98536 The Poverty Focus of Country Programs LES S O N S FR O M WO R LD BA N K E X PE R I E N C E The Poverty Focus of Country Programs: Lessons from World Bank Experience © 2015 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202–473–1000 Internet: www.worldbank.org This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202–522–2625; e-mail: pubrights@worldbank.org. ii Contents ACRONYMS AND ABBREVIATIONS .................................................................................................. VII ACKNOWLEDGMENTS ........................................................................................................................ IX OVERVIEW............................................................................................................................................ XI MANAGEMENT RESPONSE ........................................................................................................... XXIII MANAGEMENT ACTION RECORD ............................................................................................... XXVIII CHAIRPERSON’S SUMMARY: COMMITTEE ON DEVELOPMENT EFFECTIVENESS .............. XXXIV 1. ASSESSING THE WORLD BANK’S POVERTY FOCUS ......................................................... 1 Evaluation Objective and Scope ............................................................................................................................... 3 Results Chain, Evaluative Questions, and Instruments ............................................................................................ 8 2. PREPARING THE GROUND: WORLD BANK SUPPORT FOR POVERTY DATA ................ 14 State of Survey Data ............................................................................................................................................... 14 World Bank Support for Data Capacity Building ..................................................................................................... 22 Challenges beyond Household Surveys ................................................................................................................. 28 3. LAYING THE FOUNDATION: WORLD BANK SUPPORT FOR POVERTY DIAGNOSTICS . 33 Technical Quality of Poverty Assessments ............................................................................................................. 33 Characteristics of the Poor and Drivers of Poverty Reduction ................................................................................ 35 Timeliness and Dissemination ................................................................................................................................ 39 Constraints to Poverty Diagnostics ......................................................................................................................... 41 Good Practices and Lessons .................................................................................................................................. 43 4. FRAMING THE STRUCTURE: FORMULATING COUNTRY STRATEGIES .......................... 47 Factors that Condition the Poverty Focus of Country Strategies ............................................................................ 47 Consistency between Formulated Country Strategies and Poverty Diagnostics .................................................... 50 Role of Coordination and Consultation ................................................................................................................... 53 5. BUILDING OUT THE HOUSE: IMPLEMENTING COUNTRY STRATEGIES ......................... 57 Consistency between Bank Portfolio and Formulated Country Strategies .............................................................. 58 Complementarities in Implementation ..................................................................................................................... 62 Scaling Up and Portfolio Implementation ................................................................................................................ 66 Future Fidelity between Portfolio and Strategy ....................................................................................................... 68 6. OPENING THE DOORS AND WINDOWS: POVERTY FOCUS OF FEEDBACK LOOPS...... 71 iii CONTENTS Project-Level M&E .................................................................................................................................................. 71 Country-Level Results Monitoring and Learning ..................................................................................................... 73 Strengths and Weaknesses of the Feedback Loops across Countries................................................................... 75 Stakeholder Consultation and Coordination............................................................................................................ 79 7. CONCLUSIONS AND RECOMMENDATIONS ........................................................................ 82 Summary of Evaluative Findings............................................................................................................................. 82 Recommendations .................................................................................................................................................. 86 REFERENCES ...................................................................................................................................... 88 APPENDIX A. SUMMARIES OF 10 COUNTRY CASE STUDIES ........................................................ 96 APPENDIX B. EXTERNAL STAKEHOLDER SURVEY ON THE WORLD BANK’S SUPPORT FOR POVERTY REDUCTION ..................................................................................................................... 178 APPENDIX C. STAFF SURVEY ON THE WORLD BANK’S SUPPORT FOR POVERTY REDUCTION ....................................................................................................................................... 200 APPENDIX D. KEY FINDINGS FROM FOCUS GROUP DISCUSSIONS WITH STAFF .................... 227 APPENDIX E. SELECTION CRITERIA AND METHODOLOGY USED TO ASSESS THE QUALITY OF POVERTY ASSESSMENTS FROM 20 COUNTRIES ......................................................................... 237 APPENDIX F. TECHNICAL NOTE ON THE METHODOLOGY USED IN THE CASCRR/CPSCRR REVIEW .............................................................................................................................................. 244 APPENDIX G. A TECHNICAL NOTE ON CALCULATING THE PROXY FOR POVERTY FOCUS IN BANK LENDING ................................................................................................................................. 247 APPENDIX H. POVERTY DATA AVAILABILITY IN MICRO DATA CATALOG AND COST ESTIMATION OF STATISTICAL SYSTEM IMPROVEMENT ............................................................. 253 Boxes Box 1.1. Measures of Poverty ................................................................................................................. 5 Box 1.2. Defining the Poverty Focus of Bank Interventions .................................................................... 5 Box 1.3. IFC’s Poverty Focus and Results ............................................................................................. 7 Box 2.1. What are Poverty Data? ..........................................................................................................15 Box 2.2. Assessing Data Quality............................................................................................................18 Box 2.3. The Missing Mandate ..............................................................................................................21 Box 2.4. World Bank Participation in Partnerships for Building Statistical Capacity...............................28 Box 3.1. Good Practices for Poverty Assessments................................................................................34 Box 3.2. The Face of Poverty and Disadvantaged Groups ....................................................................35 iv CONTENTS Box 3.3. Literature on the Drivers of Poverty Reduction ....................................................................... 37 Box 3.4. Weak Data Limit the Scope and Robustness of Poverty Diagnostics ..................................... 42 Box 4.1. Poverty Focus of the Objectives of the Bank’s Country Strategies ......................................... 51 Box 4.2. PRSP and the Poverty Focus of the Bank’s Country Strategies ............................................. 52 Box 4.3. Selectivity and Donor Coordination ......................................................................................... 54 Box 4.4. Perceptions of Stakeholders on Bank Consultation and Coordination .................................... 54 Box 5.1. The Bank’s Response to the Great Recession and Support for Social Safety Nets ................ 59 Box 5.2. Country Assistance Strategy Formulation and Implementation: Airport Project in Egypt ........ 61 Box 5.3. Implementation Capacity and Legislation Constraints: the Case of Guatemala ...................... 61 Box 5.4. Synergy of Analytical and Financial Support in Bangladesh, Peru, and Malawi ...................... 65 Box 5.5. Piloting and Scaling Up: Two Projects in Malawi..................................................................... 67 Box 6.1. CPS Results Monitoring Systems in the Philippines and Bangladesh..................................... 75 Box 6.2. Variation in the Strength of Feedback Loops within a Country ................................................ 78 Tables Table 3.1. Years between Survey and Poverty Assessment Completion .............................................. 40 Table 3.2. Fewer than Half the CAS/CPSs were Preceded by PERs within Three Years ..................... 40 Figures Figure 1.1. Extreme Poverty Fell Steadily During 1990–2011 ................................................................. 2 Figure 1.2. Evaluation Chain of the Poverty Focus of Country Programs ............................................... 9 Figure 2.1. After Rising through 2002, the Number of New Household Income and Expenditure Surveys Leveled Off............................................................................................................................................ 17 Figure 2.2. Challenges Facing Bank Staff Conducting PAs and PSIAs ................................................ 20 Figure 2.3. Main Constraints to Poverty Data Collection: Perspectives from World Bank Staff............. 23 Figure 2.4. Main Constraints to Poverty Data Collection: Perspectives from World Bank Clients ......... 24 Figure 2.5. Capable Statistical Systems Produce More Surveys .......................................................... 25 Figure 5.1. Share of Investment Lending with Themes Directly Focused on Poverty ........................... 63 Figure 5.2. Share of Development Policy Lending with Themes Directly Focused on Poverty ............. 64 Figure 5.3. Share of Development Policy Lending in Total Lending ...................................................... 65 Figure 6.1. IEG Project M&E Ratings .................................................................................................... 72 Figure 6.2. Country Monitoring of Poverty-Related Indicators ............................................................... 74 v Acronyms and Abbreviations AAA analytic and advisory activities ADP Accelerated Data Program ARDE Annual Review of Development Effectiveness CAE Country Assistance Evaluation CAPMAS Central Agency for Public Mobilization and Statistics CAS Country Assistance Strategy CASCR Country Assistance Strategy Completion Report CASCRR Country Assistance Strategy Completion Report Review CASPR CAS Progress Report CCT conditional cash transfer CPF Country Partnership Framework CPS Country Partnership Strategy CPSCRR Country Partnership Strategy Completion Report Review DDI Data Documentation Initiative DFID Department for International Development (U.K.) DHS Demographic and Health Survey DPL development policy loan DPO development policy operation DQAF Data Quality Assessment Framework ESW economic and sector work EU European Union FCS fragile and conflict state GDP gross domestic product HIES Household Income and Expenditure Survey IBRD International Bank for Reconstruction and Development ICP International Comparison Program ICR Implementation Completion Report ICRR Implementation Completion Report Review IDA International Development Association IEG Independent Evaluation Group IFC International Finance Corporation IHSN International Household Survey Network IL investment lending IMF International Monetary Fund INEI Instituto Nacional de Estadística e Informática IRLDP Irrigation, Rural Livelihoods, and Agricultural Development Project LECS Lao Expenditure and Consumption Survey LSMS Living Standards Measurement Study M&E monitoring and evaluation MDB multilateral development bank MDG Millennium Development Goal MIC middle-income country MICS Multiple Indicator Cluster Surveys MSME micro, small, and medium enterprises NSDS National Strategy for Development Statistics NSO national statistics office ODA official development assistance vii ACRONYMS AND ABBREVIATIONS OECS Organization of Eastern Caribbean States OED Operations Evaluation Department PA Poverty Assessment PDR Peoples Democratic Republic PER Public Expenditure Review PPP purchasing power parity PRSC Poverty Reduction Support Credit PRSO Poverty Reduction Support Operation PRSP Poverty Reduction Strategy Paper PSIA Poverty and Social Impact Analysis QAG Quality Assurance Group SCD Systematic Country Diagnostic SCI Statistical Capacity Indicator SSN social safety net TFSCB Trust Fund for Statistical Capacity Building UNICEF United Nations Children’s Fund USAID United States Agency for International Development All dollar amounts are U.S. dollars unless otherwise indicated. viii Acknowledgments This evaluation is a product of the Independent Evaluation Group (IEG). The evaluation team was led by Xubei Luo under the supervision of Mark Sundberg (Manager, IEGPS) and the general direction of Emmanuel Jimenez (former Director, IEGPS), Nick York (Acting Director, IEGPS), and Caroline Heider (Director General, IEG). Eric Bell and Brett J. MC Farlan Libresco co-led the evaluation at the Approach Paper stage. Marcelo Selowsky provided input on the evaluation design in the early stages and comments on the main finding in the final stages. Shahrokh Fardoust prepared synthesis notes and provided input to the overall messages. Brian Aimes reviewed the survey results and triangulated the finding with multiple sources. Ravi Kanbur contributed to the design of the evaluation instruments and served as advisor to the team. Aline Dukuze was responsible for the administrative aspects. William Hurlbut, Barbara Rice, and Cheryl Toksoz provided editorial support. Bruce Ross-Larson provided editorial comments and inputs to the overall messaging. The 10 country case studies were conducted by Roger Grawe with support from Siban Shahana (Bangladesh), Ehtisham Ahmad with support from Mona Fayed (Egypt), Ulrich Lachler with support from Vivian Guzmán (Guatemala), Alan Piazza with support from Thiphavong Boupha (Lao People’s Democratic Republic), Kathie Krumm with support from Chance Ngamanya Benson Mwabutwa (Malawi), Kathie Krumm with support from Chukwuemeka Ugochukwu (Nigeria), Javier Bronfman and Marcelo Selowsky (Peru), Xubei Luo and Samantha Mignotte (the Philippines), Mauricio Carrizosa with support from Georgiana Necolau (Romania), and Brian Aimes with support from Ibrahima Ndiaye (Senegal). Izlem Yenice provided synthesis on IFC’s poverty focus and results. Virgilio Galdo provided input on data at the early stage, and Javier Bronfman carried out a stocktaking of the data availability. Robert Yin provided support to the design of the templates of the country case studies and comments to the pilot cases. Shahrokh Fardoust reviewed all the 10 country case studies and provided synthesis of the case studies. Bahar Salimova, with support from Samantha Mignotte, led the design and coordination of the internal Bank staff survey and external multi-stakeholder survey. Richard Tobin provided input on survey design. The external multi- stakeholder survey was conducted by an independent survey firm, ICF International. Javier Bronfman and Samantha Mignotte provided support on the survey data analysis. Zhan Shi collected information of external stakeholders. Bahar Salimova led the design and moderated the focus group meetings with select Bank ix ACKNOWLEDGMENTS staff, with input from Brian Aimes. Mauricio Carrizosa with support from Samantha Mignotte conducted a systematic review of the Country Assistance / Partnership Strategy Completion Report Reviews and the Country Assistance / Partnership Evaluations. Alan Piazza provided an in-depth assessment of the quality of the Poverty Assessment in 20 countries. Eric Swanson carried out a review of poverty data. Rasmus Heltberg, with support from Anna Amato, provided input on global partnership for statistical capacity building. Javier Bronfman, Samantha Mignotte, Aghassi Mkrtchyan, Moritz Piatti, and Jesse Torrence conducted the World Bank portfolio analysis and reviewed the Development Policy Operations and Investment Lending, Public Expenditure Reviews, and Poverty and Social Impact Analysis. Peer reviewers were Shaida Badiee (Open Data Watch), Alan Gelb (Senior fellow, Center for Global Development), Nora Lustig (Samuel Z. Stone Professor of Latin American Economics, Tulane University), and Martin Ravallion (Edmond D. Villani Chair of Economics, Georgetown University). The team is grateful to IEG colleagues who have provided helpful comments and suggestions, including Kostya Atanesyan, Elena Bardasi, Geeta Batra, Erik Bloom, Kenneth Chomitz, Marie Gaarder, Giuseppe Iarossi, Malathi S. Jayawickrama, Nidhi Khattri, Anjali Kumar, Midori Makino, Pia Schneider, Andrew Stone, William Sutton, Clay Westcott, Monika Weber-Fahr, and Fang Xu; and to Bank staff who have shared their insight at different stages, including Kathleen Beegle, Shaohua Chen, Olivier Dupriez, Johan Mistiaen, Umar Serajuddin, and Nobuo Yoshida. IEG extends its sincerest thanks to all of the stakeholders and Bank staff who participated in the interviews in the country case studies and focus group discussions, who provided feedback to the survey pre-tests, and who responded to the surveys, and to the country teams who helped facilitated the IEG country visits and stakeholder surveys. x Overview Highlights The World Bank Group in 2013 made the elimination of extreme poverty by 2030 a central institutional focus and purpose. This evaluation examines how, and how well, the Bank Group has focused its support on poverty reduction over the past decade, and what lessons to draw from this moving forward. The lessons aim to strengthen the Bank’s country diagnostics, improve the design of country strategies, and build greater learning opportunities from program experience. Using country case studies, surveys, focus group meetings, systematic reviews of Bank products, and other instruments, the evaluation examines the consistency of poverty focus in each of four links in a causal chain: data, diagnostics, strategy formulation, and strategy implementation through lending and nonlending instruments. It reviews the adequacy of the information base and usefulness of the analytical underpinnings that support country strategy formulation and implementation. It also evaluates the consistency of the poverty focus throughout the evaluation chain and the strength and weakness of feedback loops. The evaluation finds that the Bank’s work on data and diagnostics was generally robust, but with significant gaps in coverage and timeliness. Areas that require attention include improving country coverage and data capacity, reflecting the findings of the diagnostics in country strategy formulation, enhancing the consistency of the poverty focus in strategy implementation, and strengthening monitoring and feedback loops. Three main findings emerge.  Creating knowledge. The World Bank provides an important public good in supporting and reporting global poverty data, and producing high-quality poverty diagnostics. The Bank can better perform this role by investing more in sustainable data collection and by adopting data reporting standards as a part of its mission. On diagnostics, it should strengthen analysis of institutional issues and sociopolitical constraints, and improve the actionability of policy recommendations.  Understanding context. The Bank operates in a complex environment, and the choice of portfolio is conditioned by the strategic focus of a government client as well as the Bank’s comparative advantage. The government commitment to poverty reduction is a key factor in the fidelity between implementation and the formulated country strategy. When a country is not fully committed to poverty reduction, the Bank often faces a tough choice between disengaging from significant lending or engagement in areas that may be only tangentially related to poverty reduction. High-quality and timely diagnostics, policy xi OVERVIEW dialogue, and technical assistance should help identify entry points and lay the groundwork for greater impact.  Leveraging resources. Given the small size of Bank resources relative to the economies it seeks to influence, the effectiveness of Bank interventions in helping clients reduce poverty will increasingly depend on how it uses instruments as pilots and as catalysts to leverage resources from development partners and other stakeholders. Strengthening the complementarity among diagnostic work, technical assistance, and lending instruments, and among policy and investment lending instruments, can help to scale up efforts and achieve more sustainable, long-term impact. Poverty reduction has been a strategic share. The twin goals can be sustainably objective of the World Bank Group since achieved only through leveraging other the 1970s, when President Robert S. public and private sources of McNamara first made it a priority. development finance, and by using President James D. Wolfensohn later these sources more effectively to emphasized the importance of the stimulate growth and build the assets of mandate. In 2013 President Jim Yong the poor. Kim extended the vision by setting two goals (commonly known as the “twin Rationale, Objective, Scope, and goals”): to reduce extreme poverty— Structure of the Evaluation defined as income of less than $1.25 per day—to 3 percent by 2030, and to The World Bank Group recently promote income growth of the bottom endorsed a new Country Partnership 40 percent of the population. This was Framework (CPF) to define country the first time the Bank set explicit engagement. Each CPF will draw upon numerical targets and called attention to analytical work—a Systematic Country issues of equity and distribution. Diagnostic (SCD)—intended as a rigorous and independent diagnostic About 1 billion people still live in exercise conducted by Bank country extreme poverty, however, despite teams in consultation with national much progress since the 1970s and the authorities and other stakeholders. The launch of the 2001 Millennium SCD aims to become a reference point Declaration. Progress has been for client consultations on priorities for extremely uneven across countries and the Bank’s country engagement. localities. Achieving the World Bank’s Conducted upstream of the CPF, the goal of eliminating extreme poverty will SCD should help identify important require mobilizing far more resources challenges and opportunities within the than its own small and diminishing xii OVERVIEW country’s context to inform strategic given the 2011 evaluation, Assessing discussion of priorities for World Bank IFC’s Poverty Focus and Results, but it Group support. refers to its findings. This evaluation draws lessons from how The overarching question for this the Bank has designed and carried out evaluation is: “How, and how well, has its country programs to support poverty the Bank focused its programs on reduction during the past decade. It also reducing poverty in partner countries?” aims to provide analysis to improve The evaluation covers International how SCDs can inform the CPFs to Development Association (IDA) and achieve the twin goals. It explores the International Bank for Reconstruction analytic underpinnings of Bank support and Development (IBRD) assistance to poverty reduction, both income and during fiscal years 2004–2012. The non-income dimensions, and whether evaluation is structured along the and how the Bank has focused its results chain—from data to diagnostics, programs on poverty (box 0.1 describes to strategy formulation and “poverty focus”). This evaluation implementation, and to learning from focuses on the process by which the experience, or the “feedback loops” Bank has engaged with countries to (figure 1). The main evaluative support poverty reduction. Although questions are: poverty outcomes are noted, no effort is made to attribute outcomes to the Bank due • Did the Bank have the to the technical difficulty of attribution. appropriate data to understand The report aims to provide lessons to the nature of poverty and help sharpen the effectiveness of provide an information base for country programs as the post-2015 robust analytical