffalo was 15 percent and the expected net present value (ENPV) was NPR 54,000 (US$ 0.53 thousand). Livestock-pig. Efficiency analysis found that financial returns to investments made in pig production was 20 percent with FNPV of NPR 70,000 (US$ 0.7 thousand). The EIRR was estimated to be 14 percent with ENPV of NPR 20,000 (US$ 0.2 thousand). Livestock-goat. PAF II supported income-generating activities in goat rearing. With the total investment of NPR 63,000 (US$ 0.62 thousand) from the project to the sub-projects in goat production, the FNPV was estimated to be NPR 62,000 (US$ 0.61 thousand) with the FIRR of 17 percent. The EIRR for PAF II investment in goat rearing sub-projects was also found to be 16 percent with the ENPV of NPR 36,000 (US$ 0.35 thousand). Page 51 of 64 The World Bank Poverty Alleviation Fund II (P105860) Retail business. This model presents a representative retail business firm supported by PAF II under income generating sub-projects component. With investment of NPR 70,000 (US$ 0.68 thousand) from PAF II, the FNPV of the investment was NPR 27,000 (US$ 0.27 thousand) and the FIRR to the investment was 18 percent. The EIRR and the ENPV to the investment were 11 percent and NPR 14,000 (US$ 0.14 thousand), respectively. Vegetables farming. Vegetable-producing farmers groups and cooperatives were also prioritized in receiving grants from the PAF II under the component of income generating sub-projects. With the total investment of NPR 20,000 (US$ 0.2 thousand), the FIRR and the FNPV were estimated to be 26 percent and NPR 15,000 (US$ 0.15 thousand), respectively. The EIRR and ENPV for the investment were 19 percent and NPR 8,000 (US$ 0.08 thousand), respectively. Irrigation. Irrigation facilities were greatly supported in project areas. About of 28 percent of total spending under small-scale infrastructure went for irrigation support. The economic returns from such irrigation facilities was substantial. With a total investment of NPR 1 million (US$ 98 thousand), the financial and economic returns from the investment were 16 percent and 22 percent, respectively. The FNPV and ENPV of the investment were found to be NPR 572,000 (US$ 5.6 thousand) and NPR 1,116,000 (US$ 10.9 thousand), respectively. Micro hydro. One of the key small-scale infrastructures supported by the project was the construction of a micro hydro facility. With a total investment of NPR 5.1 million (US$ 0.05 million) in a micro hydro sub- project, PAF incurred FNPV of NPR 8.2 million (US$ 0.08 million) with FIRR of 22 percent. The economic returns from the investment was even higher as the EIRR was 47 percent and the ENPV was NPR 19.1 million (US$ 0.19 million). Water supply. The project also made significant investment in water supply sub-projects under the component for small-scale village and community infrastructure. The average size of PAF II investment in water supply sub-projects was NPR 0.96 million. The FIRR from water supply sub-project was 6 percent. However, the economic returns from investment made in the water supply sub-project was 21 percent. The F/ENPVs of PAF II investment in water supply sub-projects were found to be NPR 0.40 million (US$ 4,000) and NPR 0.9 million (US$ 9,000), respectively. Rice mill. The project also made some investments in a rice mill sub-project under the component for small-scale village and community infrastructure. With the total investment of NPR 0.6 million (US$ 0.05 million), the project incurred FNPV of NPR 0.21 million (US$ 2.0 thousand) with FIRR of 14 percent. The economic returns from the investment was even higher as the EIRR was 15 percent and the ENPV was NPR 0.28 million (US$ 2.8 thousand). Other investments. The project has also made investments in many other sub-projects and income generating activities in improving the living conditions and livelihoods of poor households. Due to lack of sufficient data to quantify the benefits from these investments, the efficiency analysis did not develop an investment model for those activities. The comprehensive impact evaluation report of the project, however, highlighted benefits from some other investments that were not included in this economic analysis. Overall EFA The ex-post economic analysis at project completion found that the financial and economic returns to the project investments in income generating activities and small infrastructures were 18 percent and 17 percent, respectively. The FNPV and ENPV under this scenario were about NPR 14,837 million (US$ 145 Page 52 of 64 The World Bank Poverty Alleviation Fund II (P105860) million) and NPR 11,978 million (US$ 117 million), respectively. The analysis confirmed the project’s financial viability of income generating and small infrastructure investments as measured by FIRR at 18 percent; ERR