work? agenda is launched, which will likely • With the given data, did the require more ambitious measures and Bank’s analytic work adequately actions. address poverty issues and identify policy priorities for To keep the analysis tractable and poverty reduction? focused, this evaluation excludes several • Did Bank country strategies important aspects of poverty, such as adopt the findings of analytical intrahousehold dynamics or work on poverty to help set distributional consequences within the priorities for and guide policy household, and intertemporal tradeoffs dialogue and lending? related to climate change and • Did interventions—operations, environmental poverty. It does not technical assistance, and capacity cover the International Finance building—reflect the strategic Corporation (IFC) and the Multilateral priorities for poverty reduction? Investment Guarantee Agency (MIGA), xiii OVERVIEW Box 1. Defining Poverty Focus of Bank Interventions Poverty is the result of economic, political, and social processes that interact and frequently reinforce each other in ways that exacerbate the deprivation in which poor people live (World Bank 2010e). It is well recognized that poverty eradication depends on both stimulating growth and providing basic social services to the poor (World Bank 1990). Many of the Bank’s interventions can be considered relevant to poverty in some way. However, defining the poverty focus and the degree to which development support contributes to poverty outcomes is not straightforward. Some interventions, such as for safety nets, may contribute directly and immediately to reducing income poverty; others, such as support for education, may do so with a long lag; and still others, such as improvements to the investment climate, may contribute indirectly or only in the presence of other policies or dynamics in the economy. It is beyond the scope of this evaluation to examine the impact of Bank-supported interventions on poverty reduction through all channels. Instead, the evaluation groups the Bank’s interventions into two broad categories—directly or indirectly focused on poverty reduction. Direct poverty focus is broadly defined as the activities that are designed and implemented to target or provide a disproportionate first-round benefit to the poorer segment of the population. This rough measure is indicative of strategic emphasis but no normative conclusion should be drawn from the relative weights. The evaluation recognizes that the links between Bank interventions and poverty reduction are complex and country-specific, and the Bank faces trade-offs in selecting projects with direct linkages (such as social safety nets) or indirect linkages (such as financial reforms) with poverty reduction. It does not judge whether the interventions with direct poverty focus have stronger impact on poverty reduction than those with indirect poverty focus. The report does not make a normative statement of whether there should be a larger or smaller share of interventions directly or indirectly focusing on poverty reduction as the binding constraints vary across countries. Instead, it assesses the extent to which lending operations, technical assistance, capacity building, analytical work, and policy dialogue were focused on the direct type of poverty interventions as identified in the poverty diagnostics. Obviously the proper mix of direct and indirect interventions that maximizes their joint impact on poverty reduction is highly country specific. • Did the Bank collect and draw • 10 country case studies lessons from poverty-reduction • An internal survey with Bank interventions to strengthen staff feedback loops and improve the • An external survey with client effectiveness of its country government officials and strategies and programs? If so, nongovernment stakeholders in how did it do so? 20 countries • A series of focus group meetings The analysis relies on the following with Bank staff working closely instruments: on poverty xiv OVERVIEW Figure 1. Evaluation Chain for the Poverty Focus of Country Programs • An assessment of the quality of survey data in the past decades. The Poverty Assessments in 20 stock of household surveys greatly countries increased over the past three decades: in • A review of the Country the 2000s, 40–50 surveys were Partnership Strategy Completion conducted each year, up from 20–30 Report Reviews and Country surveys in the 1990s and less than 10 per Assistance Evaluations year through most of the 1980s. • Stocktaking exercises of PovCalnet data, Development The Bank is seen as a global leader and Policy Lending (DPL), valued development partner, providing Investment Lending (IL), Public technical expertise and building capacity Expenditure Reviews (PERs), and to support its country clients in their Poverty and Social Impact efforts to improve poverty data Analyses (PSIAs). availability, accessibility, and quality. Through programs like the Living Standards Measurement Study, the Bank is a major contributor to the collection of Support for Poverty Data credible data and improvement of The quality of data and its accessibility methodologies to estimate poverty. play a critical role in measuring poverty, However, progress is uneven across identifying the poor, and monitoring countries. Often data are most scarce progress in poverty reduction. Many where the challenges of poverty are most countries, with the Bank’s support, have severe, such as in many low-income made significant progress in poverty xv OVERVIEW countries and most fragile and conflict- political constraints to data access affected economies. undermine their value, as in the Arab Republic of Egypt, where data access In general, the Bank works well with has been limited. In countries where other international partners in poverty is a politically sensitive topic, supporting the government to conduct restrictions on data remain a major household surveys, but there is obstacle to analyzing the magnitude, considerable variation across countries. nature, and distribution of poverty. The portrait of national poverty is a critical input into broad policymaking and strategy formulation, particularly Support for Poverty Diagnostics when supplemented with strong The development community sees the knowledge of the country context. Bank as a leader in providing poverty Nevertheless, the lack of good-quality, diagnostics. Generally, Bank poverty timely poverty data and the issues of diagnostics are of high technical quality. data accessibility remain major The best work is done in countries with constraints to carrying out robust good-quality data that is available in a diagnostics and policy dialogue. timely manner. Poverty Assessments “Data are the lifeblood of decision broadly make good use of available making,” notes A World that Counts: quantitative data in order to derive Mobilising the Data Revolution for poverty incidence indicators, identify Sustainable Development, a 2014 report to the key drivers of poverty, develop a the United Nations. Good data are poverty profile and, in some cases, essential to identify the poor and their develop a poverty map. They often characteristics, measure changes in examine income and non-income poverty over time, and assess the poverty at the national and regional effectiveness of interventions to reduce levels and across social groups, tailoring poverty. Going forward, the need for to country conditions. Two Poverty data to measure achievement of the Assessments—Ethiopia: Well-Being and Bank’s twin goals will increase. But the Poverty in Ethiopia (2005) and Indonesia: sustainability of data efforts is Making the New Indonesia Work for the challenging in many countries where Poor (2006)—stand out as excellent other claims on resources take priority. examples of good practice. Unsustained support can jeopardize However, the Bank’s analytical work on data progress. In Guatemala, the quality poverty often does not adequately of household survey data worsened address the important social and after the Bank and donor-supported political factors that contribute to MECOVI (Programa para el Mejoramiento poverty and impede efforts to reduce it. de las Encuestas y la Medición de Robust and independent poverty Condiciones de Vida) ended. Elsewhere, xvi OVERVIEW diagnostics that identify social and consistent with poverty diagnostics. For political parameters as well as example, high-quality poverty distributional issues, institutional diagnostics in Malawi informed country capacity, and excluded communities, strategy formulation, leading to a new are better prepared to provide relevant focus area for poverty reduction work: and actionable policy recommendations. nutrition programs to fight against child stunting. Insufficient alignment between the timing of analytic work with policy The Bank’s strategy is more poverty- cycles (the annual budget, five-year focused when the client government is plans, and the like) and weak public committed to poverty reduction. When dissemination can further undermine client governments lack such political the strategic impact of analytic work. commitment and do not have a clear, Strengthening programmatic planning poverty-focused strategy, the Bank has can help better align the timing between less success in formulating its own supply (the household survey cycle, strategy. This is particularly true in which conditions the cycle of the middle-income countries, which have a diagnostic work) and demand (Country limited need for Bank financing. In a Program Strategy cycle). Providing challenging environment with deeply succinct summaries tailored to rooted, vested interests and weak policymakers and more closely commitment to poverty reduction, the partnering with government clients and Bank can identify entry points for other stakeholders can help increase impact through high-quality and timely public awareness and transparency of diagnostics, policy dialogue, and the diagnostics, and hence the impact of technical assistance. One example is the the analysis on policymaking. Philippines. During periods of low government commitment to poverty reduction, the Bank focused its support The Bank’s Country Strategy on identifying appropriate areas for Formulation additional interventions, piloting Overall, the Bank’s country strategies conditional cash transfer programs, and address the poverty reduction objective. laying the ground work for reform. A majority of the Bank staff and When new space opened for reforms, government officials in the external the Bank seized the opportunity to survey reported that the Bank’s country formulate country programs with strategies addressed the main causes of stronger poverty focus and scaled up poverty. The evidence generally successful pilot projects. confirms this view; Country Assistance In low-income countries, the Poverty Strategies (CASs) and Country Reduction Strategy Paper (PRSP) Partnership Strategies (CPSs) are largely process, which draws on the poverty xvii OVERVIEW diagnostics, focuses strategies on poverty development partners (including country (both income and non-income). It clients, donors, and the private sector), increases cohesion in sectoral strategies and ensuring that projects are designed and the overall macroeconomic with appropriate piloting and monitoring framework, and improves coordination and evaluation (M&E) to assess whether among donors and recipient countries. In they warrant replication and scaling up. Bangladesh, Lao People’s Democratic Republic (Lao PDR), and Senegal, the The Bank’s Strategy Implementation Bank’s strategies and interventions were based on pillars identified in the The Bank’s country strategies and the governments’ PRSPs. interventions supported by its lending and nonlending portfolio (advisory and Making development outcomes more technical assistance) broadly reflect the effective in reaching the poor and more client countries’ poverty reduction sustainable requires greater focus on strategies and the development inclusive growth. When the government’s priorities of country clients. A majority own strategy is not clear or not focused of Bank staff and external on poverty reduction, the Bank’s strategy stakeholders—including local civil can fill the gap and be opportunistic. It society and government officials— can engage in areas where the country’s believe that the Bank’s lending and own development strategy aligns with nonlending instruments address the the goal of reducing poverty and also poverty focus of the Bank’s strategies. reflects the Bank’s comparative Analytic and advisory activities, advantage. Stronger attention to the particularly in middle-income countries, challenges that the extreme poor face— are generally considered to have strong particularly the non-income aspects such systemic impact on strategy formulation as child malnutrition and other and lending, contributing to poverty irreversible human capital damage—is reduction. When the Bank’s lending and crucial for sustainable outcomes in nonlending instruments complement poverty reduction. each other, support to country clients The Bank’s financial resources are tends to be more effective and well typically small relative to the economies calibrated to the local country needs. it seeks to influence, limiting its direct There is often deviation between the relevance for poverty reduction. Two key formulated strategy (in writing) and strategies for sustaining the Bank’s implementation of the strategy. In part, poverty reduction outcomes are this can occur for good reasons, for leveraging its own resources (including example adjusting to changes in the the World Bank, International Finance external or domestic environment (such Corporation, and Multilateral Investment as the global financial crisis, commodity Guarantee Agency) with those of other xviii OVERVIEW price shocks, natural disasters, or When the government follows a reform changes in political direction). But agenda and the economy grows rapidly, deviations can also be due to a partner the Bank can focus on interventions that country’s weak commitment to poverty are more directly poverty-focused, such reduction, limited implementation as improving basic service delivery with capacity, or legislative constraints. targeted and differentiated actions, to When a country is not fully committed speed up poverty reduction among to poverty reduction, the Bank often extremely poor and isolated groups. As faces a tough choice between suggested by country case studies, disengaging from significant lending, or when growth is weak or imbalanced, continued lending in areas that the Bank needs to spread its tangentially related to poverty reduction interventions to both support broad- priorities. With robust data in place, based and inclusive growth and to high-quality and timely diagnostics, directly address poverty priorities. policy dialogue, and technical assistance can help identify entry points and lay The Bank can improve the deployment the groundwork for impact. of its instruments so that they complement each other to strengthen The 2008–2009 global crisis resulted in collective impacts on poverty reduction. considerable, sudden shifts in the Bank’s The safety net programs in Bangladesh portfolio across the affected countries as provide a good example of using the the portfolios were being implemented. synergy between instruments and The Bank’s total commitments (for IBRD scaling up to expand resource and IDA lending) more than doubled, deployment. However, project lending from $25 billion in 2008 to $59 billion in is often viewed on its own terms instead 2010, with a sharp increase in budget of as a catalyst to leverage far greater support (DPLs) in IBRD countries, non-Bank resources. The which was generally efficient in complementarity between policy providing for rapid increases in loan lending and investment lending is yet to sizes and disbursement amounts (IEG be fully exploited. 2009). However, there are indications in some cases, such as Guatemala, that the An important change in the mindset is a shift of lending away from poverty- much stronger and explicit emphasis on focused interventions and toward DPLs scaling up: projects that address poverty after the global crisis was mainly due to need to be viewed as opportunities to a preference for fast-disbursing budget crowd-in resources from the public and support over investment or project private sectors, as well as from other lending, because of the latter’s greater development partners, not simply in implementation difficulties. terms of circumscribed interventions. The impact of the Bank should be catalytic and beyond the individual xix OVERVIEW intervention. This mindset needs to be Bank resources. At the strategy level, built into dialogue, planning, design, CPS results monitoring covers a wide and the intelligent use of pilots and range of poverty-related areas, with sequential planning, so the Bank can education and health receiving the most inspire action and induce policy change. attention and infrastructure often One way to scale up successful, receiving the least. Most reports tend to individual poverty-focused investment focus on the process, and there is no clear projects (particularly pilots) is to use evidence that the CAS Completion policy dialogue and DPLs to influence Reports (CASCRs) and the M&E the allocation of critical resources. frameworks for CAS/CPSs monitor and evaluate poverty outcomes. This is a missed opportunity to leverage The Bank’s Feedback Loops development finance for more sustained The Bank generates information and impact on poverty reduction. learning about poverty reduction from Deficiencies in M&E design are its programs. Requirements and frequently identified as shortcomings in processes are in place for this purpose Bank support at project entry. According through M&E; however, they are not to the Results and Performance of the always well implemented. The Bank’s World Bank Group 2014 (IEG 2015), only feedback loops—from results to data about 30 percent of projects’ M&E analysis to diagnostics to strategy frameworks were rated satisfactory in formulation and implementation—have FY2008–2010 and this declined generally been weak, with large marginally in FY2011–2013. From the variation across countries. country case studies in general, the Project and program-level M&E need to linkages are stronger between data and inform the design and implementation diagnostics, but linkages are weaker with of country strategies and provide a basis respect to how data and diagnostics feed for scaling up to better leveraging of into strategy formulation and resources. implementation. For example, in Nigeria, feedback loops on poverty reduction, At the project level, there is some from data to diagnostics to strategy experience with project piloting and formulation and implementation, have scaling up. However, Implementation been incomplete. Needed improvements Completion Reports, the main feedback in data and diagnostics are not in place, instrument on individual interventions, mainly due to lack of local demand. are rarely used to reflect on improving the design and implementation of the The strength of feedback loops in a country strategy or to provide a basis country varies across sectors. In for scaling up by better leveraging non- Bangladesh, feedback loops are strong in social protection and education xx OVERVIEW programs but weak in infrastructure formality to a real tool, can help the interventions. The enabling factors for Bank and development partners learn strong feedback loops include a from its successes and failures. combination of government commitment, staff commitment, and Recommendations Bank management support to measure results, increase technical expertise The Bank has had notable success in through staff training, and provide focusing its support on poverty technical assistance. reduction in its country programs with a combination of lending and nonlending Most of the Bank’s country strategies instruments. The Bank’s work on data were developed through some kind of and diagnostics is generally strong, but participatory consultations with there is substantial variation across government and nongovernment countries and room to improve strategy stakeholders. For example, in formulation and implementation. If the Guatemala, all three country strategies Bank is to effectively support underwent extensive CAS consultation achievement of the twin goals, the areas processes. However, in the majority of that require attention include reflecting cases, there was no clear evidence that the findings of the diagnostics in the such consultations had a meaningful formulation of country strategy, effect on the design or implementation enhancing the consistency of the of Bank strategies. In some instances poverty focus between the formulated consultations appeared to be more of a strategy and its implementation, and “box-ticking” exercise. For example, in strengthening the monitoring and Peru, multiple consultations were feedback loops. conducted with the incoming and outgoing authorities, the private sector, The findings support the following and civil society for the preparation of recommendations in five areas to guide CPS 2007. However, the major direction improvements of the Bank’s future work of the CPS regarding poverty is isolated on poverty reduction, particularly in the from the topics discussed at the design and implementation of the SCDs consultations. and CPFs. In the IEG evaluation Learning and DATA Results in World Bank Operations: How the 1. Ensure that poverty data Bank Learns (2015), lack of institutional development and reporting needs are incentives was identified as one of the comprehensively addressed in the SCD biggest obstacles to knowledge sharing and country policy dialogue to identify and learning. Changing incentives and gaps, steps to fill them, and requisite culture to emphasize learning, and financing arrangements. moving M&E from the status of a xxi OVERVIEW 2. Advocate and organize support to 7. In the country strategy address the sustainably improve the capability of mix between indirect poverty national statistical agencies, both interventions (inclusive growth) and internal operational support and in direct poverty interventions (social partnership with external agencies. safety nets, access to basic services) with attention to their sequencing to achieve 3. The Bank Group should take a the Bank’s twin goals. stronger lead in strengthening mechanisms for quality and STRATEGY IMPLEMENTATION AND FEEDBACK LOOPS transparency on poverty data, motivate 8. Develop and adopt explicit country compliance, and regularly evaluation protocols for piloted disseminate data. interventions to capture lessons from experience on poverty reduction, with a DIAGNOSTICS view towards opportunities for scaling 4. Strengthen the Bank Group’s up successful interventions. poverty diagnostic work to ensure that it incorporates relevant social and 9. Ensure attention at project inception political dimensions of poverty analysis. to evaluability through (1) developing standards for baseline measurement, (2) 5. Focus poverty analysis on actionable explicit linking of the baseline to priorities for policy interventions to indicators relevant to project objectives, accelerate poverty reduction and including any that refer to poverty or develop the SCD discussion of linkages inclusion impacts, and (3) robust between recommended actions and planning for monitoring data required their expected impact on poverty for ex-post evaluation. reduction. Strategy Formulation 6. Pursue the recommended actions on poverty from the SCD through CPF country strategies. xxii Management Response The World Bank commends the Independent Evaluation Group (IEG) for this evaluation of lessons and experience with the poverty focus of country programs. This evaluation is timely. The new country engagement model, including the Systematic Country Diagnostic (SCD) and the Country Partnership Framework (CPF), aims to enhance the focus on the goals of eliminating extreme poverty and promoting shared prosperity in a sustainable way, and should address various findings and recommendations of the report. A strong SCD requires high-quality data and greater use of evidence to guide policy dialogue and interventions. Improving the poverty focus of World Bank Group programs thus remains an ambitious task, but one that is well worth pursuing. Management looks at the report as an opportunity to reflect on a number of questions: (1) Why are data lacking in some countries but not in others? (2) Why has diagnostic work influenced policy changes in some contexts but not in others? (3) How important is the quality and relevance of analysis to the degree of success achieved? (4) How important has client engagement been; and what is the role of Country Management Units in this process? A better understanding of these issues would help strengthen the Bank’s country diagnostics, improve the design of country strategies, ensure a more evidence-based approach to the Bank’s programs and interventions, and increase learning. Specific Comments Main findings of the IEG report. Management concurs with the main findings related to (i) creating knowledge, (ii) understanding context, and (iii) leveraging resources, and offers specific comments on these topics: • Data collection and reporting. The Bank can improve its performance by investing in sustainable data collection and by adopting data reporting standards as a part of its mission. • Poverty diagnostics. The Bank should strengthen analysis of institutional issues and sociopolitical constraints, and improve the actionability of policy recommendations. • Strategy formulation. High-quality and timely diagnostics, policy dialogue, and technical assistance should help identify entry points and lay the groundwork for greater impact. xxiii • Strategy implementation. Strengthening complementarity among diagnostic work, technical assistance, and lending instruments, can help to scale up efforts and achieve more sustainable, long-term impact. DATA COLLECTION AND REPORTING Criteria for selecting country case studies. The report would have benefited from clearer criteria for selecting countries for case studies. The small sample of countries is not representative. Management flagged this issue to the team at the Approach Paper stage. Interpretation of the findings from such a small sample requires caution. Uneven progress across countries. Management acknowledges that data are often scarce where challenges of poverty are most severe. For instance, this is the case in many low-income countries and economies affected by fragility, conflict, and violence. Data accessibility. Management appreciates the report’s recognition of the need for effective collaboration with partners to strengthen data quality and availability. The report also notes the need to improve access to good-quality and timely data on poverty to support an informed policy dialogue. The analysis of the current state of data quality and availability is useful. The report mentions the Demographic and Health Survey as significantly improving the international stock of high-quality, comparable demographic, and health data. Similar improvements are necessary to improve the quality of household income or consumption data. The report could have explored how support for such efforts differs among development partners. The IEG report is silent, for example, on the Bank’s approach to supporting the collection of household survey data and how to balance international comparability with national data needs. Measuring the effectiveness of Bank interventions. As noted in the report, good data are essential not only to identify the poor and their characteristics, but also to measure progress. The report suggests “improving the actionability of policy recommendations for poverty reduction and strengthening the linkages between recommended actions and the expected impact on poverty.” This calls for more disaggregated analytical work on the poor across spatial and other characteristics. In spite of the positive assessment of the quality of diagnostics, and recent progress on this front, the report also notes the variation in the quality of analysis and, at times, a lack of actionable guidance. The report also notes issues with outcome measurement and data limitations when trying to attribute poverty reduction outcomes to Bank- supported interventions. Management broadly agrees with this analysis. The SCD has been developed to address these issues, i.e., improving the quality of xxiv diagnostics, identifying data gaps as well as priority areas to achieve the Bank Group’s goals, and creating demand for better data quality and poverty-focused Bank interventions through the CPF process. The CPF will determine Bank Group support based on the priorities identified through the SCD and client demand, taking into account the Bank Group’s comparative advantage and complementarity with other partners. POVERTY DIAGNOSTICS Addressing social and political factors. Management appreciates IEG’s recognition that Bank poverty diagnostics are of high technical quality. Management agrees there is significant room for improvement in analytical work to address social and political constraints to poverty reduction. The Poverty and Social Impact Analysis is one mechanism for doing this systematically. It is particularly useful for assessing the impact of policy reforms. The SCD process also requires addressing social and political impediments to poverty reduction and greater inclusion. Alignment with policy cycle. The insufficient alignment between the timing of analytic work with the policy cycle has been identified as a factor limiting strategic impact on policy making. Management concurs with the recommendation to improve programmatic planning to better align timing and dissemination of analytic work to increase public awareness, transparency, and impact on policymaking. STRATEGY FORMULATION Political commitment to poverty. Management appreciates the assessment that the Bank’s country strategies have addressed the overall objective of poverty reduction. The report also notes that poverty-focused strategies have been more successful in countries that have the requisite political commitment. The report highlights the difficulty of pursuing a “poverty-focused” approach when there is a weak government commitment to poverty reduction or a disagreement on priorities between the Bank and government. Such differences limit the Bank’s options. It can either disengage from major lending or limit engagement to politically feasible entry points with the hope that a future government will be more committed to this agenda. The report calls for more strategic use of lending and nonlending work or more pilot work that can be scaled up under more favorable circumstances. However, the dilemma the Bank often faces is not one necessarily based on engagement with countries on poverty issues, but rather on the Bank’s added value to the agenda. If a government chooses to emphasize approaches that aim at broader “development” with mostly indirect (but perhaps large) impacts on poverty, the report seems to suggest that this is a “lack of commitment” to the poverty agenda, xxv equivalent to a government where an elite directs investments to serve its own interests. Management feels that these are fundamentally different situations. Inclusive growth. Management recognizes the need to strengthen its focus on inclusive growth by complementing the government’s own development strategy and by paying greater attention to the challenges related to extreme poverty. This need has been articulated in the Bank’s commitment to the goals through a more selective and creative engagement with clients. The report puts considerable emphasis on inequality and exclusion as challenges to development strategies. Inclusiveness of the development process and the underlying institutional or economic factors need more attention. New engagement model. The report misses an important opportunity to discuss how its recommendations relate to the Bank’s new, country engagement model. Overall, the reference to the new context is minimal, despite its stated intention to “provide analysis to improve the ways that SCDs can inform the CPFs to achieve the twin goals.” Prioritizing country program interventions based on diagnostics is critical in ensuring the poverty focus of country programs. The IEG recommendations could have been better tuned to the new context, taking into consideration a more in-depth analysis of past experience with poverty diagnostics underpinning country programs. STRATEGY IMPLEMENTATION Prioritization of poverty reduction by clients. Management welcomes the report’s positive remarks regarding implementation of the Bank’s country strategies, which reflect the poverty reduction and development priorities of client countries. Management also agrees that there is room for better aligning strategy and implementation. Management agrees that stronger analytic and advisory activities may play an important role in shifting attention to poverty objectives and identifying critical areas of engagement. Balanced approach. The report says that in cases when growth is weak or uneven, the Bank needs to balance its approach to broad-based and inclusive growth with its approach to addressing poverty. The report recommends the Bank enhance the coherence of country strategies by identifying an appropriate mix of interventions to reduce poverty, including the sequencing of interventions to promote inclusive growth and reach excluded communities. This recommendation is too general and too broad. It is not clear how such recommendations can be implemented. The report is not clear on its recommendations related to analytical tools for prioritizing interventions. xxvi Catalytic interventions. Management welcomes the report’s emphasis on leveraging the World Bank’s interventions, including using development policy lending to influence the allocation of critical resources or crowding-in resources from the public and private sectors. Strengthening capacity. Management very much welcomes the report’s recommendation to improve the capabilities of national statistics’ agencies. This is critical for sustainable improvements in data quality. Project-level monitoring and evaluation (M&E). Management agrees with IEG on the importance of strengthening M&E at the project level. The report seems to suggest, however, stronger links between project-level outcomes and poverty reduction. While much needs to be done to sharpen the alignment of project monitoring indicators to reflect intended outcomes at the right level of the results chain, “poverty reduction” is usually a much higher level outcome than one project can or should aim to attain. Recommendations The recommendations made in the report specify several constraints. Management concurs with the challenge of improving data quality and availability and the need to strengthen client capacity in this area. It also agrees with addressing the missing links in the results chain—from data to solutions—in the Bank’s work with clients. The report’s emphasis on the importance of political will and commitment within client countries is welcome as are the insights on specific areas for improvement. However, some of the recommendations provide guidance that is too generic and, therefore, difficult to measure. The following Management Action Record matrix contains the Management response to specific IEG findings and recommendations. xxvii Management Action Record Acceptance by IEG Findings and Conclusions IEG Recommendations Management Management Response The quality of data and its Ensure that poverty data World Bank: At the country level, SCDs will be accessibility play a critical role development and reporting Agrees documenting key data gaps that in measuring poverty, needs are comprehensively constrain the analysis. The Poverty identifying the poor, and addressed in the Systematic Global Practice is engaged in an effort monitoring progress in Country Diagnostic (SCD) to make this a more systematic poverty reduction. In the past and country policy dialogue process by providing better guidance. decade, some 30 low- or to identify gaps, steps to fill Once critical gaps are identified, the middle-income countries have them, and requisite Country Partnership Framework not had household surveys financing arrangements. (CPF) is an instrument that identifies (which does not allow to Bank interventions, based on client establish a poverty profile) and demand and in coordination with some 20 more have had only support from other development one survey (which does not partners. allow to track poverty trends). The lack of good-quality, timely poverty data and the issues of data accessibility remain major constraints to carrying out robust diagnostics and policy dialogue. xxviii MANAGEMENT ACTION RECORD With the Bank’s focus on the Advocate and organize World Bank: Management intends to encourage twin goals, the Sustainable support to sustainably Agrees stronger country engagement, Development Goals gaining improve the capability of through the CPF process and international traction, and the national statistical agencies, potentially in collaboration with other pressing demand for results, both internal operational development partners, for building data needs are increasing. support and in partnership statistical capacity and addressing Sustainability of data efforts is with external agencies. data needs identified in the SCD. challenging in many countries where capacity is weak and other claims on resources take priority. Unsustained support will jeopardize data progress. The Bank plays a leadership The World Bank Group World Bank: The Bank Group is engaged in role in poverty data standards, should take a stronger lead Agrees strengthening data quality and support for data collection, in strengthening transparency at various levels. Within and data dissemination. Unlike mechanisms for quality and countries, the Bank Group staff the International Monetary transparency on poverty provide technical assistance and Fund, the absence of the data, motivate country capacity building and promote open mandate to require the compliance, and regularly data policies. There are ongoing collection and dissemination of disseminate data. efforts to curate, archive, and vet data standardized data on quality, its dissemination, and household income and poverty enhanced availability. constrains the Bank’s ability to achieve the twin goals. The Bank is exploring mechanisms to create standards on data access, xxix MANAGEMENT ACTION RECORD quality, and timeliness that countries can voluntarily adhere to. The Bank’s analytical work on Strengthen the Bank World Bank: A sizeable number of poverty poverty often does not Group’s poverty diagnostic Agrees advisory and analytical activities, in adequately address the work to ensure that it the past but especially in more recent important social and political incorporates relevant social times, has examined the economic, factors that contribute to and political dimensions of social, and political drivers of poverty poverty and impede efforts to poverty analysis. reduction and shared prosperity. They reduce it. Robust and have also articulated policy actions independent poverty that would lead to achieving the diagnostics that identify social goals. Poverty analytic work is an and political parameters as important input into the SCD process. well as distributional issues, Strengthening the social and political institutional capacity, and aspects of poverty diagnostics is also excluded communities provide an intent of the Poverty and Social more relevant and actionable Impact Analysis work which focuses policy recommendations. on the implications of policy reforms. This is now also being addressed by the Poverty Global Practice through its Global Solutions Areas— particularly the Fiscal and Social Policies and Markets and Institutions for Poverty Reduction and Shared Prosperity. xxx MANAGEMENT ACTION RECORD Poverty diagnostics often Focus poverty analysis on World Bank: The SCD has been developed exactly provide strong technical actionable priorities for Agrees to help address these issues. analysis but lack actionable policy interventions to Therefore, this recommendation will guidance, thus limiting their accelerate poverty be addressed through implementation direct relevance for strategy reduction. and develop the of the new country engagement and policy design. Poverty SCD discussion of linkages framework as part of the prioritization diagnostics are better at between recommended undertaken in the SCD. The Poverty informing the formulation of actions and their expected Global Practice also addresses this poverty-focused country impact on poverty issue at the sectoral level by strategies when they provide reduction. supporting other Global Practices in actionable policy the operationalization of the goals recommendations. within sectoral programs. Country strategies need to be Pursue the recommended World Bank: Under the new country engagement underpinned by solid analysis. actions on poverty from the Agrees model, the SCD provides a critical Building a portrait of national SCD through CPF country input into the CPF formulation poverty drawn from data is a strategies. process, providing a necessary space critical input into broad policy for CMU to take into account the making and strategy country’s own development strategy, formulation. the client’s demand, and the Bank’s comparative advantage. Both growth and basic service In the country strategy, World Bank: The distinction between what delivery with targeted and address the mix between Partially agrees constitutes a direct as opposed to differentiated interventions are indirect poverty indirect poverty intervention is identified as key drivers of interventions (inclusive complex and difficult to poverty reduction. The growth) and direct operationalize in country strategies. appropriate balance between interventions (social safety However, the CPF should clearly xxxi MANAGEMENT ACTION RECORD ‘direct’ and ‘indirect’ poverty nets, access to basic articulate why the mix of interventions for Bank support services) with attention to recommended activities is will depend on country their sequencing to achieve appropriate, given the findings of the context and capacity. the Bank’s twin goals. SCD. Pilots are used to strengthen Develop and adopt explicit World Bank: Each piloted intervention, particularly the design, implementation, evaluation protocols for Agrees those that aim to reduce poverty or and scaling-up of projects and piloted interventions to enhance shared prosperity, should to enhance the poverty focus of capture lessons from develop a well thought-through the Bank’s projects. Early experience on poverty process and protocols for evaluating intermediate outcomes can reduction with a view its impact and create the evidence attract additional resources toward opportunities for base for scaling up. leading to scaling up. The scaling up successful impact of Bank support for interventions. poverty reduction will depend in part on successful use of pilots to leverage and crowd-in external resources to scale up. However, evidence on whether pilots are successful and serve to leverage non- Bank resources is rare. Deficiencies in the monitoring Ensure attention at project World Bank: Management agrees with the overall and evaluation (M&E) design inception to evaluability Partially agrees thrust of the recommendation to for projects are most through (i) developing strengthen the project’s M&E frequently identified as standards for baseline framework, but finds the specific shortcomings in Bank support measurement, (ii) explicit actions recommended hard or xxxii MANAGEMENT ACTION RECORD at entry. Often what gets linking of the baseline to impractical to translate into measured are the project indicators relevant to monitorable actions. milestones (mainly outputs or project objectives, including intermediate outcomes) but any that refer to poverty or Establishment of key monitoring often not the final outcomes. inclusion impacts, and (iii) indicators that reflect intended project Early attention to project robust planning for outcomes, with baseline values, at the evaluability should strengthen monitoring data required outset of the project is critical to M&E planning and execution. for ex post evaluation. assessing project impact and monitoring progress. To the extent feasible, such data monitoring should be collected throughout the project implementation in a routinized and transparent fashion. A project should also have to state, at the decision review stage, how it plans to track key indicators and outcomes throughout the project’s life. Strengthening the project’s M&E framework and its implementation is a major agenda which will continue to be addressed through enhanced guidance and training for staff, and signaling from corporate and front- line management. xxxiii Chairperson’s Summary: Committee on Development Effectiveness The Committee on Development Effectiveness (CODE) met to discuss the Independent Evaluation Group’s (IEG’s) evaluation on The Poverty Focus of Country Programs and Draft Management Response. The Committee welcomed discussion of the evaluation, which drew lessons aimed at strengthening the Bank’s country diagnostics, improving the design of country strategies, and building greater learning opportunities from program experience. Members emphasized that country dialogue needs to be at the core of the World Bank Group’s work in the new country engagement model, as previously noted in CODE and other fora. They underscored the importance of listening to clients and ensuring the institution has the tools to be adaptive, to ensure constructive country-focused dialogues, and to implement selective country approaches. Members agreed with the necessity of high quality, accessible, and timely data, particularly at the beginning of the results chain; of addressing existing knowledge gaps; and of strong analytics and diagnostics. They noted that evidence-based decision making is at the core of the institution’s goal to be a solutions bank. Members supported the call for more effective collaboration with partners to strengthen data quality and availability. Further, they emphasized that the Bank Group needs to do more to build sustained capacity vis-à-vis data/diagnostics in client countries, particularly in the International Development Association’s context. Some commented that the Board needs more engagements with Management on the data agenda. Members also agreed that early attention to project evaluability should strengthen monitoring and evaluation planning and execution. It was recognized that the Bank Group is still at an early stage in implementing the country engagement model and, in many ways, the institution is learning by doing, and the tools underpinning the Systematic Country Diagnostic (SCD) and Country Partnership Framework (CPF) are being continually updated. Members welcomed Management’s concurrence with the main findings. They recognized that the new country engagement model, centered on the SCD and CPF, addresses some of the issues covered by the recommendations. Members asked Management to work with IEG in developing an action plan to address areas where it may be difficult to implement the recommendations. Several speakers also commented on the importance of enhancing the Bank Group’s poverty focus. They noted the multidimensional nature of poverty, the differences in poverty in low-income countries and middle-income countries, and accordingly the xxxiv CHAIRPERSON’S SUMMARY: COMMITTEE ON DEVELOPMENT EFFECTIVENESS necessities of having a broad toolkit; of engaging a broad range of stakeholders; and of embracing country approaches. Alejandro T. Foxley CHAIRPERSON xxxv 1. Assessing the World Bank’s Poverty Focus Reducing poverty has been a strategic objective of the World Bank Group since the 1970s, when President Robert S. McNamara first made it a priority. At the turn of the millennium, President James D. Wolfensohn again stressed the importance of this mandate. In 2013 President Jim Yong Kim extended this objective, setting two goals (the “twin goals”) to reduce extreme poverty to 3 percent of the world population by 2030 and, for the first time, including a distributional goal, to “share prosperity” by promoting the income growth of the poorest 40 percent. Over the past quarter century, and particularly since the advent of the Millennium Development Goals (MDGs) in 2001, the world’s rapid economic growth significantly reduced rates of extreme poverty and improved social indicators in many developing countries. A major milestone was accomplishing the first MDG— to halve the 1990 level of extreme poverty by 2015—five years early, lifting some 700 million people out of poverty. Globally, the proportion of people living in extreme poverty fell from 43 percent in 1990 to 19 percent in 2010, and further to 17 percent in 2011 (figure 1.1). Despite this progress, about 1 billion people still live in extreme poverty, and progress has been extremely uneven across Regions, countries, and localities (figure 1.1). The number of extreme poor fell during 1990–2011 in all Regions except Sub- Saharan Africa, where the number grew by more than 125 million, even as poverty incidence fell from 57 percent to 47 percent. All Regions except Sub-Saharan Africa are now expected to halve extreme poverty by 2015. Largely because of rapid population growth, the Sub-Saharan share of the world’s poor swelled from 15 percent in 1990 to 42 percent in 2011. Continuing success in reducing extreme poverty will become more difficult. Poverty is becoming more concentrated geographically: as of 2011, 1 billion people still lived in extreme poverty, with 415 million in Sub-Saharan Africa, 399 million in South Asia, 158 million in East Asia and the Pacific, and the remaining 2 percent scattered across the other Regions. About 17 percent of the world’s poor live in 36 fragile and conflict-affected states (FCSs) (World Bank 2014a). 