at 17 percent, with a positive Net Present Value (NPV) estimated at US$ 117 million. Both the economic and financial returns were net of returns from employment as the imputed wage costs were deducted for each investment. Thus, actual benefits would be even higher when the returns from employment were accounted. While the investments captured by the investment models constructed under the efficiency analysis accounted for two-third of total project costs, the other one-third of the project cost, for which benefits were not quantified, were also expected to generate benefits to the beneficiaries and improve overall efficiency of investments through better management. However, under a conservative scenario, the efficiency analysis accounted the total project costs, but benefits from components other than income generating and small infrastructures were not accounted. With this conservative scenario, which deficiently underestimates investment benefits, the project was considered robust because the economic returns were 12 percent with NPV of NPR 4,715 million (US$ 46 million). Therefore, the overall cost-effectiveness of the project was rated SUBSTANTIAL, and it was largely based on conservative estimates of returns from non-quantified benefits of investment. As the project was implemented for a longer period than the original planned project period, the efficiency analysis tested the sensitivity of returns to this delayed implementation and found that both economic and financial returns declined but remained economically viable. Table A4a: Economic and financial returns from the ex-post efficiency analysis Page 53 of 64 The World Bank Poverty Alleviation Fund II (P105860) Investments Actual investment ENPV ('000 NPR) EIRR (%) FNPV ('000 NPR) FIRR (%) ('000 NPR) A. Income generating activities Poultry 57 70 14% 130 17% Livestock-cow- 60 81 17% 149 21% buffalo Livestock-pig 45 66 21% 100 25% Livestock-goat 78 36 16% 54 18% Retail business 70 13 13% 35 17% Vegetable 20 8 19% 15 26% farming Total Income 9,455,616 7,851,027 16% 13,839,583 20% generating activities B. Infrastructure Irrigation 1,000 1,116 22% 572 16% Micro hydro 5,050 19,048 47% 8,244 22% Rice mill 645 285 15% 207 14% Water Supply 955 916 21% (382) 6% Total 7,540,241 9,763,552 24% 1,355,862 13% Infrastructure Total Investment 16,995,857 11,978,133 17% 14,837,188 18% in income generating activities and infrastructure components Overall project 24,985,000 4,715,275 12% 7,574,330 13% Page 54 of 64 The World Bank Poverty Alleviation Fund II (P105860) ANNEX 5. BORROWER, CO-FINANCIER AND OTHER PARTNER/STAKEHOLDER COMMENTS Borrower Comments from Poverty Alleviation Fund 1. Para 64. Weak fiduciary capacity and governance of PAF. After CIAA case there was a high staff turnover, but adequate number of staff were recruited as per the PAF By-laws. CIAA is a constitutional body which supports to maximize governance accountability and apex body of corruption control. It has supported to increase transparency in decision making process. Ultimately, CIAA has supported to strengthen overall governance system. 2. Fiduciary compliance section. The submission of financial reports was delayed because of lack of human resources in the finance unit and frequent turnover. 3. Para 93. The mentioned points happened at the starting phase and before the CIAA case but it was significantly improved during the last fiscal year: (i) grant register was updated, (ii) source wise expenditure is booked as per the grant agreements, (iii) most of the unreconciled items in BRS is now reconciled, (v) appropriate documents were obtained and scrutinized during payment procedure, (vii) submission of internal audit report was delayed because of delay in hiring an internal auditor, (viii) new building was constructed as per the mutual agreement between PAF and Nepal Trust, (ix) now, timely reimbursed to GoN, (x) Bank reconciliation is prepared in a timely manner now and uncleared cheques were cancelled on time, (xii) proper record keeping of assets is being done, (xv) physical verification of assets is done in a timely manner and reports submitted to the concerned oversight agencies, and (xvi) due to the lack of sufficient manpower and high staff turnover deadline was not met with past agreed actions. 4. Risk to development outcome section. The context of the country has changed significantly. It has been nearly four years since the Federal Democratic Constitution was promulgated in Nepal. As this country has shifted to federal governance from the unitary practice, local governments have now been granted an exclusive list of 22 different powers by the constitution itself. After local elections, the elected local representatives are assigned to offices where they start service and development delivery from local level. Local governments can identify their necessities, collect revenue, plan projects, layout budgets and implement these actions at the local levels. Although poverty reduction is a shared responsibility of three tiers of government, local governments have a special and vital role towards poverty reduction activities. 