1 An additional 1.7 billion people (or about 30 percent of the population in developing countries) are considered vulnerable to falling into extreme poverty because they live on $1.25 to $2.50 a day. 1 CHAPTER 1 ASSESSING THE WORLD BANK’S POVERTY FOCUS Figure 1.1. Extreme Poverty Fell Steadily during 1990–2011 Source: PovcalNet, the online tool for poverty measurement developed by the World Bank Development Research Group. Note: AFR = Africa Region; EAP = East Asia and Pacific Region; ECA = Europe and Central Asia Region; LCR = Latin America and the Caribbean Region; MNA = Middle East and North Africa Region; SAR = South Asia Region. Global poverty projections are highly sensitive to the underlying assumptions about growth and changes in income distribution. According to Bank research, if developing countries grow at the same annual average rate as they have during the past 20 years, global poverty will remain at around 6.8 percent of the world population in 2030 (World Bank 2014a). This implies that 202 million people (or 19 percent) in Africa, and a disproportionately high share of people in FCSs, will continue to be trapped in poverty. And there is strong evidence that poverty reduction becomes more difficult when the poverty rate is lower, because remote or intractable populations are more costly to reach. Achieving the Bank’s twin goals will require concerted action by developing countries, the Bank, and other members of the development community. Under plausible assumptions, growth alone will be insufficient to reach the targets, so new solutions will be needed that go beyond stimulating growth. Distributional changes will almost certainly be necessary to achieve the shared prosperity goal, 2 and sustainable elimination of extreme poverty will need to address the distribution of assets and asset returns. Meanwhile, the financial influence of the Bank Group and other development partners has steadily decreased, dwarfed by much larger international capital flows, mainly from the private sector. Official development assistance (ODA) as a percentage of gross domestic product (GDP) is small and has 2 CHAPTER 1 ASSESSING THE WORLD BANK’S POVERTY FOCUS declined over time, particularly in International Bank for Reconstruction and Development (IBRD) countries. Bank lending as a share of government spending is also small, particularly in IBRD countries, and is further diminishing in most developing countries. Given their reduced financial influence and in the absence of major new recapitalization of the multilateral development banks (MDBs), the World Bank and other development finance institutions can scale up their efforts to reduce poverty only by better leveraging their resources with government clients, global private funds, and other capital flows to developing countries. The comparative advantages of the Bank include its reputation for strong analytics, and its role as a “knowledge bank” to provide intellectual leadership and use its analytical work to influence policymaking. But the Bank works with governments that vary in their political commitment to poverty reduction and their base of domestic support. For this evaluation, political economy concerns consist mainly of the government’s commitment to poverty reduction, and its political will to measure and understand the nature of poverty and to identify and address key obstacles. Consideration of the specific country context and its political economy issues is critical to increasing the traction for poverty-reduction reforms: where there is strong commitment and capacity, the process of supporting poverty reduction is much more straightforward. But without strong government commitment, even the best-designed poverty-reduction strategies are bound to fail to achieve the desired results. This evaluation draws on the lessons from the poverty focus of recent country programs to show how Systematic Country Diagnostics (SCDs) can inform Country Partnership Frameworks (CPFs) to achieve the twin goals. Each CPF, which will define the Bank Group’s country engagements, will draw upon an SCD, which is designed to be a rigorous and independent exercise that Bank country teams conduct in consultation with national authorities and other stakeholders. The SCD will then become a reference point for client consultations on the priorities for Bank Group country engagements. The design and implementation of the SCD will identify the challenges and opportunities in each country. Evaluation Objective and Scope Since 2000 the Bank has monitored its contributions to development and poverty reduction, and in 2011 it published the first Corporate Scorecard and World Bank for Results report, assessing country results and organizational performance (World 3 CHAPTER 1 ASSESSING THE WORLD BANK’S POVERTY FOCUS Bank 2011a, 2011b). The Global Monitoring Report, published annually since 2004 by the International Monetary Fund (IMF) and World Bank, has provided additional information on progress in the MDGs (World Bank and IMF 2004). Despite this work, there has not been a comprehensive, independent evaluation of how the Bank has operationalized its poverty mission (World Bank 2005a). 3 This evaluation seeks to fill this gap. It assesses how the Bank’s programs have been designed and positioned to support partner country efforts to reduce poverty. The emphasis is on poverty, but also recognizing the importance of distribution, particularly in the context of the twin goals. 4 The evaluation for all International Development Association (IDA) and IBRD countries, as well as “blend” countries that get assistance from both, covers FY2004–2012. 5 The evaluation also identifies obstacles the Bank encounters when moving from data, to diagnostics, to strategy development and implementation. The findings aim to sharpen the effectiveness of country programs as the Bank starts to implement the post-2015 agenda, which will likely call for more ambitious action on poverty reduction. The emphasis of the evaluation is on the process by which the Bank has engaged with country governments to support poverty reduction. It assesses the Bank’s engagement with countries to generate and share data, prepare poverty diagnostics, use those diagnostics to formulate and implement strategy, and to monitor and evaluate feedback loops to inform future strategies. It does not assess the impact of the Bank’s intervention on poverty or how much poverty outcomes can be attributed to the Bank. Instead, it assesses how Bank programs have been designed and positioned to support country client efforts to reduce poverty. The evaluation examines both income and non-income dimensions of poverty (Box 1.1). This is important to the extent that the degree of poverty and its improvement over time may differ using different indicators, as the progress of some dimensions may lag others. This evaluation uses the income (or consumption) poverty lines to define the poor and examines the non-income dimensions of the challenges that the poor face. For ease of comparison across countries, the evaluation uses the extreme poverty line of $1.25 a day (at purchasing power parity 2005) for the income dimension and information from the World Development Indicators (WDI) database for the non-income dimensions, such as health, education, and access to basic services. For country studies, the evaluation uses the national poverty line as the primary threshold for income-based poverty. 6 To keep the analysis tractable and focused, this evaluation will exclude several important aspects of poverty, such as intrahousehold dynamics or distributional consequences within the household, and intertemporal tradeoffs related to climate change and environmental poverty. 4 CHAPTER 1 ASSESSING THE WORLD BANK’S POVERTY FOCUS Box 1.1. Measures of Poverty The evaluation examines both income and non-income dimensions of poverty. There is widespread consensus that poverty is a multidimensional phenomenon—besides low levels of income and consumption, poor people often suffer from fear, humiliation, and ill health. They may experience multiple forms of hardship simultaneously. However, there is much less consensus on whether it is useful to aggregate across different dimensions to construct a multidimensional measure of welfare and, if so, how to do so in a way that is conceptually sound and readily interpretable. Using income (or consumption) poverty lines and the Multidimensional Poverty Index (MPI) could result in divergent results in several countries. The hot debates about the MPI—including the opportunity of zooming in to see the different dimensions in which a person is deprived and then “adding up” that person’s deprivations; and the challenges of justification of the choice of sub indicators and weights—illustrate the pros and cons of using this composite indicator to measure poverty. Source: World Bank (2014b), Green (2010), Alkire (2010), and Ravallion (2010). All development assistance eventually has some relevance to poverty reduction, so the evaluation excludes some aspects of development to keep the analysis tractable and focused. It does not consider intrahousehold dynamics (such as distributional consequences within the household)7 or intertemporal trade-offs (such as climate change and environmental poverty). It also avoids the traditional portfolio review, because the entire portfolio of Bank interventions is aimed to either directly or indirectly be relevant to poverty. Further, it is concerned more with the poverty focus of Bank-supported interventions at the country level than with the achievement of specific project objectives. To support poverty reduction, the Bank needs a range of interventions supporting growth and equitable distribution, from improving basic services delivery to the poor to broad-based growth and competitiveness. The appropriate mix depends on the country context (box 1.2). Since the Bank Group had no comprehensive strategy for reducing poverty at the outset of the period reviewed, the evaluation does not examine the effectiveness of the Bank’s corporate strategy.8 The Bank’s technical support to regional and global agencies or to international forums on poverty is also excluded. Box 1.2. Defining the Poverty Focus of Bank Interventions Poverty is the result of economic, political, and social processes that interact and frequently reinforce each other in ways that exacerbate the deprivation in which poor people live (World Bank 2010a). It is well recognized that poverty eradication depends on both stimulating growth and providing basic social services to the poor (World Bank 5 CHAPTER 1 ASSESSING THE WORLD BANK’S POVERTY FOCUS 1990). Many of the Bank’s interventions can be considered relevant to poverty in some way. However, defining the poverty focus and the degree to which development support contributes to poverty outcomes is not straightforward. Some interventions, such as for safety nets, may contribute directly and immediately to reducing income poverty; others, such as support for education, may do so with a long lag; and still others, such as improvements to the investment climate, may contribute indirectly or only in the presence of other policies or dynamics in the economy. It is beyond the scope of the evaluation to examine the impact of Bank-supported interventions on poverty reduction in the short term and long term. Instead, the evaluation groups the Bank’s interventions into two broad categories—directly or indirectly focused on poverty reduction. Direct poverty focus is broadly defined as the activities that are designed and implemented to provide a disproportionate first-round benefit to the poorer segment of the population. The evaluation recognizes that the links between Bank interventions and poverty reduction are complex and country-specific, and the Bank faces trade-offs in selecting projects with direct linkages (such as social safety nets) or indirect linkages (such as financial reforms) with poverty reduction. It does not judge whether the interventions with direct poverty focus have a stronger impact on poverty reduction than those with indirect poverty focus. The report does not judge whether there should be a larger or smaller share of interventions directly or indirectly focusing on poverty reduction as the binding constraints vary across countries. Instead, it assesses the extent to which lending operations, technical assistance, capacity building, convening power, analytical work, and dialogue were focused on poverty reduction and its key country-specific dimensions as identified in the poverty diagnostics. The proper mix of interventions with direct and indirect poverty focus depends on the specific country situation. The development community increasingly recognizes the private sector’s crucial role in reducing poverty by creating jobs and providing services and opportunities for the poor. 9 The International Finance Corporation (IFC), the Bank Group’s private sector arm, is the largest development institution focused on the private sector. It provides direct financing and technical assistance to private enterprises and shares the Bank’s poverty-reduction mission. The poverty focus of IFC’s interventions are not assessed in this report since the Independent Evaluation Group (IEG) evaluation Assessing IFC’s Poverty Focus and Results (IEG 2011a) examines this in depth. The 2011 evaluation concludes that there is not enough clarity about what poverty means within the IFC strategic context. Although IFC’s interventions are designed to contribute to growth, distributional aspects are often not integrated into project designs. Most IFC investment projects generate satisfactory economic returns but do not provide evidence of how its interventions reach and affect the poor. The evaluation points toward a need for closer collaboration and greater synergies with the World Bank to better address poverty and distributional issues, beyond company-level impacts. Box 1.3 summarizes the key findings of the evaluation. 6 CHAPTER 1 ASSESSING THE WORLD BANK’S POVERTY FOCUS Box 1.3. IFC’s Poverty Focus and Results A 2011 IEG evaluation assessed the relevance and effectiveness of the poverty focus in IFC over fiscal years 2000 to 2010. In the context of IFC’s business model, the evaluation assessed poverty focus in terms of how IFC’s strategies, projects, and results measurement framework contribute to growth and to distributional patterns of growth that create opportunities for the poor. The evaluation found that IFC’s strategic priorities on frontier areas (areas with large concentrations of poverty) and sectors such as infrastructure, agribusiness, health, education, and financial markets are consistent with support to an inclusive growth pattern, but improvements are needed in three areas: Strategic Level. IFC aims to end poverty through private sector development. Three of IFC’s corporate strategic priorities—frontier markets, including International Development Association (IDA) and frontier regions in middle-income countries (MICs); targeted sectors with potential engagement of the poor; and certain types of financial services, focusing on micro, small, and medium enterprises (MSMEs)—have guided its poverty focus since 2004. The evaluation found that IFC investments in IDA countries have increased over the years; however, these investments were allocated in a few IDA countries, such as India, Nigeria, Pakistan, and Vietnam. The investments need to be allocated in more IDA countries. In MICs, IFC focuses on frontier regions with large concentrations of poverty. However, IEG analysis showed that the largest density of poor people was not in locations with the highest poverty rates. While the strategic sectors are broadly defined in terms consistent with a pro-poor orientation, they need to be designed and implemented in ways that actually enhance opportunities and the impact on poor people. IFC’s strategic directions related to poverty have a special focus on support to MSMEs, through financial intermediaries, which grew eightfold in the 2000s. The efficacy and welfare impacts of these types of interventions are not known. Operational Level. The majority of IFC projects are designed to contribute to growth, but IFC found it challenging to incorporate distributional issues in its interventions. Fewer than half of projects reviewed included evidence of poverty reduction and distributional aspects in project objectives, targeting of interventions, characteristics of intended beneficiaries, or tracking of impacts. Projects that paid attention to distribution issues performed at least as well as other projects on development and investment outcomes; this suggests that poverty focus need not come at the expense of financial success. The relatively high proportion of projects that do not generate identifiable opportunities for the poor suggests that they rely primarily on the pace of growth for poverty reduction, at a time when IFC’s strategies point to more attention to the pattern of growth. Impact through a Poverty Lens. IFC’s monitoring and evaluation (M&E) framework did not quantify benefits to the poor and there were no indicators for measuring a project’s effect on poverty. Given the limited attention to distributional issues in the M&E framework, IEG used a poverty index to characterize project benefits based on their contribution to growth and inclusion of the poor. The analysis indicated that the majority of investment projects generated economic returns but did not provide evidence of identifiable benefits to the poor. This further suggests that IFC primarily relies on the pace of growth for poverty reduction. 7 CHAPTER 1 ASSESSING THE WORLD BANK’S POVERTY FOCUS Source: IEG (2011a). Results Chain, Evaluative Questions, and Instruments The results chain through which the Bank supports economic development to reduce poverty can be long and differs significantly from country to country depending on initial endowments, social structures, quality of governance, economic systems, and global circumstances. It is also widely recognized that development outcomes are the result of the capacity and ownership of client countries. The support of multiple partners and interventions across sectors and time often complicates the attribution of results to a single partner or initiative. Still, in practice the Bank uses its levers of support—analytical work, lending, convening power, technical assistance, capacity building, and policy dialogue—to influence national policies and programs, and to help translate growth, greater access to opportunities, and poverty alleviation mechanisms (such as social safety nets) into poverty reduction. Bank programs and projects are required to specify results that are expected from the actions and policies they support, and most have a results framework. The overarching question for this evaluation is: “How, and how well, has the Bank focused its programs on reducing poverty in partner countries?” Using country programs as the primary unit of analysis, the evaluation reviews the Bank Group’s Country Assistance Strategies (CASs) and Country Partnership Strategies (CPSs) to determine the extent of their focus on poverty and the poor and assesses them based on their consistency with analytical work on poverty for that country. The evaluation is structured around five questions underlying this main line of inquiry: • Did the Bank have the appropriate data to understand the nature of poverty and provide an information base for robust analytical work on poverty? • With the given data, did the Bank’s analytic work adequately address poverty issues and identify policy priorities for poverty reduction? • Did Bank country strategies adopt the findings of analytical work on poverty to help set priorities for and guide policy dialogue and lending? • Did interventions—operations, technical assistance, and capacity building— reflect the strategic priorities for poverty reduction? • Did the Bank collect and draw lessons from poverty-reduction interventions to strengthen feedback loops and improve the effectiveness of its country strategies and programs? If so, how did it do so? 8 CHAPTER 1 ASSESSING THE WORLD BANK’S POVERTY FOCUS The report is structured along the evaluative chain from data, to diagnostics, to strategy formulation and implementation, and to the feedback loops (figure 1.2). The evaluation fully assesses each of the components in the process of country program work related to poverty reduction, tracing through data, diagnostics, country strategy formulation and implementation, and feedback loops in the respective chapters. The analysis relies on 10 country case studies, an internal survey with Bank staff, an external survey with client government officials and nongovernment stakeholders in 20 countries, a series of focus group meetings with Bank staff who work on poverty, an assessment of the quality of Poverty Assessments (PAs) in 20 countries, a review of the Country Partnership Strategy Completion Report Reviews (CPSCRRs) and Country Assistance Evaluations (CAEs), and stocktaking exercises of PovcalNet data, development policy operations, Investment Lending (IL), Public Expenditure Reviews (PERs), and Poverty and Social Impact Analyses (PSIAs). Country case studies. The 10 country case studies drew in-depth lessons from the Bank’s support of government efforts to reduce poverty (appendix A). 10 The countries were selected from a population of 144 countries comprising all IDA, IBRD, and blend countries. 11 The selection is purposive and does not aim to fully represent the various categories of countries. It tries to cover a range of countries at different income levels to reflect the differing approaches and challenges to poverty reduction in countries at different levels of development. The case study countries were selected to roughly reflect regional balance. During the selection process, Figure 1.2. Evaluation Chain of the Poverty Focus of Country Programs 9 CHAPTER 1 ASSESSING THE WORLD BANK’S POVERTY FOCUS countries were first grouped according to (i) regions; (ii) income level; and (iii) whether or not they are classified as FCS. 12 An emphasis was placed on countries with significant Bank engagement (lending and nonlending activities) 13. To provide lessons that reflect a wide range of operational experience, the final selection of countries also took into consideration variations in the number of poor and in poverty rates, and potential lessons for learning in consultation with external experts and senior Bank staff. The 10 countries selected for study are Arab Republic of Egypt, Bangladesh, Guatemala, Lao People’s Democratic Republic (Lao PDR), Malawi, Nigeria, Peru, the Philippines, Romania, and Senegal. Focusing on FY2004–2012, each case study consisted of desk reviews, structured interviews with Bank staff, and in-country consultations with stakeholders. External stakeholder survey. This survey 14 collected views and feedback from government officials, civil society, academia, the donor/international community, and the private sector. The list of respondents was randomly generated for each country from the list of respondents of the Bank’s stakeholders shared by the respective Country Office teams (which in some cases were drawn from the list of recent Bank’s country client surveys), IEG evaluation interview lists, and complemented by research conducted by the independent survey firm. Heavier weight in the respondent list was given to the government officials, since they are the World Bank’s primary counterparts. The 20 countries surveyed covered all six Regions and reflect a balance between types of countries by lending group (that is, IBRD, IDA, and blend), accessibility of data, and FCS and non-FCS. 15 A 27 percent response rate is on a par with similar stakeholder surveys administered by international organizations. 16 (appendix B). Internal staff survey. This survey sought insight into staff perspectives of the constraints on data and the drivers of the difference in quality of poverty data and diagnostic work. The survey focused on three areas: constraints to obtaining poverty data; best practices and challenges to creating poverty diagnostic work; and challenges of translating poverty diagnostic work into country strategies. The survey was sent to all staff at grades GF and above who were working on poverty issues in the Bank’s operations and research departments. 17 An analysis of the distribution of survey respondents shows that it largely mirrors the distribution of the population in terms staff technical mapping, grade levels, and location (headquarters vs. country office). A response rate of 21 percent is on a par with similar surveys administered within the Bank 18 (appendix C). Focus group meetings. The six focus group meetings gathered in-depth information about how the availability of poverty data affects the Bank’s poverty diagnostics, and whether and how they translate into country strategies (appendix D). The 10 CHAPTER 1 ASSESSING THE WORLD BANK’S POVERTY FOCUS representatives at the meetings included three distinct groups: task team leaders or co-leaders of Poverty Assessments/Poverty Updates, country economists of countries that have not had a poverty assessment since 2009, and task team leaders or co-leaders of CPSs. The 18 countries represented in the discussions were purposively selected to cover all six Regions considering the balance between the types of countries by lending group, accessibility of data, and FCS and non-FCS countries. The selected countries were Afghanistan, Botswana, Cambodia, the Democratic Republic of Congo, Jordan, India, Indonesia, Kosovo, the Kyrgyz Republic, Moldova, Pakistan, the Republic of Yemen, Russia, South Sudan, Sri Lanka, and Papua New Guinea, along with West Bank, Gaza, and the member countries of the Organization of Eastern Caribbean States. 19 Assessment of the quality of Poverty Assessments/Poverty Updates. This assessment sought to identify gaps in the quality of Bank poverty diagnostics across Regions, across those countries with the greatest poverty-reduction challenges (in absolute or relative terms), and countries with weak data 20 (appendix E). Twenty countries were purposively selected for this assessment to provide equal coverage of each of the Bank’s six Regions (four countries each for Sub-Saharan Africa and East Asia and Pacific, and three countries each for the remaining four Regions); countries with greater rates of poverty either as a proportion of the developing world’s poor, or as a share of country population; and at least one country with weak data in each Region.21 Review of CASCRRs/CPSCRRs and CAEs/CPEs. This review assessed the poverty focus of country strategies and their M&E systems as reflected in the completion report. The analysis was a systematic review of all of IEG’s CPSCRRs since FY2004. The review draws on the comparison of the CPSCRRs from 66 countries, with two CPSCRRs each in the past decade to examine the trends in the poverty focus of country strategies and the 14 CAEs/CPEs relevant to the period of evaluation (appendix F). 1Given the extremely poor data on poverty and growth in many of the 36 countries on the World Bank’s FCS list in 2014, this number is a crude estimate. 2 See more details in World Bank 2014a. 3 The2004 Operations Evaluation Department (now IEG) Annual Review of Performance looked into the Bank’s contribution to poverty reduction, but in a narrow way (see paragraph 14). evaluation focuses on poverty and IEG plans to conduct an evaluation on shared 4 This prosperity separately in FY2017. 11 CHAPTER 1 ASSESSING THE WORLD BANK’S POVERTY FOCUS 5 The population includes 144 countries, listed as Bank borrowers as of October 2013. 6This was done, first, because the evaluation focuses on improving the Bank’s support at the level of the country program. The basis of this dialogue is the national poverty lines of the countries involved. The second reason is practical: much of the country-specific analytical work and dialogue is based on the national poverty line, without reference to international poverty lines. For countries that have adopted two poverty lines a nationally defined standard poverty line and a nationally defined extreme poverty line, the evaluation emphasizes the extreme poverty line, as defined by each country. 7A separate forthcoming evaluation on poverty and gender is examining intrahousehold aspects of social safety nets. 8In practice, the Bank has favored a broad, multisector and multistakeholder approach to achieving development results and poverty reduction. There are multiple development or poverty strategies at the regional and sector levels across the Bank. In addition, the Bank has not precisely committed to any particular component of the MDGs. 9“[P]rivate investment has become the dominant mode of capital transfer worldwide” (World Bank 2014b). “The private sector is the main engine of growth.” OECD 2007. “[P]rivate sector is already central to the lives of the poor and has the power to make those lives better” (UNDP 2004). 10Given the scope of work and resource constraints, the evaluation focused on 10 country case studies to draw in-depth lessons. The countries are purposively selected and not statistically representative. Every effort is made to ensure no bias is introduced in the selection process. 11 The IBRD, IDA, and blend country classification is up to date as of October 2013. 12 Fragileand conflict-affected states have either: (a) a harmonized average CPIA country rating of 3.2 or less, or (b) the presence of a UN and/or regional peace-keeping or peace- building mission during the past three years. This list is up to date as of FY14 and includes only IDA-eligible countries and nonmember or inactive territories and countries without CPIA data. It excludes IBRD-only countries for which the CPIA scores are not currently disclosed. 13Significant levels of Bank engagement are indicated by a minimum of 25 approved projects between FY2004 and 2012, since 25 projects is the median number of projects for the population of the 144 countries during this period. 14 To ensure confidentiality and anonymity of the responses as well as respondent bias, IEG commissioned ICF International, an independent survey firm, to conduct a web-based survey to gather opinions and perceptions of the World Bank’s poverty through a survey in 20 client countries. See appendix B for details. 15 Thecountries selected are China, Democratic Republic of Congo, the Dominican Republic, Ethiopia, Ghana, Haiti, Honduras, India, Jordan, Kosovo, the Kyrgyz Republic, Mexico, Morocco, Nepal, Papua New Guinea, South Africa, Sierra Leone, Rwanda, the Russian Federation, and the Republic of Yemen. 12 CHAPTER 1 ASSESSING THE WORLD BANK’S POVERTY FOCUS 16For example, the World Bank’s 2013 Access to Information Stakeholder Survey had a 27 percent response rate and the Client Feedback Survey of FY13 Analytical & Advisory Activities had a 31 percent response rate. 17 Thesurvey was conducted from April 15 to May 13, 2014, and was sent to 4,150 Bank staff, of which 866 responded (a rate of 21 percent). To focus on staff with close experience in operations and country strategies, staff mapped to procurement, human resources, information and technology, business solutions, World Bank Institute, and IEG were exluded. The survey was confidential and anonymous. See appendix C for details. 18 Forexample, DEC’s research paper on the influence of World Bank Research fielded a survey among all senior operations staff at grades GG and above and received a 19 percent response rate (555 respondents out of 2,900) (see Research at Work: Assessing the Influence of World Bank Research. Development Economics Unit. World Bank. 2012.) In addition, the WHO Stakeholder Perception Surveys set the response rate for the internal staff survey at 20 percent as a good threshold (see http://www.who.int/about/who_perception_survey_2012.pdf) 19 Although all of the participants from relevant focus group categories in each country were invited to participate, the final list depended on the staff availability in June-August 2014. The total number of focus group participants was 22 staff members. 20 Forreasons of comparison, this evaluation refers to data availability as the survey data availability in the World Bank’s PovCalNet. Countries with weak data broadly refers to countries that have, at most, one round of household surveys available in PovCalNet during the period of 2000–2012. evaluation also examines in country case studies the quality of other analytical work 21 The on poverty, including from the quality of public expenditure reviews, poverty and social impact analyses, and the poverty chapters of the Country Economic Memoradum. 13 2. Preparing the Ground: World Bank Support for Poverty Data Highlights  The availability of poverty survey data has steadily improved in recent decades, but progress has been highly uneven, and data quality remains weak in many developing countries.  Partial coverage, lags in periodicity, and poor timeliness characterize countries with weak data capacity and constrain understanding of poverty profiles and trends.  Consistency and comparability of poverty data across indicators remains a problem. Gaps between income and consumption data from household surveys and from national accounts are also a problem.  The World Bank made poverty data more widely accessible by advocating for open data and developing tools to document, catalog, and disseminate micro data. But it would increase its effectiveness through a more coherent and systematic approach to data collection and reporting.  The World Bank has made important contributions to improving statistical systems, and has provided technical assistance, often in collaboration with donors. However, sustainability of improvements when financing ends remains an issue. Reliable and timely data are key to measuring poverty. Acquiring such data is the first step in understanding the nature and magnitude of the challenges of poverty reduction, which is essential for policy planning and design of targeted interventions for poverty reduction. This chapter reviews the state of country-level poverty data. It examines whether the Bank has sufficient information to produce robust diagnostic work, and if the Bank has helped to expand and improve the collection, dissemination, and use of poverty data. The chapter also discusses the challenges in data beyond household surveys. The evidence is drawn primarily from the country case studies, staff survey, external stakeholder survey, and focus group meetings, which were all triangulated with information from the Bank’s micro data catalog and other data sources. State of Survey Data Poor people are subject to many kinds of deprivation. This evaluation uses the income (or consumption) measure to set the poverty line, and examines the causes and consequences of poverty in multiple dimensions. It also considers a range of other indicators that characterize and influence the welfare of individuals (box 2.1). 14 CHAPTER 2 PREPARING THE GROUND: WORLD BANK SUPPORT FOR POVERTY DATA Box 2-1. What are Poverty Data? Many kinds of data are needed to assess the conditions of poor people and to design programs that improve their circumstances. Surveys and administrative sources provide many non-income measures of poverty, such as educational outcomes, health status, living conditions, labor force participation, and environmental conditions. Demographic data from censuses and civil registration and vital statistics systems are useful in constructing sampling frames and sample weights, and for normalizing data from administrative records. Data from consumer price surveys and income aggregates from national accounts are needed to extrapolate income and consumption data between survey rounds. Global price indexes based on purchasing power parities are needed to calculate international poverty lines. Well-functioning, complete national statistical systems should be able to produce all these statistics through national censuses, administrative records, and national and international survey programs. However, many developing countries need help to produce statistics that meet the standards required for international reporting. The Multiple Indicator Cluster Surveys (MICS) sponsored by UNICEF and Demographic and Health Survey (DHS) sponsored by the U.S. Agency for International Development have helped fill this gap. Labor force surveys also yield relevant information on employment status, occupation, and wages. The International Labor Organization provides standards and guidelines for labor force surveys and compiles results from national surveys. Efforts are also under way to create a harmonized global database of labor force surveys. Comprehensive surveys such as the Living Standards Measurement Study (LSMS) collect data on key non-income characteristics of households, household members, and their communities. These are used to construct poverty diagnostics. The LSMS questionnaire cannot accommodate all the items needed to reproduce the output of a DHS, MICS, or labor force survey, and surveys cannot replace all the data from administrative records. Coordinating statistical activities in a country is necessary to produce data to generate the statistics needed for a comprehensive view of poverty. During the past three decades, the availability, quality, and accessibility of Household Income and Expenditure Surveys (HIES) have improved for most developing countries. The World Bank has supported capacity building for data collection and management and poverty estimation in its country clients. In the stakeholder survey for this evaluation, 86 percent of government and 90 percent of nongovernment stakeholders believed that the Bank has helped to improve the quality of poverty data. 15 CHAPTER 2 PREPARING THE GROUND: WORLD BANK SUPPORT FOR POVERTY DATA Without more frequent data collection, changes in poverty and income distribution can be measured only indirectly. Data quality issues, particularly comparability across survey rounds and the availability of panel data, remain serious constraints to understanding poverty in many countries. 1 Data accessibility and transparency also remain a problem, particularly for fragile and conflict-affected states (FCS) and in countries where poverty is a politically sensitive and unwelcome topic. Although there is little consensus on the best way to cost the requirements of the global statistical system, an estimate by Demombynes and Sandefur (2014) asks for about $300 million per year to close the financing gap for surveys for countries below $2,000 gross domestic product per capita in purchasing power parity (PPP) dollars (appendix H). DATA AVAILABILITY After publishing the World Development Report 2000/2001: Attacking Poverty, 2 the World Bank highlighted the need to expand the coverage and improve the quality and accessibility of household income surveys, calling for more detailed questionnaires. The stock of household surveys, which capture the most common sources of poverty statistics including income, consumption, and expenditure data, greatly increased during the past three decades. 3 During the 2000s, some 40–50 surveys were conducted each year, up from the 20–30 surveys in the 1990s and less than 10 per year through most of the 1980s. Surveys have also become more sophisticated: LSMSs have incorporated new modules to collect information on non- income correlates and subjective measures of poverty. Multiple rounds of surveys allow for the assessment of poverty trends, whereas a single round permits only static profiling. 4 But monitoring progress requires continual data updates. After reaching a peak in 2002, the number of new income and expenditure surveys added to the Bank’s PovcalNet database leveled off at about 50 per year, and the number of new surveys recorded within a five-year period peaked at 229 in 2006. More surveys may exist, but long delays between the initiation of surveys and the recording and processing of micro data caused a decline in the number of surveys available in recent years. As a result, no data are available in PovcalNet after 2012, and country coverage going back to 2008 is reduced (figure 2.1). Despite the increase in survey coverage, 22 low-or middle-income countries do not have income or consumption data in the past three decades (or since the 1980s) in PovcalNet. Some 30 countries have not had household surveys and some 20 more have had only one survey in the past decade (or since 2000). The lack of survey data limits the ability of these countries to measure the profiles and trends of their poverty.5 More than one-third of these countries are in Sub-Saharan Africa, and more 16 CHAPTER 2 PREPARING THE GROUND: WORLD BANK SUPPORT FOR POVERTY DATA than four-fifths are small or FCS. Among the 30 to 50 new surveys added each year since 2000, 80 percent were in middle-income countries. In Latin America and the Caribbean, the average gap between two surveys is slightly more than two years, but in Sub-Saharan Africa (where poverty is concentrated) the gap is almost nine years. The staff survey confirms that wide data gaps are a big concern to development practitioners in FCS. Figure 2.1. After Rising through 2002, the Number of New Household Income and Expenditure Surveys Leveled Off Source: PovcalNet. DATA QUALITY Producing high-quality data requires significant effort (box 2.2). Challenges in survey data can range from design and implementation problems in a single survey round (which prevent the identification of a comprehensive poverty profile) to the lack of survey comparability over time (which hinders the examination of poverty trends), and to problems in constructing panel data (which limit the possibility of tracking poverty mobility and vulnerability). High-quality panel survey data are important for policy relevant-analysis (such as chronic versus transient poverty and the vulnerability of different types of households) and to trace the medium- or long- term impact of interventions on well-being. Obtaining accurate information on sources of income and on consumption expenditures is a particular challenge. The reliability of consumption data is affected by the timing of surveys and the recall period or the type of diaries used to record expenditures. The consistency and comparability of data can also be improved by adhering to international standards for classifying consumption components. 6 17 CHAPTER 2 PREPARING THE GROUND: WORLD BANK SUPPORT FOR POVERTY DATA The responsibility for data quality lies with national statistical offices (NSOs) and the governments that authorize and fund them. Evidence points to the need to improve their capacity and performance. The 2014 report by the UN Secretary- General’s Independent Expert Advisory Group on the Data Revolution for Sustainable Development recommends that “data quality should be protected and improved by strengthening NSOs” (IEAG 2014: 23). It declares that “a robust framework for quality assurance is required, particularly for official data. This includes internal systems as well as periodic audits by professional and independent third parties. Existing tools for improving the quality of statistical data should be used and strengthened, and data should be classified using commonly agreed criteria and quality benchmarks” (IEAG 2014: 22). The World Bank has contributed to improving data quality in many countries. With that assistance, overall household survey data quality has improved. In a majority of case study countries, data comparability, disaggregation, and depth have improved. For instance, during the past two decades, the improved quality of the Lao Expenditure and Consumption Survey data provided a strong information basis for poverty diagnostics in the 2000s. 7 Box 2-2. Assessing Data Quality The quality of poverty estimates and the diagnostic indicators and policies based on them depend on the quality of the underlying data. The best assurance of data quality is knowledge of how they were produced. The International Monetary Fund developed a generic data quality assessment framework (DQAF) that can be applied to many of the outputs of an official statistical agency (IMF 2012). The prerequisite for data quality is a legal and institutional environment that authorizes the work of a statistical office and provides adequate resources. The DQAF then defines five dimensions of data quality:  Assurance of integrity demonstrated by impartiality, transparency, and ethical behavior  Methodological soundness, including the use of appropriate standards and classification systems  Accuracy and reliability of source data and the processes used to compile statistics  Serviceability, which looks at the periodicity, timeliness, and consistency of statistics within a dataset  Accessibility of data and metadata to the final users. The lack of a DQAF or equivalent does not imply that quality has been neglected. Expert advice from World Bank teams and others has helped improve survey programs. The International Household Survey Network and its Accelerated Data Program have encouraged countries to provide robust information on the design and execution of surveys, using the Data Documentation Initiative (DDI) standard (see website at 18 CHAPTER 2 PREPARING THE GROUND: WORLD BANK SUPPORT FOR POVERTY DATA http://www.ddialliance.org). Full compliance with the DDI provides evidence of the quality of micro data sets. Progress on improving data quality is evident in some of the country cases, which is making the development of rigorous analytics possible. Peru and Romania now conduct high-quality, nationally representative household surveys annually. Also, some low-income countries with capacity constraints produce high-quality poverty data with high comparability over time. In Malawi, the last two rounds of household surveys were fully comparable and representative at the national level and (three) sub regional levels, as well as urban/rural zones. In Bangladesh, the Household Income Expenditure Survey used a comparable set of questions and sample frames in the last three rounds. This allowed detailed examination of the dynamics and determinants of poverty incidence, supporting rigorous analysis. But the five-year gap between surveys leaves analysts and policymakers uncertain about the impact of new policies and programs. In some other case study countries, data are much weaker and often not comparable over time, presenting challenges to analysis. Senegal conducted three nationally, regionally, and zonally (rural/urban) disaggregated household surveys between 2000 and 2012, but they are not fully comparable. 8 In the Philippines, inconsistencies between the definitions of “rural” and “urban” in successive surveys complicated, or even invalidated, the comparability of the data over time. In Nigeria, the Harmonized Nigeria Living Standard Survey (HNLSS) 2009/10 data were inadequate because of changes in the method for estimating expenditure and the way the survey was carried out. Overall, therefore, progress on data quality appears to have been uneven. Bank staff view weak survey data as the main constraint to carrying out robust poverty diagnostics. For instance, more than one-third of the staff survey respondents indicated that the lack of sufficient data is an obstacle for creating Poverty Assessments and Poverty and Social Impact Analyses (figure 2.2). 19 CHAPTER 2 PREPARING THE GROUND: WORLD BANK SUPPORT FOR POVERTY DATA Figure 2.2. Challenges Facing Bank Staff Conducting PAs and PSIAs Source: Bank Staff Survey. Note: PA = Poverty Assessment; PSIA = Poverty and Social Impact Analysis. The challenges are even more daunting when considering the need to measure the income growth rate of different segments of the population to monitor progress in reducing extreme poverty and improving shared prosperity. As reported in the Global Monitoring Report 2014/15: Ending Poverty and Sharing Prosperity, the Bank has information about the growth rate of the bottom 40 percent during the period 2006– 2012 for only 86 countries. The missing information in the upper tail of the distribution in the HIES (and in some countries also the lower tail, related to, for example, informality) is an obstacle to measuring poverty and for measuring changes in expenditure or income distribution over time. DATA ACCESSIBILITY Data accessibility has improved considerably in the past decade. With support from other agencies and partner countries, the World Bank’s Open Data Initiative has been influential, encouraging the release of a variety of statistical information.9 The Accelerated Data Program has assisted partner countries in retrieving and documenting surveys that had been lost or forgotten. Information about these surveys is now available to the public through the International Household Survey Network (IHSN) catalog and internally to Bank staff through the central Micro data Catalog. As countries have adopted open data polices, statistical offices have become more willing to make their poverty data and documentation available through websites and in other public venues. In Peru, for example, the statistical agency provides unrestricted access to survey data, which is reported using multiple methodologies to 20 CHAPTER 2 PREPARING THE GROUND: WORLD BANK SUPPORT FOR POVERTY DATA allow for comparability over time. Data from Malawi’s 2010 Integrated Household Survey were made public within 12 months of completing fieldwork. But the absence of the mandate to require the collection and dissemination of standardized data on household income and poverty constrains the Bank’s ability to achieve the twin goals (Box 2.3). Restrictions on data access often remain severe. In Bangladesh, although the main results and documentation of the HIES are provided on the website, micro data cannot be downloaded, thus restricting public access. 