5. PAF has created around 32,000 COs and nearly 6,000 community-based small infrastructure projects. The 70th board meeting of PAF, which was chaired by the Rt. Honorable Prime Minister, has decided to hand over these organizations to concerned local governments in a systematic way. The meeting also passed management and operation guidelines for local level regarding COs and small infrastructures. Detailed descriptions of each CO and community small infrastructures have been prepared and circulated to the concerned local level. This endeavor will sensitize and support to internalize PAF programs at the local level. A Page 55 of 64 The World Bank Poverty Alleviation Fund II (P105860) 6. According to the Plan of Action endorsed by the cabinet, all PAF programs will be transferred to the Ministry of Land Management, Cooperatives and Poverty Alleviation (MoLMCPA) after July 16, 2019. PAF activities will be monitored, supervised and coordinated by a division of MoLMCPA after shifting the responsibility for the sustainability of COs and small infrastructures to the division. At the local level, discussions are ongoing to mobilize COs for production maximization and federated into cooperatives. Some local level governments have tried to replicate the best practices of successful COs. The GoN has allocated 50 million rupees for the remaining activities of PAF, which will be utilized for related training programs and other necessary actions. 7. Lessons and recommendations. Lesson 3: The institutional design of the PAF, as an independent agency proved to be appropriate during PAF I, however during PAF II, several shortcomings were exposed. Organization and development is a dynamic process. Timely intervention in Organization and Development is needed for a vibrant organization. There is a good practice in PAF: the Chair delegates the authority to the Vice Chair. The delegation of authority has supported timely decision making and smooth implementation of the programs. A few inconsistencies and lapses are seen in PAF Act and Rules, hence, it needs to be amended accordingly. Furthermore, if there is a sound relationship and coordination between the Executive Director and Vice Chair, there would not be any conflict or problem. From my perspective, there are no such difficulties in PAF II. Mr. Chhabi Rijal Joint Secretary Executive Director Poverty Alleviation Fund Comments from Office of Prime Minister and Council of Ministers (OPMCM) 1. Poverty Alleviation Fund (PAF) is a semi-autonomous agency established to address rural poverty and social exclusion related problems. At the time of requesting the World Bank for its financial and technical assistance, the situation was very different than that of today. A decade-long (1996-2006) conflict had just been ended. Then local bodies were administrated by the bureaucrats in absence of the people’s elected representatives. So, community driven development (CDD) instruments with the professional assistance from the non-governmental organizations were the best options to adopt. The modality worked very well. Most of the targeted outputs have been well achieved. It was successful in improving the living standards of poor and marginalized populations. 2. Towards the end of the project, following Nepal’s shift to the federal structure and establishment of local levels (municipalities) in the leadership of the people’s representatives, the mode of operation has been expected to be changed with the greater engagement of the local levels in the implementation of the project activities. So, after the successfully completion of the PAF II, all the community organizations (COs) have been handed over to the local levels along with the revolving funds that those COs hold. Further, arrangement has been made to ensure the sustainable use of all those local infrastructures, developed during the project period, in the leadership of the concerned Page 56 of 64 The World Bank Poverty Alleviation Fund II (P105860) COs and supervision of the local levels. Here in after, local levels will mobilize community to improve living conditions, livelihoods and empowerment of the poor people. 