10 In the Philippines, income and expenditure surveys are publicly available, but the release of poverty estimates can take up to two years after survey completion. In Senegal, even sector ministry and subnational government officials often lacked access to the poverty data or did not know that the data existed. Development partners in Senegal noted the lack of official access to poverty data and instead sought to access it informally from Bank staff. Box 2-3. The Missing Mandate The IMF is the international financial institution charged with, among other things, collecting and reporting key international finance, public finance, and trade data from all of its member countries. Under the Articles of Agreements VIII General Obligations of Members, the IMF may require member countries to furnish information as it deems necessary, including national data such as total exports and imports of merchandise, international balance of payments, international investment position, national income, and price indices. This helps the IMF provide a critical public good, collecting and publishing information on monetary and financial issues, including financial risk and credit ratings, policy formulation during financial crises, periodic identification of policy priorities and financial health, assessment of basic research on macroeconomic stability, and basic research on international and public finance. The IMF also has taken several important steps to enhance transparency and openness, including setting voluntary standards for the dissemination of economic and financial data. Over 95 percent of the IMF's member countries participate in the General Data Dissemination System or Special Data Dissemination Standard. High quality, standardized, and comprehensive financial data support the IMF’s mission, and serve as the hallmark of its international contribution to knowledge on of the global economy. By contrast, the World Bank Group has no such requirement for collection and dissemination of standardized poverty data. While many countries collect such data, and the World Bank has become a global leader supporting poverty data development and analysis, there remain major gaps in country coverage, timeliness of data, and quality standards. These gaps seriously constrain the Bank’s ability to generate knowledge on poverty. The Bank’s capability of analyzing public policy issues, measuring the human dimensions of risk from global shocks, or conducting research and strengthening targeted poverty reduction strategies has been compromised. Adopting a clear mandate, establishing strong standards and monitoring mechanisms, putting in place a well- defined strategy and incentives to support the collection and reporting of high-quality 21 CHAPTER 2 PREPARING THE GROUND: WORLD BANK SUPPORT FOR POVERTY DATA poverty data, are needed to support the Bank’s commitment to “eliminating extreme poverty and boosting shared prosperity.” Source: Cited from Articles of Agreement of the International Monetary Fund (IMF), (https://www.imf.org/external/pubs/ft/aa/#art8) and from IMF Standards for Data Dissemination (http://www.imf.org/external/np/exr/facts/data.htm). Political will is an obstacle to making poverty data public. Where poverty is a sensitive topic, the constraints are worse. In Egypt, the Bank has only partial access to the Central Agency for Public Mobilization and Statistics data through selected consultants identified by the government. Often, and particularly in FCS economies, the release of poverty data hinges on whether the conclusions of the survey are favorable to political constituencies. While the Bank managed to produce some analytical work of high quality, data challenges limited their scope and possible impact. Even in countries without legal or political restrictions, the dissemination of micro data can be inhibited by concerns that the data might breach the confidentiality of respondents. New methods of making data anonymous prevent the identification of individuals, and the World Bank is helping Senegal and other countries to implement those methods. World Bank Support for Data Capacity Building The Bank is generally seen by its country clients and development partners as a global leader and valued partner in building capacity to improve the availability, quality, and accessibility of poverty data. World Bank staff and government officials mention insufficient capacity and government budget as key obstacles to collecting poverty data (figures 2.3 and 2.4). Client demand for support with data capacity building is strong, and the Bank is well positioned to help meet that demand. The Bank has been a major contributor to poverty data collection and to methodology improvement through programs like the LSMS since the 1980s, and more recently the LSMS-Integrated Surveys on Agriculture. To address deficiencies in the documentation of survey data, the World Bank organized the IHSN in 2004, a partnership of household survey sponsors. Secretariat responsibilities are shared between the World Bank and the Partnership in Statistics for Development in the 21st Century. 11 A complementary program, the Accelerated Data Program (ADP), was set up to provide technical assistance to countries adopting IHSN tools and standards to better manage their household survey data. A recent evaluation shows IHSN and ADP have built strong partnerships and an enthusiastic following (Thomson, Eele, and Schmieding 2013). 22 CHAPTER 2 PREPARING THE GROUND: WORLD BANK SUPPORT FOR POVERTY DATA In 1999 the World Bank established a Trust Fund for Statistical Capacity Building (TFSCB) to strengthen the capacity of national statistical systems in developing countries. Of the 107 countries with National Strategies for Development Statistics (NSDS), 80 received grants (World Bank 2014c). A large share of TFSCB grants went to countries developing and implementing the NSDS. Other projects have also supported a wide range of activities for implementing the NSDS. Peru, for example, received support for survey data collection on Afro-Peruvians, for improved coverage, quality, and timeliness of vital statistics, and for implementing its open data program. The Philippines received grants for capacity building in the rural sector and for improving the quality of its national accounts. Figure 2.3. Main Constraints to Poverty Data Collection: Perspectives from World Bank Staff Source: Bank Staff Survey. 23 CHAPTER 2 PREPARING THE GROUND: WORLD BANK SUPPORT FOR POVERTY DATA Figure 2.4. Main Constraints to Poverty Data Collection: Perspectives from World Bank Clients (Governments) Source: External Stakeholder Survey. TECHNICAL SUPPORT FOR DATA CAPACITY BUILDING To monitor and report on country capacity to provide economic and social statistics, the World Bank maintains the Bulletin Board on Statistical Capacity. 12 A Statistical Capacity Indicator (SCI) uses a weighted average of scores on 25 items, and rates countries for their “statistical methodology,” “source data,” and “periodicity and timeliness.” A higher score indicates a stronger statistical system and greater ability to carry out the work needed to produce reliable statistics. Countries with higher adjusted overall SCI scores have more surveys of sufficient quality (figure 2.5). 13 The six countries with adjusted scores of 90 or above each had more than 10 surveys per country in PovcalNet between 2000 and 2012; those with adjusted scores below 50 had slightly more than one each. With support from other development partners, the World Bank provides technical support to improve the availability, quality, and timely access of data, which is evident in the country case studies. In Bangladesh, the Bank’s long-term engagement with statistical agencies since the early 1990s strengthened local statistical capacity. In Guatemala, the Bank’s capacity-building effort helped create a critical mass of technical expertise in the National Statistics Institute and the government’s planning secretariat, resulting in a vast improvement in the quality and transparency of data. In Lao PDR, the Bank provided extensive technical and financial support to improve the capacity of the statistical agency, improving the 24 CHAPTER 2 PREPARING THE GROUND: WORLD BANK SUPPORT FOR POVERTY DATA Lao Expenditure and Consumption Survey and administrative data, which allowed for deeper and more multidimensional poverty diagnostic work. Figure 2.5. Capable Statistical Systems Produce More Surveys Source: Bulletin Board on Statistical Capacity. Note: SCI overall scores were adjusted to remove two indicators related to poverty surveys. The adjusted SCI overall scores were re-normed to a range of 0 to 100. The monitoring program for the proposed Sustainable Development Goals is expected to create greater demand for relevant statistics. New methods of data collection—including remote sensing and “big data” derived from cellphone records and social media—are expected along with new modes of international support and coordination. 14 In this context, the World Bank’s poverty data work is part of a much larger effort to support improvement of data sources for measuring poverty and obtaining the evidence for effective policymaking. The Bank also builds consensus among technical and political players to improve survey design, implementation, and poverty estimation methodologies. In Peru, several changes in methodologies for estimating the poverty line compromised the credibility of poverty data in the early 2000s. A Bank team, in addition to providing technical support, worked closely with Peruvian researchers and public servants to set up an advisory committee in 2007 to help reach consensus on the best methodological practices to produce comparable poverty estimates. As a result, the Instituto Nacional de Estadística e Informática (INEI) issued comparable poverty figures, and public trust was restored. Building sustainable institutional capacity alongside technical capacity was key to INEI’s success. 25 CHAPTER 2 PREPARING THE GROUND: WORLD BANK SUPPORT FOR POVERTY DATA IMPORTANCE OF POLITICAL WILL Political will is a major constraint to poverty data development in some countries. Nearly 40 percent of Bank staff cited the lack of political will as an obstacle to obtaining data for assessing poverty levels. 15 In Nigeria, official commitment was lacking and the government failed to finance the HNLSS 2009/2010. Insisting on cost sharing, the World Bank and U.K. Department for International Development withdrew support and the survey was unsuccessful. In Egypt, poverty headcount data (and inequality estimates) have been tied to the political discourse of successive governments. 16 Data accessibility for the general public had been a major issue at least until the late 2000s—assumptions of the low domestic labor mobility and the accuracy of the poverty estimations in the major shantytowns and areas around metropolitan areas were questionable. 17 These challenges also apply to census data. In Nigeria, there are inevitable tensions regarding population figures, given their importance in determining the allocation of petroleum revenues across regions and urban/rural zones. Political tensions are a major factor, along with technical challenges, in having outdated population weights, which affected the poverty figures. SUSTAINABILITY ISSUES The sustainability of poverty data efforts without the active presence of the Bank is questionable in most low-income countries and in some middle-income countries with low capacity, especially when success depends on key individuals in national statistical offices. When Guatemala’s MECOVI (Programa para el Mejoramiento de las Encuestas y la Medición de Condiciones de Vida) program ended in 2010 and external funding lapsed, many skilled consultants were let go. Technical problems associated with the 2011 ENCOVI (Encuesta Nacional de Condiciones de Vida) survey compromised the comparability of poverty statistics. In view of the government’s current fiscal constraints and the absence of external support for capacity building, stakeholders are now concerned about the sustainability of Guatemala’s statistical capacity. Similarly in Malawi, capacity remains thin because capacity-building efforts have not sustained domestic capacity, with considerable staff turnover and limited domestic skills. Donor financing is not a panacea for capacity constraints. In some countries, it has had unexpected negative consequences for clients, breeding a culture of donor dependency. Some developing countries have low use, demand, and investment in statistics, with donors (including the Bank) driving national surveys. The timing and 26 CHAPTER 2 PREPARING THE GROUND: WORLD BANK SUPPORT FOR POVERTY DATA focus of surveys in such dependent situations is determined more by donor needs than by country priorities. Much of the Bank’s survey work has been trust-funded, even for the flagship LSMS program, and donor interests have led the LSMS to focus on Africa and agriculture. Reliance on trust-funded surveys has limited the Bank’s ability to scale up the coverage and frequency of poverty monitoring. This will become increasingly important, given the data needs for the World Bank Group to monitor progress against the twin goals and to produce robust analytic work in the SCDs. DONOR COORDINATION Partnerships with bilateral and multilateral agencies have been integral to the Bank’s support for statistical capacity building since the early 2000s (box 2.4). A 2011 IEG evaluation of these global partnership programs found that they are generally working well (IEG 2011b). Based on interviews with both recipient countries and donor partners, the statistical capacity support generally meets user requests, is consistent with recipient country priorities, and is responsive to changing circumstances. The Bank has played a convening role in some countries. In Guatemala starting in the late 1990s, the government pushed hard to strengthen the public institutions in charge of carrying out living standards surveys and generating poverty-related data. Support for enhancing the National Statistics Institute’s capacity to map poverty and proxy means tests contributed to the government’s targeting system for its conditional cash transfer program, Mi Familia Progresa. 18 The support also contributed to greater data transparency and helped achieve public consensus on the measurement of poverty based on objective technical criteria. This facilitated the production of poverty statistics and diagnostics. 27 CHAPTER 2 PREPARING THE GROUND: WORLD BANK SUPPORT FOR POVERTY DATA Box 2-4. World Bank Participation in Partnerships for Building Statistical Capacity The most prominent statistics partnerships have been the Marrakech Action Plan for Statistics, Partnership in Statistics for Development in the 21st Century, Trust Fund for Statistical Capacity Building, Statistics for Results Facility, and International Household Survey Network (IHSN), as well as the Living Standards Measurement Study (LSMS) and LSMS-Integrated Surveys on Agriculture programs and the International Comparison Project. The major achievements of these partnerships include developing national strategies for statistics, easing access to data, coordinating definitions and protocols, and expanding census coverage to more countries. Collaboration with the United Nations and other partners has also grown. There has been somewhat less progress on implementing national strategies for statistics and ensuring the use of statistics for policy and planning. According to interviews with staff, the Bank’s leadership in open data since 2010 has been strengthened through partnerships within the Bank and with countries, other agencies, and the private sector. The partnership with Google has helped improve instant access to a number of datasets in the Bank’s open data website through Web searches. The Bank has many roles in partnerships to build statistical capacity—financial contributor, trustee, convener, chair of governing bodies, implementing agency, and a host of secretariats—and has made strong and relevant contributions over many years. In these partnerships, the Bank has drawn on such comparative advantages as strong technical skills, recognized technical leadership (including the long-running LSMS program), and connections to operations that sometimes financed statistics projects in client countries. Bank participation also enjoyed support from management and received funding from the Development Grant Facility, the corporate budget line for financing new partnerships. Partnership programs such as the IHSN can provide technical assistance to countries to improve their survey programs. Increasing the uptake and country ownership of statistics and surveys leads to the gradual transition out of donor dependency. Work with global and country partners can increase investments to improve the availability and quality of survey data (including income and non-income poverty data and related measures for shared prosperity) and nurture mechanisms such as administrative data and project-level data. One aim should be enhancing support for long-term data generation, particularly in low-income countries, by strengthening local capacity and demand for data and coordinating more with donor partners. Source: Challenges beyond Household Surveys More than just household survey data are critical to public planning and supporting allocative efficiency in budgets and service delivery. Data gaps extend to international cross-country comparison initiatives as well—and the challenges are greatest in countries where the information is needed most. In some instances, data from different sources produce different numbers—World Bank and International 28 CHAPTER 2 PREPARING THE GROUND: WORLD BANK SUPPORT FOR POVERTY DATA Monetary Fund data often differ, for example. Other data sources can also produce confusion for policymakers and development partners. 19 The International Comparison Program (ICP) produces the estimates of purchasing power parities (PPPs) used to monitor global poverty. Price levels for countries not participating in the ICP are estimated by regression models. In 2005, 146 countries participated (fewer participated in earlier years). 20 Country coverage in the 2011 ICP data collection rose to 199 countries and territories, so comparable price data are available for most countries. But changes between survey rounds and analytical methods create uncertainty about comparability with earlier estimates and thus about the time trends of extreme poverty. The lack of public access to the PPP data means that alternative methodologies cannot be tested and independent calculations of PPPs cannot be done. Timely and accurate population data are essential for establishing the sample frame for household survey data collection and for translating poverty rates into statistics on the number of poor. If the population census is obsolete, particularly in a country that is growing or otherwise demographically changing, it will produce an outdated sampling frame and biased results that can undermine public service delivery and planning systems. In Bangladesh, a country with a rapidly changing demographic structure, the poverty estimate is 10 million (or 7 percent) lower using the latest 2011 population census than when using the population projection with the 2001 census (World Bank 2015b). Similarly in the Philippines, analysis of changes in the agricultural sector have been particularly troubled because the Census of Agriculture is supposed to be conducted only once every 10 years, but the most recent was conducted in 2002. The obsolete census weakened the accuracy of the sample frame and biased the poverty estimates. Administrative data from the operational records of the education and health systems and other government programs are important for understanding the full picture of poverty. 21 But many countries lack complete civil registration systems, which are the principal source of vital statistics and intercensal demographic data. Civil registration also provides birth records and marriage certificates that are evidence of citizenship and legal rights. Although indicators from administrative records cannot be used to measure correlations across characteristics at the individual level, they provide valuable control totals for the aggregates derived from surveys. When available at regular intervals, they are particularly useful for measuring outcomes at higher frequencies than surveys can provide. The most problematic discrepancy affecting income poverty statistics is the large difference between the level and growth of per capita income and consumption in 29 CHAPTER 2 PREPARING THE GROUND: WORLD BANK SUPPORT FOR POVERTY DATA the national accounts and in income and expenditure surveys. Some differences are definitional: consumption and income aggregates from the national accounts include concepts that are not in household surveys, such as the value of bank intermediation services. Others may arise from the difficulty in sampling very rich and very poor people. Typically growth rates of consumption in surveys are one-third to one-half those in the national accounts, which is troublesome because national account growth rates are used to extrapolate surveys forward (World Bank 2015b). The Bank could play a stronger role in improving the quality and consistency of data for all dimensions of poverty through partnering with multiple agencies. 1 Considering the many challenges in poor countries, including conflict, violence, and all sorts of deprivation, weak data can hinder the understanding of the poverty situation, even if it might not be the immediate binding constraint to an effective poverty reduction strategy. 2According to figures from PovcalNet, the number of surveys has declined since 2010. Rather than implying a decline in surveys collected, this may be related to World Bank data processing and reflective of the time it takes to clean data and make them available in the PovcalNet platform. PovcalNet (database), World Bank, Washington, DC (accessed October 2014), http://iresearch.worldbank.org/PovcalNet/index.htm. 3In this report, unless otherwise specified, data availability refers to data available in the Bank’s poverty data platform, PovcalNet. The increase in the number of surveys in PovcalNet may be due either to more surveys being conducted or to improvements in PovcalNet’s ability to collect information from countries. As of November 2014, the database included 854 household surveys from 117 low- and middle-income countries, compared to 82 surveys from 79 countries at the end of the 1980s and 452 surveys from 123 countries at the end of the 1990s. 4Many countries also had nationally, regionally, and zonally (rural/urban) representative household surveys, with subnational information available at the state or district levels. 5Dating back to 1980, PovcalNet has data for 150 countries, of which 33 are now classified as high-income economies, including some former “developing” or “transition” economies such as Chile, Uruguay, Russia, and other eastern European countries, leaving 117 low- or middle-income economies. The World Development Indicators lists 215 countries and territories, of which 139 were classified as low- or middle-income in 2014. 6The international standard for the classification of consumption is the Classification of Individual Consumption According to Purpose (COICOP). See United Nations Statistics Division, “Detailed structure and explanatory notes,” http://unstats.un.org/unsd/cr/ registry/regcst.asp?Cl=5 . 7In the past 20 years, Lao PDR has improved the implementation and analysis of its household income and expenditure surveys, broadened the range of topics covered by the surveys, and introduced additional methods to better capture cost differences between urban and rural areas. 30 CHAPTER 2 PREPARING THE GROUND: WORLD BANK SUPPORT FOR POVERTY DATA 8 In Senegal, the household surveys were conducted at different times of the year and rural sector well-being changes significantly after the annual harvest (October–December). The first round of the survey (ESPS I) is particularly problematic as it was conducted following the 2005 harvest, which likely resulted in an underestimation of the level of poverty. When compared with information from the subsequent survey rounds, the data yields inaccurate poverty trends. 9Countries have undertaken a variety of initiatives creating open access to government information. See, for example, the recent Open Data Foundation review of 13 case studies (Davies 2014). International agencies have also adopted policies of free distribution of their databases. The IMF, for example, has announced that as of January 1, 2015, its statistical databases will be made available for free. 10 See the Bangladesh Bureau of Statistics website, http://www.bbs.gov.bd. 11IHSN membership includes more than 20 organizations with a management group consisting of representatives from World Bank, UNICEF, ILO, U.K. Department for International Development, UNSD, PARIS21, UIS, FAO, and WHO. 12For more information on the Statistical Capacity Indicator, visit the Bulletin Board on Statistical Capacity, http://data.worldbank.org/data-catalog/bulletin-board-on-statistical- capacity. 13SCI overall scores were adjusted to remove two indicators related to poverty surveys. The adjusted SCI overall scores were re-normed to a range of 0 to 100. The two items referred specifically to production of poverty statistics―the availability of a recent HIES and the timeliness of poverty estimates measured at the international poverty line. 14 See, for example, McKinsey Global Institute 2011. 15“Staff working on Blend countries were more likely (50 percent) to identify insufficient political will within government as a main constraint as compared to staff working on IDA (40 percent) and IBRD countries (34 percent).” See appendix C, paragraph 13. 16Matching the labor force surveys or Egypt Labor Market Panel Surveys with the HIES could have provided additional information on poverty. 17There is limited information on whether the Egypt HIES captures the popoluation living in informal areas. According to some estimates, the population living in the ancient cemetery area near Cairo is on the order of 500,000 (see http://www.huffingtonpost.com/2014/10/29/city-of-the-dead-cairo_n_6044616.htm). On people living in informal areas, see, for example, GTZ (2009), Cairo’s Informal Areas Between Urban Challenges and Hidden Potentials: Facts. Voices. Visions. 18The program is sponsored by the World Bank, the Inter-American Development Bank (IDB), and the United Nations Economic Commission for Latin America and the Caribbean (CEPAL), along with support from a number of other institutions and bilateral donors (including UNDP, USAID, UNICEF, ILO, the Soros Foundation, Canada, Denmark, Germany, Japan, Norway, and Sweden). Gini coefficients in several countries that use the Bank’s PovcalNet and the Standardized 19 World Income Inequality Database (SWIID) do not match. The difference in Gini coefficients 31 CHAPTER 2 PREPARING THE GROUND: WORLD BANK SUPPORT FOR POVERTY DATA in the same year can be as high as 10 percentage points for some countries (such as Kenya and Zambia) and trends of change can differ. While both numbers can be justified based on the technical assumptions made, the gaps are nevertheless confusing for the broad development community and policymakers attempting to understand the levels and trends of changes in income/consumption distributions. See Ferreira and Lustig (forthcoming). 20Price levels for 21 out of 36 fragile states were based on a regression model (World Bank 2015a). 21Statistics based on administrative data often differ from those derived from surveys for many reasons. The enrollment statistics in educational management information systems are usually based on beginning-of-the-year enrollments. Survey data, such as those collected by the DHS and MICS, record school attendance at an arbitrary mid-year date. Both may be correct, they simply are measuring different events. Records of the health system record the incidence of disease but only for those seeking care, leaving out the unserved. 32 3. Laying the Foundation: World Bank Support for Poverty Diagnostics Highlights  The Bank is considered a leader and valued partner in poverty diagnostics by governments and other stakeholders, providing an important global public good.  The Bank’s Poverty Assessments (PAs) are generally of high technical quality and make good use of available data to derive poverty indicators, poverty profiles, and identify poverty drivers, but at times they inadequately synthesize knowledge on poverty in the country.  The rigorousness of poverty diagnostics is often constrained by the quality and accessibility of data.  Poverty Assessments could have had more impact if the diagnostics were better timed with political cycles and disseminated, and if their policy recommendations were more actionable or specific. Poverty data provide the basis for identifying the poor and measuring poverty. But understanding the constraints to reducing poverty requires solid analysis. Good poverty diagnostics can not only pinpoint the symptoms of poverty but also identify its causes and provide policy options. This chapter evaluates the technical quality of the Bank’s poverty diagnostics, examines whether the diagnostics have provided the needed guidance to country programs on poverty reduction, and identifies the constraints. The evidence for this chapter is drawn primarily from a review of the Bank’s Poverty Assessments (PAs) in 20 countries, supplemented by relevant findings from country case studies, staff and stakeholder surveys, and focus group meetings. Technical Quality of Poverty Assessments The Bank’s PAs are generally of high technical quality. Governments and other stakeholders largely consider the Bank a global leader and valued partner in the development of poverty diagnostics, acknowledging the Bank’s comparative advantage in providing global knowledge products. 1 Three-quarters of government and nongovernment stakeholders noted that PAs and Poverty and Social Impact Analyses (PSIAs) provide well-grounded analysis and identify key constraints to poverty reduction. 2 Bank staff expressed similar opinions about the quality of the Bank’s poverty diagnostics. The views of both groups were corroborated by an in- depth review of PAs in 20 countries and by the 10 country case studies. 33 CHAPTER 3 LAYING THE FOUNDATION: WORLD BANK SUPPORT FOR POVERTY DIAGNOSTICS The criteria to evaluate the quality of the 20 PAs are consistent with the World Bank’s 2004 Guidance Note on Poverty Assessments by examining to what extent the assessments: • Provide background information on available poverty surveys and data • Make good use of available survey and other data to provide a clear understanding of the extent and drivers of poverty • Assess the adequacy of the countries’ poverty-reduction institutions, programs, and funding • Evaluate poverty monitoring and evaluation arrangements • Propose specific and actionable recommendations for reducing poverty • Influence the countries’ poverty-reduction strategies and programs, help build in-country capacity, and support joint work and partnerships. Although the 20 PAs varied considerably in their depth of analysis and coverage of topics, the review consistently found three positive attributes. 3 Most of the PAs reviewed make good use of the information available, produce clear poverty profiles, and identify the main constraints to poverty reduction. Nearly all made use of the available data and explored the obstacles to poverty reduction nationwide, in the regions most affected, and across social groups. The review identified the Ethiopia and Indonesia PAs as examples of good practice with strong poverty diagnostics (box 3.1). Box 3-1. Good Practices for Poverty Assessments Well-Being and Poverty in Ethiopia: The Role of Agriculture and Agency (2005) and Making the New Indonesia Work for the Poor (2006) were particularly well suited to inform poverty- reduction policies and programs. Both assessments:  Explicitly drew upon many data sources, provided full descriptions of the available information base, and concisely integrated other existing knowledge  Made good use of all available surveys and other information through clear presentation of the extent and drivers of poverty, identification of excluded groups, detailed analysis of empowerment and governance, compelling assessments of remoteness and gender inequalities, and presentational features that showed how the poor experience poverty  Integrated information about the governments’ key poverty-reduction institutions, strategies, funding, and programs (though neither assessment provided a discrete summary of those aspects)  Included annexes with specific recommendations for capacity building and improving poverty-reduction monitoring and evaluation 34 CHAPTER 3 LAYING THE FOUNDATION: WORLD BANK SUPPORT FOR POVERTY DIAGNOSTICS  Provided a clear statement of objectives, a succinct “Striking Statistics” section, and prioritized, rigorous, and highly specific policy recommendations across multiple sectors  Made strong efforts to engage government, development partners, academics, and civil society during Poverty Assessment preparation. Source: Review of Poverty Assessments for 20 countries conducted for this evaluation. Characteristics of the Poor and Drivers of Poverty Reduction IDENTIFYING THE POOR The Bank’s diagnostic work typically describes a country’s poverty profile. When survey data are of good quality and representative at disaggregated levels, diagnostics usually estimate the national poverty headcount and, at least to some extent, explore differences across geographic areas or zones (such as urban/rural, valleys/hilltops). These diagnostics provide a picture of the characteristics of the poor. Good-practice PAs (for example, Brazil, Egypt, India, Papua New Guinea, and the Republic of Yemen) went further by undertaking or providing technical assistance for detailed poverty mapping at the local level. Some diagnostics even reported on vulnerability, seasonal poverty, transient and chronic poverty, and inequality. The level of detail and focus on heterogeneity varies by country, but is vital to targeted policy formulation to reduce poverty. Box 3-2. The Face of Poverty and Disadvantaged Groups Aggregate assessments of poverty can hide patterns of poverty incidence within a country. The social, economic, or political exclusion marginalized groups face can often intensify the severity of their deprivation. Good-practice PAs identify how multifaceted deprivation is related to the depth of poverty experienced by a certain group. For example: The 2010 Afghanistan PA found that the Kuchi (Pashtun nomads) “not only suffer from a higher prevalence of poverty, but the Kuchi poor are on average poorer compared to other groups” (World Bank, 2010b, 26) The 2011 India PA found that “scheduled tribes lag 20 years behind the general population” and “caste has been the predominant marker of deprivation and privilege in India” (World Bank 2011d, 227). The 2006 Lao PDR PA found that “ethnic minorities account for one-third of the population but make up more than half the poor” (World Bank 2006b, 129). The 2004 Moldova PA found that “many groups additionally face impediments imposed by social barriers and norms―their multiple levels of deprivation are compounded by social exclusion and discrimination within society” (World Bank 2004c, 11). 35 CHAPTER 3 LAYING THE FOUNDATION: WORLD BANK SUPPORT FOR POVERTY DIAGNOSTICS Several PAs (including Afghanistan, India, Lao PDR, Moldova, and the Republic of Yemen) went beyond the typical analysis of poverty’s determinants. They examined more deeply the well-being of marginalized ethnic minorities, indigenous peoples, tribal groups, castes, and other excluded groups, using both income and non-income indicators and reviews of sociological and anthropological literature, as described in box 3.2. Such analytic work is essential for understanding what challenges the poor and vulnerable face and how policies may affect marginalized populations. DRIVERS OF POVERTY REDUCTION The main drivers of poverty reduction identified and discussed in the majority of country cases include growth, distribution, education, health, social protection, and employment. Most poverty diagnostics examined the key drivers of income and non-income poverty at the national, regional, and zonal (rural/urban) levels, and across social groups. They considered determinants of the changes in poverty incidence over time (such as growth and distributional changes) and explored the obstacles to poverty reduction nationwide, in the regions and areas most affected, and across social groups. Intense research during the past decade highlighted the complex nature of the forces behind poverty reduction (box 3.3). Although there is now greater agreement on the types of deprivation that are bad for both poverty reduction and growth (for example, child malnutrition), there remain unresolved analytical and empirical issues on the exact nature of “inclusive growth,” and on the nature and magnitude of various trade-offs between growth and distribution. Many country case studies provide good examples of how growth and inequality play a role in poverty reduction. Bangladesh showed negative or no effect of growth on poverty reduction because of inequality in the 1990s and early 2000s, and then saw redistribution facilitate poverty reduction in later years. Romania demonstrated both the negative effects of a decline in growth on poverty in the 1990s and early 2000s (though it was mitigated by improved income distribution) and the positive effects of improved growth, leading to a decline in poverty (with a small contribution from lower inequality). Guatemala experienced modest declines in poverty accompanied by growth between 2000 and 2006, with modest reductions in inequality, but nearly no change in extreme poverty levels. From 2006 to 2011, Guatemala had the same growth rate as the preceding period, but an increase in poverty incidence and a significant decrease in extreme poverty. 4 The Philippines diagnostics discuss the growth paradox in which poverty rose despite growth, concluding that it may partially be due to “limited dynamism” of growth coupled with high degrees of inequality. 36 CHAPTER 3 LAYING THE FOUNDATION: WORLD BANK SUPPORT FOR POVERTY DIAGNOSTICS Box 3-3. Literature on the Drivers of Poverty Reduction A large amount of research has examined the growth-inequality-poverty nexus and provided theoretical and empirical evidence for the broad drivers of, and impediments to, poverty reduction. One strand of the literature has tried to answer the question of how much the poor benefit from aggregate economic growth. Seminal cross-country studies found that growth in average income is highly correlated with poverty reduction (Ravallion and Chen 1997; Ravallion 2001). This strand of the literature argues that economic growth is of central importance for poverty reduction, and growth is a primary determinant of the variation in the decline (Kraay 2006; World Bank 2005b; Gasparini and others 2007; Dollar and others 2014). Another strand of the literature emphasizes the role of distribution and the interdependence among growth, inequality, and poverty reduction. It argues that the effect of growth on poverty reduction is greater in low-inequality countries because the growth elasticity of poverty reduction in low-inequality countries is several times larger than that observed in high-inequality countries (Ravallion 1997, 2007; Lustig and others 2002; Bourguignon 2004; World Bank 2005b; Fosu 2010). The empirical finding shows that inequality not only has a negative impact on economic growth (Herzer and Vollmer 2012; Benjamin and others 2010; Knowles 2005; Voitchovsky 2005), but also on its sustainability over time (Berg and others 2012). Initial inequality levels, in particular, initial differences in human capital and social exclusion, could play an important role in determining how growth can influence poverty reduction (Ravallion 2001). Yet another strand of the literature highlights the role of growth composition in poverty reduction. It argues that growth in service sectors, which often absorb low-skilled workers, shows more poverty reduction power than that in agriculture or industry. Initial urbanization could enhance access to markets and infrastructure, thus positively influencing the poverty impact of nonagricultural growth (Ravallion and Datt 1996; Ferreira and others 2010). Some empirical studies show that growth in labor-intensive sectors contributes the most to poverty decline (Loayza and Raddatz 2006; Christiaensen and Demery 2007), and some recent decomposition exercises highlight the importance of labor income as the main factor behind poverty decline (Inchauste and others 2012). At the same time, other studies explored the heterogeneity of initial conditions in human capital accumulation and the role of growth in nonlabor-intensive sectors (Ravallion and Datt 2002). Among the 20 PAs, some themes and drivers were more important in some Regions than others. The three Europe and Central Asia countries highlighted the roles of national and international migration, employment, labor markets, and social protection systems and programs. For the three Middle East and North Africa countries, social protection systems and programs were most important. 37 CHAPTER 3 LAYING THE FOUNDATION: WORLD BANK SUPPORT FOR POVERTY DIAGNOSTICS Growth is most often identified in poverty diagnostics as the key driver of poverty reduction (in 7 of the 10 country cases). But the type of growth (or its inclusiveness) is also crucial. The 2008 Bangladesh PA characterized poverty reduction as driven by a social and economic transformation that was increasing returns to the assets of the poor, primarily in higher wages, especially for nonfarm employment. Increasing labor force participation of women and overall increases in education also contributed to reducing poverty. Education and health were repeatedly identified and analyzed as key drivers of poverty reduction in both the country cases and the PAs. Almost all PAs explored the impact of access to education and health services on poverty reduction, and additional sectoral work such as Public Expenditure Reviews (PERs) delved further into the links between education, health, and poverty reduction. Linking growth to poverty reduction requires jobs, particularly productive jobs, 5 and the pattern of growth and job creation affects the responsiveness of poverty reduction to that growth. The quality and depth of labor market analysis and recommendations for generating more jobs has varied. In the Philippines, several pieces of recent diagnostic work—particularly the 2013 Philippine Development Report 6—explored the relationship between the pattern of growth and job creation. Weak growth in productive jobs domestically was identified as a constraint to inclusive growth and poverty reduction, and the linkages between migration, remittances, and poverty incidence were highlighted. In Nigeria, the 2009 Nigeria Employment and Growth Study was critical to the Bank’s dialogue on non-oil growth. It asserted that jobs and poverty reduction were synonymous (when combined with agricultural growth) and flagged employment as a major driver for poverty reduction. But it did not make explicit links to household income poverty analysis, unbundle recommendations, or distinguish among the nonpoor, the moderately poor, and the poorest. The weak link between real wage growth and poverty reduction was not explained. A good practice for PAs is to identify the relationship between growth and poverty reduction, calculate growth incidence curves, and project poverty outcomes using the poverty-growth correlation (Bourguignon 2004). In Bangladesh, growth incidence curves differentiated the patterns of poverty reduction and inequality between the two halves of the 2000–2010 decade. In Malawi, drawing from panel survey data, the PA showed crucial differences in changes in income among different segments of the rural population: real incomes of the rural poor are falling, and only the better-off households’ experience growth in real expenditures. But in less convincing cases, such as Nigeria, PAs linked macroeconomics, growth, and poverty reduction only at a high level of generality, without clearly identifying channels for “trickling down.” 38 CHAPTER 3 LAYING THE FOUNDATION: WORLD BANK SUPPORT FOR POVERTY DIAGNOSTICS Existing diagnostic work makes a convincing case that improvements in social development contribute to poverty reduction. Social development is generally understood to encompass equity and social justice, including social inclusion, sustainable livelihoods, gender equity, and increased voice and participation. Social exclusion hampers the ability of people from disadvantaged groups to participate in social and economic life. It results from and represents structural inequities that lie at the heart of much inequality. In many developing countries, although the better- off enjoy a living standard similar to those in the developed world, the poorer segment of the population faces severe challenges in many non-income dimensions, including the persistence of child malnutrition, high maternal mortality, and low (particularly girl) school enrollment. The Bank’s diagnostic work on poverty at times offers useful information to guide targeting interventions to address social development and exclusion. In Romania, since re-engaging in the early 1990s, the Bank identified gaps in coverage, targeting, and integration of social protection arrangements and provided actionable recommendations to address issues in health service delivery. The primary aim was to restructure services toward better hospitals—with more care provided by ambulatory and primary care services—and to seek savings in such areas as expensive, unnecessary medications. In Guatemala, the poverty diagnostics focused on more direct interventions to reach the poor, such as conditional cash transfers (CCTs), and on stronger efforts to equalize access to productive infrastructure and social services. Extreme poverty seemed to respond well to CCTs introduced in the latter 2000s, providing useful analytical underpinning for the interventions. Social delivery systems determine the access to and quality of basic services for the poor, often relying on subnational governments. Local institutional capacity, particularly in the poorest municipalities, determines the effectiveness of additional transfers from the center to the front lines of service delivery. This is the case in Peru after a strong decentralization reform. In Nigeria, federalism and decentralization— overlaid with sizable resources from oil revenue—complicated the formulation and implementation of the Bank’s program. A lesson learned from the Nigeria case is that greater emphasis on governance, stronger subnational engagement, a sharper focus on results, and appropriate choices in the design and selection of lending instruments can increase the impact of pro-poor social service delivery. Timeliness and Dissemination The time between data collection and assessment completion has improved, but many assessments still work with outdated information. In the staff survey, 37 percent of respondents identified the delay between the release of analytical work and the 39 CHAPTER 3 LAYING THE FOUNDATION: WORLD BANK SUPPORT FOR POVERTY DIAGNOSTICS drafting of country strategies as one of the most important constraints to developing country strategies. 7 There were no significant variations across respondents mapped to the Bank’s Regions. 8 Shorter and more regular poverty notes and updates were seen as having greater potential for being used (both by the clients and the country teams) than longer but less timely PAs. For example, a large PA right before the Country Partnership Strategy process starts may have less impact than regular poverty updates submitted to the country team. Thirteen of the 20 PAs were completed within three years of a survey to estimate the poverty headcount (table 3.1). Table 3.1. Years between Survey and Poverty Assessment Completion Number Lag of (years) Countries Countries 1 1 Yemen, Rep. Afghanistan; Congo, Dem. Rep.; Egypt, Arab Rep.; Guyana; Indonesia; Kyrgyz 2 7 Republic; Moldova 3 5 Armenia, Bangladesh, Columbia, Lao PDR, Nigeria 4 2 Brazil, Iraq 5 1 Mozambique 6 3 China, Ethiopia, India 8 1 Papua New Guinea Source: Systematic Review of Poverty Assessments for 20 countries. Understanding the impact of public expenditure and revenue policies on the poor is central to informing policymaking. But fewer than half of the country strategies in FY2004–2013 were preceded by PERs in the previous three years (table 3.2). Two- thirds of PERs discussed reorienting public spending to benefit the poor, and two- fifths conducted (or referred to) incidence analysis and looked into the distributional impacts of public policies. Of the 146 PSIAs since FY2004, roughly 40 percent were explicitly referred to when designing budget lending policy operations. Table 3.2. Fewer than Half the CAS/CPSs were Preceded by PERs within Three Years Two Years or Three Years or Region Same Year One Year or Less Less Less AFR 6 (9%) 14 (22%) 27 (42%) 34 (53%) ECA 4 (8%) 15 (29%) 25 (48%) 32 (62%) LCR 5 (13%) 9 (23%) 11 (28%) 13 (33%) EAP 2 (9%) 3 (14%) 6 (27%) 7 (32%) MNA 2 (13%) 2 (13%) 2 (13%) 3 (20%) SAR 0 (0%) 1 (8%) 1 (8%) 2 (17%) 40 CHAPTER 3 LAYING THE FOUNDATION: WORLD BANK SUPPORT FOR POVERTY DIAGNOSTICS Total 19 (9%) 44 (21%) 72 (35%) 91 (44%) Source: Staff review of the CAS and PER during FY2004–2013. Note: AFR = Africa Region; EAP = East Asia and Pacific Region; ECA = Europe and Central Asia Region; LCR = Latin America and the Caribbean Region; MNA = Middle East and North Africa Region; SAR = South Asia Region. Reaching a wide array of stakeholders can improve the impact of poverty diagnostics on strategy formulation. The PAs would have benefitted from better communication and wider dissemination. The July 2004 Guidance Note on Poverty Assessments called for wide dissemination of “the results of poverty work within the Bank and outside” to “ensure strong linkages between PAs, the development of CASs, and the design of lending operations and nonlending activities” (World Bank 2004a). There are several examples of successful partnerships with the government that strengthened the dissemination of poverty diagnostics. The report Poverty in Lao PDR 2008: Lao Expenditure and Consumption Survey 1992/03–2007/08 benefitted from extensive consultation and dissemination, such as a launch workshop, dissemination to universities and provinces, and online access, including statistics and information in Laotian. In the Philippines, the key findings of Bank analytical work are frequently disseminated to both stakeholders and the public through events in the World Bank Manila Office and in World Bank Knowledge for Development Centers throughout the country. 9 In several instances, it appears the Bank did not adequately focus on communicating with stakeholders, which limited the impact and effectiveness of the diagnostics. In Nigeria, Bank outreach was quite limited. No one interviewed outside the Bank was aware of any of the Bank’s three PAs produced since 2004. 10 Lack of time and budget, along with lack of support from the professional communications team to create targeted messages were cited by focus group participants as main constraints to effective dissemination of findings and collaboration with government counterparts. The focus group participants also noted that more institutional recognition is often given to large poverty assessments, and the Bank does too little to disseminate the findings of its smaller poverty related work, either internally or to its country clients. Constraints to Poverty Diagnostics Data remain a severe constraint to the depth and rigor of poverty diagnostics in some countries. For example, of the 20 PAs reviewed, 10 had at most one round of household surveys since 2000 and could not establish poverty trends. 