3. The Government of Nepal is always thankful towards the World Bank and its team for successfully implementation of PAF I and PAF II. Further comments on ICR assessment are the following; 4. Paragraph 62, the last sentence – “Moreover, throughout the second half of PAF II the Board started to lose its strength and convening power and was not as strong as in its earlier years.” – I don’t think so. 5. Paragraph 66-68 – I couldn’t understand what you want to highlight here. Do you want to say that CIAA affected the smooth running of PAF? If so, please be clear that CIAA is a constitutional body with the authority to investigate the abuse of authority. It’s not a good idea to indicate this institution as an obstacle on running any project. Just imagine what will happen in absence of such organization. 6. Paragraph 67, 2nd sentence – “However, the decision-making process at PAF became more challenging, as new staff were hesitant to make decisions and there was high turnover of and difficulty in recruiting competent finance staff.” It seems like all these happened because of CIAA; but in reality, that’s not true. As long as one follows the established rules and regulations, there is no need to afraid of CIAA. So, this sentence is not appropriate to write here. 7. Paragraph 111 – “At the time of first restructuring of second AF, an attempt was made to negotiate with PAF and GoN to cancel new CO formation and divert funds towards consolidation. However, the PAF Board did not accept the suggestion as the funds were already committed to new districts.” – Was there any agenda presented in the PAF Board regarding this and the decision of rejecting that? If not, better to delete these sentences as PAF Board is autonomous on taking its decision. 8. Paragraph 91, 93 and 94 – better to consult with PAF officials either these are true or just exaggeration. Mr. Mahesh Bhattarai Under Secretary OPMCM The Government of Nepal Comments from the International Fund for Agricultural Development (co-financier) 1. Outcome a. Relevance: Para 19, last sentence gives an impression that the PDO and modus operandi lost their significance at the end of project once the federalism was put into effect. Actually, PAF was criticized for inadequately adhering the provision of Local Self-Governance Act 1995 and Page 57 of 64 The World Bank Poverty Alleviation Fund II (P105860) its By-laws 1997 in failing to coordinate with the local level before the federalism process started in the country. Please review. b. Effectiveness: The PAF has achieved commendable outreach taking into account membership (22 percent Dalit, 32 percent Janjati, 3 percent Muslims and 43 percent others). The number of members fell slightly for the first two categories, while the holding of key positions was even more commendable, with 30 percent, 30 percent, 3 percent and 37 percent, respectively. This also applies to gender and social inclusion as 80 percent of the members were women. Please consider. c. Institutional strengthening: PAF missed the boat of gradually graduating the empowered COs and linking with other initiatives in the absence of a well-thought exit strategy from the beginning. This should be taken as a good lesson learned for future interventions both by the World Bank and IFAD. d. Private sector financing was far less than expected. It should be realized that linking of the empowered COs with the financial institutions would have leveraged more private sector financing. Mobilization of COs in and around the revolving fund hindered this process. 2. Key factors during Implementation: The following can be drawn as lessons for future as the key factors: a. Started with six poorest districts in PAF I, the expansion was rapid while reaching out to 59 districts and a very large number of VDCs. b. ‘Autonomy’ was granted for implementation of activities on the ground considering the conflict situation at the beginning. Efforts towards coordination were very weak at the district and VDC levels even after the signing of Comprehensive Peace Accord in 2006. c. More could have been done in the institutional transformation of the empowered COs into viable rural institution capable of linking with other development initiatives and also with the Fis for assured access to resources. 3. Lessons and recommendations a. Recommendation 2 can be reformulated in the spirit of: (i) segregating the savings from the revolving fund with defined percentage, and (ii) crafting an institutional transformation strategy as a part of the exit strategy to make COs a viable rural institution linked with financial institutions. Page 58 of 64 The World Bank Poverty Alleviation Fund II (P105860) ANNEX 6. SUPPORTING DOCUMENTS Bronfman J, Takada M, Monreal J. 2018. Nepal Poverty Alleviation Fund Quantitative Impact Evaluation. Washington DC: World Bank. Government of Nepal. 2003. The Tenth Plan: Poverty Reduction Strategy Paper 2002–2007. Kathmandu: National Planning Commission. International Monetary Fund. 2007. Nepal: Poverty Reduction Strategy Paper Progress Report. Washington DC: International Monetary Fund Micro credit ratings international limited (MCRIL) 2016. Review of Performance of Revolving Fund under PAF Poverty Alleviation Fund. 2016. Community Organization Maturity Assessment. Kathmandu: Government of Nepal. Poverty Alleviation Fund. 2018. Comprehensive Impact Evaluation of Poverty Alleviation Fund. Kathmandu: Government of Nepal. Poverty Alleviation Fund. 2019. Project Completion Report (PCR). Kathmandu: Government of Nepal. Poverty Alleviation Fund. 2007. Poverty Alleviation Fund Act 2063. Kathmandu: Government of Nepal. World Bank. 2003. Nepal — Country Assistance Strategy. Washington DC: World Bank. ———. 2004a. Nepal — Poverty Alleviation Fund. Project Appraisal Document Report 28606-NEP. Washington DC: World Bank. ———. 2004b. Nepal — Poverty Alleviation Fund. Project Development Grant Agreement H091-NEP. Washington DC: World Bank. ———. 2007a. Nepal — Interim Strategy Note. Report 38119-NEP. Washington DC: World Bank. ———. 2007b. “Nepal — Poverty Alleviation Fund Project II. Project Appraisal Document Report 41317–NP. Washington DC: World Bank. ———. 2009a. Nepal — Poverty Alleviation Fund Project. Implementation Completion and Results Report ICR1197. Washington DC: World Bank. ———. 2009b. Nepal — Interim Strategy Note. Report 48297-NEP. Washington DC: World Bank. ———. 2011a. Nepal — Poverty Alleviation Fund Project I. Independent Evaluation Group, Implementation and Completion Results Report Review ICRR13295. Washington DC: World Bank. ———. 2011b. Nepal — Interim Strategy Note. Report 63381-NEP. Washington DC: World Bank. ———. 2016. World Development Indicators: 2019. Washington DC: World Bank. --------. 2016. PAF institutional review. Washington DC: World Bank (unpublished). ______. 2018. Nepal — Country Partnership Framework. Report 121029-NP. Washington DC: World Bank. ______. 2018. Nepal Poverty Alleviation Fund Quantitative Impact evaluation” Washington DC: World Bank. Page 59 of 64 The World Bank Poverty Alleviation Fund II (P105860) World Bank and DFID (Department for International Development). 2004. Unequal Citizens: Gender, Caste and Ethnic Exclusion in Nepal. Washington DC: World Bank and Department for International Development, Kathmandu, Nepal. ANNEX 7. COMPREHENSIVE IMPACT EVALUATION Comprehensive impact evaluation (CIE), conducted during December 2017–May 2018 by an independent consultant group under the overall supervision of the National Planning Commission, consisted of both quantitative and qualitative analyses. Quantitative analysis was based on data obtained from baseline and follow-up household surveys conducted in different years. A quasi-experimental impact evaluation based on difference-in-difference estimators was adopted to quantify impacts on selected outcome indicators, and difference-in-difference estimations were calculated. The CIE qualitative analysis was based on results of 36 focus group discussions, key informant interviews and stakeholder group meetings. The study went beyond the results framework indicators, which mostly focused on output indicators, to better gauge PAF impact on the ground by triangulating different sources of evidence to assess achievements and impacts of PAF. In order to assess the impacts of PAF II, the project conducted a series of household surveys in various districts, according to the roll out schedule. Table A7a shows the coverage and years when the surveys were carried out. The survey procedure involved conducting a census of all households in the selected villages (or settlements), defined as Primary Sample Unit (PSU). Then a multi-module detailed household questionnaire was administered to 15 randomly selected households in each PSU. The survey questionnaire was adapted from the Nepal Living Standards Survey and included detailed information on consumption and income, socio-economic and demographic issues, including education, health and nutrition, physical assets, migration and remittances, employment, social environment, community relationship, voice and participation. All these contained critical information to study a household’s well being, and therefore, identified their evolution as the project got implemented. Both the baseline and subsequent follow-up surveys included the same key questions and also gathered basic information on the actual treatment status (PAF intervention) and non-treatment status (comparison) at both the household and PSU levels. Table A7a: Household surveys carried out by PAF 1st follow up 2nd follow up Household Phase Group Baseline survey survey sample Group II:5 6 districts Rautahat, Rolpa, PAF II 2007 2010 2014 2,089 Dailekh, Jumla, Humla and Doti Group III: 5 districts Taplejung, PAF II 2011 2017 1,372 Khotang, Dhanusha, Dhading, Bardiya Group IV: 5 new districts; Sunsari, PAF II 2011 2018 1,394 Dolakha, Gulmi, Surkhet, and Kailali Total 4,855 data collection for Group II, households belonging to PAF I and PAF II were surveyed. The VDC variable was used to isolate 5 During the sample corresponding to PAF II, as each VDC would only have either one of the projects. Page 60 of 64 The World Bank Poverty Alleviation Fund II (P105860) Each sample contained different households with different relationships to PAF II. On this basis, two different treatment and control groups were defined to analyze the impact of the program on different levels of participation. These treatments were defined as follows:  Treatment 1_RF (revolving fund): as “fully benefitted from PAF interventions”, i.e. being a CO member and accessing money from the revolving fund for income generating activities. The analysis focused on the reception of the revolving fund, as this is the core element of the PAF approach, in addition to social mobilization, and capacity and skill development. In other words, under this treatment definition, the treatment groups were those households who were CO members and received revolving funds for income generating activities; while the control groups were CO members who never received revolving funds.  Treatment 2_CO: as “being CO member (regardless of the reception of money from the revolving fund).” Community organization membership status was also considered important to estimate the impacts especially on the dimension of social cohesion. In other words, under this definition, treated groups were those who participated in COs for the entire period, and control groups were those who did not participate in COs for the entire period. The treatment and control groups, and the number of observations for each survey are shown in Table A7b. Table A7b. Treatment and control groups Survey group Group II Group III Group IV Treatment 1,600 684 145 Treatment 1 Controls 489 688 1,249 Treatment 1,262 263 66 Treatment 2 Controls 339 421 79 In this context, the CIE 2018 qualitative impact evaluation aimed at providing quantitative evidence of PAF’s achievements and impacts by using the available household survey data. Going beyond the results framework, the impact evaluations focused on several key indicators in line with the GoN’s priority emphasis on achieving Sustainable Development Goals and fighting poverty. Table A7c. Key dimensions and indicators studied Total expenditure is the total annual amount spent Total expenditure per capita (real term) This has sub-categories, including food (NPR) (log) expenditure, productive investment, human development and debt expenditure Economic Household Food expenditure per capita (real term) Expenditure for purchasing food development expenditure (NPR) (log) Food production expenditure represents Food production expenditure per capita the expenditure that would have been (real term) (NPR) (log) made on food if it had not been produced by the household Productive investment expenditure per Productive investment includes capita (real term) (NPR) (log) investments in agricultural inputs (such Page 61 of 64 The World Bank Poverty Alleviation Fund II (P105860) as seeds, fertilizers, insecticides, labor, bullock, irrigation) livestock, trade and land; expenditure on fodder, straw, veterinary services; buying animals (cow, bullock, buffalo, goat, sheep, pig, chicken, duck); renting/ sharecropping/mortgaging in-land Human development includes Human development expenditure per expenditures in education, health capita (real term) (NPR) (log) services and medicines Debt expenditure per capita (real term) Debt expenditure is the annual amount (NPR) (log) of interest paid Annual interest rate for the loan from Debt Annual interest rate for the existing neighbors, money lenders and relatives repayment debt over the past 12 months Total land (hectare) Total land in hectares Total livestock value (NPR) (real) Includes all household livestock Assets Asset bicycle Binary variable for the possession Asset radio Binary variable for the possession Asset phone Binary variable for the possession Household farm Household self-agriculture Household Household share crops member Household wage agriculture employment Household non-farm Employment and jobs Number of household members working Household self non-farm and Jobs including on each job category defined migration Household wage non-farm (primary job Household duties only) Household migration Household migration to urban areas Household international migration Women keep income Binary variables that indicate whether Women women get to keep their income and empowerme Women asked when property sold whether they are consulted when a nt property is to be sold Social Development Access to services disputes Land disputes Dispute variables are also binary variables that indicate whether the Social capital household was involved in any of the Water disputes disputes specified Percent of school enrolment (5–15 Percent of kids ages 5–15 years enrolled Education Human years) in school development Percent of children with birth Percent of children with birth Health complications complications Page 62 of 64 The World Bank Poverty Alleviation Fund II (P105860) Number of months in a year when Food Months of food sufficiency households had sufficient food for Security household consumption To estimate the impact of PAF on treated households, the studies calculated the difference-in-difference estimator. This estimation methodology permits to isolate the impact of the project when comparing the difference between treated and control households before and after treatment. Difference-in- difference is calculated using multiple regression with different control variables to better isolate the effect of the intervention and eventually detect a causal effect. The objective of this methodology is capturing the effects of participating in the project, controlling for time tendencies, initial differences between treatment and control groups, and other household characteristics that may affect the outcomes of interest. Additional control variables allow to better estimate the causal effect, especially when treatment is not assigned randomly. The main estimation model used was the following: = + + + ( ∗ ) + + Where Y is the outcome of interest, is a binary variable that takes the value of 1 if the observation corresponds to the treated group and 0 if it corresponds to the control group; is a binary variable that takes the value of 1 if the observation corresponds to the last follow up of the survey and 0 otherwise; ∗ is the interaction between both previously defined variables (so it takes the value of 1 if the observation was treated and it is the last follow up of the survey); and contains baseline control variables, specifically ecological zone belt dummies for mountain and hill zone, sex and literacy of the head of the household, caste, food expenditure of the household, months of food sufficiency, total household residents, total value of household livestock, and whether the household owns a phone and TV. We decided to use these baseline variables to control the differences in the initial characteristics of the households; these variables are key profiles in the demographic, social and economic dimension. The parameter α is the constant term, β the treatment group specific effect (accounting for the average permanent differences between treatment and control households before the intervention), γ is the time trend (captures the effect of the passage of time in the absence of intervention), corresponds to the parameter associated to baseline control variables and δ is the coefficient of interest: the true effect of the treatment. In order to assess the robustness of the estimations and given the data limitations (i.e. attrition) several methodologies were also implemented, such as the estimation of the average treatment effect with propensity score matching, and difference-in-difference with propensity score matching. These methodologies allow for control of covariates that assures that the analysis considers the lack of balance in the panel data. In particular, propensity score matching allows to compare each result of each treatment household with the outcome of its most similar control household, which ensures that the covariate balance between both groups (treatment and control) needed to conduct an accurate evaluation is met. The main results of the analyses are summarized as follows:  In general, the impact evaluation showed positive and statistically significant effects on household expenditure, with heterogeneous effects depending on the different surveys. Page 63 of 64 The World Bank Poverty Alleviation Fund II (P105860)  The results were overall positive and the PAF intervention helped CO members to accumulate assets in various forms.  The results on employment and jobs were varied and depended on household location. PAF helped poor households diversify their livelihoods and income generating activities, depending on their natural and economic endowments, while reducing the dependency on remittance.  The results showed that being a CO member improved women’s economic and social empowerment as well as participation in household decision-making when compared to those who were not CO members.  The impact estimations on human development varied depending on the indicator studied and the survey used. On food security and health, the impact estimations were positive and statistically significant in one of the surveys but not in the other. No significant effects were found on education (school enrolment). Page 64 of 64