11 The depth, rigor, and usefulness of diagnostics vary considerably across countries, related to a 41 CHAPTER 3 LAYING THE FOUNDATION: WORLD BANK SUPPORT FOR POVERTY DIAGNOSTICS significant extent to the variation in the availability, quality, and accessibility of data. Although some PA task teams tried to use other sources of information and draw on their country knowledge, the reliability of results was tenuous. One-third of the respondents to the Bank staff survey identified insufficient poverty data as the main constraint to carrying out PAs and PSIAs. Not surprisingly, concerns about data were stronger among staff working on fragile and conflict-affected states and countries with limited data. In several instances, weak data limited the ability to draw meaningful conclusions and make credible recommendations (box 3.4). Box 3-4. Weak Data Limit the Scope and Robustness of Poverty Diagnostics Nigeria accounts for about 7 percent of the world’s extreme poor, yet high-quality data on the poor are not consistently available. When Nigeria: Poverty and Vulnerability: A Preliminary Diagnostic (2004) was being prepared (it was never finalized), the most recent data were from a 1996 household consumption survey and a 2003 Core Welfare Indicator Questionnaire. The assessment provided only limited insights into Nigeria’s poverty profile and its drivers of poverty. Weak data also limited the scope of the most recent 2013 PA. Methodological issues impaired the comparability of survey data between the 2004 and 2010 surveys, which limited the Bank’s ability to take more effective measures to reduce poverty. Limited access to the full data sets made it difficult to assess the quality of Egypt’s poverty data, at least until the 2010 revolution. Compounded by the challenges in using the administrative price and the concentration of households at relatively low levels of consumption, poverty estimates are highly sensitive to the choice of poverty lines. Using estimates from the Central Agency for Public Mobilization and Statistics, some poverty diagnostics may have missed the increasing vulnerability of migrant workers in the informal sector in urban areas, even though the diagnostics correctly identified poor households in Upper Egypt. This may have failed to provide a robust signal to policymaking. Source: Country studies of Nigeria and Egypt; World Bank (2014d). PAs could have had more impact if the analysis of institutional aspects of poverty had been stronger; if the policy advice had been more specific and actionable; and if the analysis had gone beyond economic considerations to fully take into account the social—and particularly the political—framework for removing obstacles to poverty reduction. The majority of PAs reviewed paid inadequate attention to the political context of the economy, such as government institutions, strategies, and funding for poverty reduction. The inadequate coverage of institutional dimensions and the deeply rooted political economy likely limited the impact of the diagnostics. None of the 20 PAs provided a comprehensive discussion of the key actors and funding resources for poverty reduction, and none positioned these actors and resources in the context of government poverty-reduction strategies and programs. 42 CHAPTER 3 LAYING THE FOUNDATION: WORLD BANK SUPPORT FOR POVERTY DIAGNOSTICS For Bangladesh, the three PAs produced during 2004–2012 treated public expenditure and administration primarily as a technocratic issue and did not broaden the analysis to consider the underlying political economy. Much of the institutional, policy, and program context for poverty reduction was instead contained in poverty-related sector work (especially in education, health, and social protection), although the PAs concentrated on the drivers of poverty. For Senegal, the analysis did not explicitly discuss the social and political constraints to implementing policy. Synthesizing knowledge inside or outside the Bank is often beyond the coverage of a PA. The 2004 Guidance Note on Poverty Assessments requires “an analytical synthesis of the existing body of knowledge on (i) assessments of the poverty situation, (ii) analyses of the impact of growth and public actions on poverty, and (iii) appraisals of poverty monitoring and evaluation systems. Despite these requirements, none of the PAs included an analytical synthesis of knowledge (that is, in-country knowledge of poverty, relevant sectoral issues, or on donor programs). Only five of the 20 PAs reviewed, including the two good-practice PAs of Ethiopia and Indonesia, summarized the governments’ overall poverty-reduction strategies or explained how the PAs would contribute to the development of those strategies. This deficiency left Bank staff in a weak position to mobilize and use knowledge on poverty more comprehensively, and to know which government actors to engage in dialogue. Better engagement of beneficiaries and using existing (or commissioning new) participatory or qualitative analyses would enrich poverty profiles and inform poverty diagnostics. Of the 20 PAs, at least 15 referred to participatory analysis and qualitative information. But only five were well informed by and directly included participatory analysis and qualitative information (Democratic Republic of Congo, Indonesia, the Kyrgyz Republic, Mozambique, and Papua New Guinea). These five set standards for good practice. For example, in the Democratic Republic of Congo’s PA, the description of who the poor were, how they experienced poverty, and what their priorities were for overcoming poverty was heavily influenced by participatory analysis and a 2004 opinion survey. Good Practices and Lessons Robust poverty diagnostics in some countries have identified the challenges faced by the poor and the corresponding policy interventions. In Malawi, the Bank’s diagnostic work focused on ways to reverse stunting 12 and helped to inform the Nutrition and HIV/AIDs project that has worked to promote sustained improvements in child nutrition. But when robust and timely diagnostics are absent, the recommended policy interventions suffer. In Liberia, the Bank’s estimate of 43 CHAPTER 3 LAYING THE FOUNDATION: WORLD BANK SUPPORT FOR POVERTY DIAGNOSTICS timber output and revenue collection were too optimistic. The unbalanced support favored commercial goals at the expense of community forest management and conservation.13 The resumption of large-scale commercial logging (which did little to enhance local livelihoods) did not yield the expected benefits in terms of growth or poverty reduction. Well-executed Bank diagnostics make a real policy difference by improving information transparency and putting poverty into the national discourse with a technical perspective. In Guatemala, poverty maps and means testing became part of the government’s public programs and strategic planning. 14 In the Philippines, the World Bank’s poverty diagnostics helped set the agenda and policy discourse (IEG 2007). The Lao PDR PA identified the concentration of extreme poverty in the Priority Districts and helped target Bank support for basic infrastructure, education, health, and other social services in these areas, which disproportionately benefitted the extreme poor, ethnic minorities, and other disadvantaged groups. In some PAs, a focus on the non-income dimensions of poverty helped address key constraints to poverty reduction. In Romania, the PA supported the production of poverty and exclusion maps, using data from the census and household surveys such as information on deprivation (nutrition, durables, housing), education, health, and employment, and (from a specially designed survey) on social capital of the different segments of the population. In Cambodia, a policy and program impact assessment underpinned the Education Sector Support Project (Filmer and Schady 2006). In Georgia, the Bank sponsored work to improve the efficacy and efficiency of social protection programs and the targeting of poverty alleviation funds. Special attention to the demand side of service delivery in some countries tailored interventions to meet the needs of the poor. In Peru, pockets of extreme poverty are concentrated in rural indigenous groups, so their demand for and use of social services depends on how the services are aligned to their cultural practices. The 2005 Peru Poverty Assessment used an integrated (general equilibrium) approach to examine not only the supply side of growth and sustainability issues, but also the demand side, including social services used by the poor. The programmatic analytical and advisory assistance program RECURSO (derived from the Spanish acronym for Rendimiento de Cuentas para los Resultados Sociales, meaning Accountability for Social Responsibility) identified the needed actions to improve incentives on the supply side to provide adaptive services to the poorest segments of the population (for example, ethnic minorities and the disadvantaged). Poverty diagnostics can and have been tailored to country specifics and provided concrete recommendations. In Lao PDR, the PAs provided a good understanding of 44 CHAPTER 3 LAYING THE FOUNDATION: WORLD BANK SUPPORT FOR POVERTY DIAGNOSTICS extreme poverty and of the special concerns of poor women and upland ethnic minority groups. They also set priorities for poverty reduction measures, and provided a credible framework and menu of options that contributed to effective dialogue with the government, international development partners, and other parties. In Malawi, the Bank’s diagnostic work provided concrete recommendations to address obstacles to reducing poverty, considering the Malawian context and drawing on a broader body of analysis beyond that of the Bank.15 In Romania, a joint Bank-UNICEF report conducted a rapid assessment of the impact of the 2009 crisis. The Bank also used 2011 census data to update Romania’s poverty map, and supported development of poverty and inclusion indicators at the subnational level by including data on marginalized communities in a recently published report. These efforts responded to an increased focus on inclusion. Few diagnostic works have realized their full potential for deepening the analysis of the social and political dimensions of poverty—they miss an opportunity to deepen analysis of social inclusion and strengthen policy impact to advance the Bank’s overarching goal for greater inclusion. Social exclusion hinders the ability of people from disadvantaged groups to participate in social and economic life. It is often the reason why gains in health, education, employment, and prosperity systematically bypass people from disadvantaged groups. No single institutional arrangement for ensuring such inclusion will be optimal for all societies. How a society provides opportunities for inclusion is context- and time-dependent, linked to the political economy and power-sharing arrangements and whether the poor have a voice in determining national economic policies. 1The Bank produces many poverty-related diagnostics, including Poverty and Social Impact Analyses, Public Expenditure Reviews, and Country Economic Memoranda. But the evidence for this chapter is drawn primarily from a review of the Bank’s Poverty Assessments in 20 countries, supplemented by findings from the case studies, staff and stakeholder surveys, and focus group meetings. 2 The overall positive assessment of the quality of the PAs is broadly consistent with the key findings of the previous analysis presented in OED’s 2004 Annual Review of Development Effectiveness: The World Bank’s Contributions to Poverty Reduction and the 2003 Quality Assuarance Group Assessment Quality of ESW in FY02. 3The 20 PA countries were purposively selected to (a) provide equal coverage of each of the Bank’s six regions (four countries each for Africa and East Asia and the Pacific, and three countries each for the other four regions), (b) include countries with greater rates of poverty either as a proportion of the developing world’s poor, or as a share of country population, and (c) cover at least one weak-data country in each region. 45 CHAPTER 3 LAYING THE FOUNDATION: WORLD BANK SUPPORT FOR POVERTY DIAGNOSTICS 4The increase in poverty is thought to be linked to a global decline in remittance flows. The decline in extreme poverty is thought to be linked to public sector efforts to expand and better target social safety nets, and to ramp up social spending. The decline is also linked to a significant decline in inequality between 2006 and 2011, most notably in rural areas where most of the extreme poor are concentrated. 5A job not only produces income to support consumption and to provide resources for the future (such as providing education, health care, and assets for family members), it also contributes to self-esteem, a sense of personal security, and even social cohesion. 6Creating More and Better Jobs (2013); 2010 Philippines Development Report: Generating Inclusive Growth (2010); 2011 Philippines Development Report: Generating Inclusive Growth to Uplift the Poor (2011). See also, World Development Report 2013: Jobs (2012). 7The release of the analytical work may have been timed to inform government decision making and not the CAS timing. 8But 46 percent of staff working on FCS countries believed these delays to be a constraint, compared with 36 percent for non-FCS countries. 9There are 12 centers throughout the country in seven different universities as well as an inside of an array of research centers, such as the Congressional Policy and Budget Research Department of the House of Representatives. 10Part of the reason for the weak dissemination was the political environment, particularly given the sensitivity of the poverty numbers in the country. While outside the scope of the IEG study period, the IEG mission observed the importance the Bank team is giving to the communication of the next 2014 Nigeria Economic Report whose special topic is Poverty. 11The selection of the PAs in countries with weak data in the review is to reflect the challenges of conducting PAs with limited information. This does not reflect the share of PAs with weak data to the total of the PAs. 12In Malawi, the Bank partnered with USAID to produce a report on barriers and facilitators to infant and young child feeding. The report drew on multiple sources of information, including an anthropological study on feeding practices, covering three regions in Malawi. See: IYCN Project 2011. 13The advice from the World Bank Group led the government to believe that forest products would yield $108 million in revenues for the period 2007–11 on a timber volume of 3.3 million cubic meters. In reality, only 5 percent of forest concessions reached the production stage, while revenue collection was roughly $10 million—less than one-tenth of projections. (cited from IEG 2012b). 14These instruments either did not exist in Guatemala, or existed only in very rudimentary form, until the 2003 PA was prepared. Before then, this topic was considered too sensitive to discuss in view of the country’s historical and political circumstances. 15In Malawi, the diagnostics had breadth of coverage with a strong team leader facilitating collaboration among a cross-sectoral team, although inevitably some parts of the country team engaged more than others. 46 4. Framing the Structure: Formulating Country Strategies Highlights  The Bank’s country strategies have been broadly consistent with its poverty diagnostics and oriented toward poverty reduction.  Poverty diagnostics often provided strong technical analysis, but the lack of actionable guidance limited their direct relevance for strategy and policy design.  When serious poverty challenges combine with weak government commitment to poverty reduction, the Bank often faces a dilemma: to disengage or engage in areas tangential to poverty priorities. Political factors and uncertain opportunity then complicate the country assistance strategy.  When the Bank is relatively small financially, coordination with development partners and selectivity should help focus the Bank’s role toward areas of comparative advantage. Poverty diagnostics play a key role in identifying characteristics of the poor and the constraints they face. This chapter examines the factors that condition the poverty focus of the Bank’s country strategies. It explores the extent to which country strategies and planned operational portfolios, as indicated in the Country Assistance Strategy (CAS) and Country Partnership Strategy (CPS) documents, are consistent with the poverty diagnostics discussed in chapter 3, and discusses the role of coordination and collaboration. The evidence is drawn primarily from the country case studies and a review of Country Assistance Strategy Completion Report Reviews (CASCRRs) and Country Partnership Strategy Completion Report Reviews (CPSCRRs), 1 triangulated by findings from the staff survey, stakeholder survey, and focus group meetings. Factors that Condition the Poverty Focus of Country Strategies ACTIONABLE POLICY RECOMMENDATIONS OF POVERTY DIAGNOSTICS Poverty diagnostics better inform the formulation of poverty-focused country strategies when they provide actionable policy recommendations. The Lao PDR 2010 PA, for example, was of high technical quality and was closely tailored to country conditions. It produced a consolidated list of poverty-reduction measures that were mapped to, and consistent with, the government’s National Socioeconomic Development Plan. Based on the analysis, the PA recommended construction of complete primary schools and emphasized raising schooling completion and continuation rates, expanding education access to disadvantaged groups, and 47 CHAPTER 4 FRAMING THE STRUCTURE: FORMULATING COUNTRY STRATEGIES increasing funding for recurrent expenditures. These suggestions were addressed in the CAS through the 2010 Education for All ―Fast Track Initiative, which focused on primary education for “the most educationally disadvantaged students.” Similarly, the PA’s recommendation to target Priority Districts (despite their lower populations) was addressed by the 2011 Poverty Reduction Fund 2 project, which focused on 38 districts (many or most of them Priority Districts). However, few PAs reviewed provided policymakers with adequate information about the costs and benefits of particular recommendations. For example, the Guatemala PAs (2003 and 2009), despite providing specific, prioritized, and time- bound recommendations, lacked a clear results chain and quantitative links between recommended measures and reduced poverty. In much of the diagnostic work, it was possible to predict the impact of certain interventions on particular outcomes (such as increased spending on education or health). But the results chain linking Bank operations to social indicators and then to poverty had a large element of uncertainty. There is clearly a need to better prepare consolidated and prioritized summary lists of recommendations and to better define costs, timing, administrative responsibilities, and funding sources for those recommendations. GOVERNMENT COMMITMENT Government commitment to poverty reduction is a key determinant of the focus on poverty in strategies and project selection. Of the staff survey respondents, two- fifths perceived government disinterest as a major obstacle to translating diagnostic work into Bank strategy. This result is also supported by the country studies. In Romania, Bank diagnostics found high geographic income disparities across urban, rural, and regional groups in the 2000s. The Bank’s Rural and Regional Development Loan is focused on social and economic regeneration and is based on poverty diagnostics. However, country fragmentation and delegation of power to sub regional entities complicated the efforts. The loan was not prepared due to the lack of Romanian counterparts to champion it, though it was listed in the 2006 CAS. In Egypt, despite some progress toward the Millennium Development Goals, substantial regional disparities and high concentrations of poverty persist. Public Expenditure Reviews or Public Investment Reviews would have been highly relevant to poverty-reduction strategy formulation and related policy design. Yet prior to the 2010 revolution, the government did not allow the Bank to do such analytic work on a regular basis. The Bank tried to maintain a working relationship and engage a strategically important client. Egypt did not prioritize borrowing from the Bank for poverty reduction or social sector development. Although the strategy formulated by the Bank over the years attempted to include some focus on poverty, governments in Egypt during the evaluation period (FY2004–2012) generally 48 CHAPTER 4 FRAMING THE STRUCTURE: FORMULATING COUNTRY STRATEGIES resisted these efforts. There needs to be better selectivity in terms of the Bank’s engagement in poverty reduction. The Bank operates in a complex environment in which the poverty focus of government strategy conditions the Bank’s ability to focus its country strategy on poverty. In cases where there is weak government commitment to poverty- reduction objectives, the Bank faces a challenge in promoting a poverty focus. Weak commitment can arise for several reasons: for example perceived risks to political support arising from ‘pro-poor policies’, elite capture and cronyism, or even ideological commitment to ‘trickle-down’ theories of development. Politically feasible opportunities for Bank engagement in such cases will depend critically on country context. There is no recipe book. But tools to strengthen engagement include high-quality and timely diagnostics, sustained policy dialogue, and technical assistance to better orient the country program toward addressing principle constraints to poverty reduction. The Bank can remain prepared through piloting projects, and identifying windows of opportunity to scale support when circumstances permit. The Philippines provides an example of this point. It is a middle-income country with challenging political economy issues and deeply rooted vested interests, and it has been challenging for the Bank to appropriately balance supporting government priorities and institutionalizing difficult reforms needed to foster more sustained and inclusive growth. During periods of low government commitment to poverty reduction, the Bank focused its support on key interventions (see paragraph 4.16) and helped to identify appropriate areas for additional interventions, laying the ground work for reform. The Bank used Discussion Notes and formal and informal dialogue as vehicles to provide sectoral and thematic analyses that identified key challenges and prioritized actions to help shape and inform policy discussions. The political changes that occurred after the 2010 presidential election opened new space for reforms and improvements in the political and institutional environment of the Philippines, enabling the Bank to bring greater poverty focus to the country program and scaling up the successfully piloted conditional cash transfer (CCT) programs. Obstacles to focusing on poverty can also include conflicting objectives in the Bank strategy. In Bangladesh, for example, the 2011 CAS had a broad sectoral mix of investments, including assistance to the transformative $1.2 billion Padma bridge project, which clearly addressed the regional integration highlighted in the 2008 PA. After the project’s approval, allegations of corruption involving a senior government official led to a review and renegotiation. The Bank then cancelled the project. 2 It also withdrew from other infrastructure and energy investments, including the 49 CHAPTER 4 FRAMING THE STRUCTURE: FORMULATING COUNTRY STRATEGIES proposed Poverty Reduction Support Credit. In this case, the Bank made a tough choice and sacrificed its poverty-reduction objective for its governance objective. Consistency between Formulated Country Strategies and Poverty Diagnostics Overall Bank strategies (CASs and CPSs) have been broadly consistent with underlying poverty data and diagnostics, though the priorities, mode of intervention, and instruments used vary across countries. A majority of respondents to the staff survey (72 percent) believe that the Bank’s country strategies address the causes of poverty, although to varying degrees. Growth, employment/poverty reduction, and social service delivery (specifically for health and education) or modest variants on these are typically the main pillars of the Bank’s country strategies (box 4.1). The Bank’s program in Romania consistently used the same three pillars of Bank assistance throughout the past decade: growth/employment, public sector development/reform, and poverty reduction/inclusion. During the 2000s, the Bank’s strategies in Nigeria focused on similar issues such as human development (particularly social service delivery in health and education), agricultural productivity, growth, and governance. In Malawi, the 2006 CAS objectives addressed poverty through vulnerability, agriculture, education, and nutrition after the 2006 poverty assessment. In supporting a government’s poverty-reduction strategy, the Bank’s country strategies link achievement of country priorities to lending and analytical activities by the Bank. The Bank can influence country priorities related to poverty reduction by using high-quality and timely diagnostics to underpin its policy dialogue with the government. The Bank can also deploy its resources to support the elements of country strategy that will have the greatest impact on poverty reduction. Both the general strategy and the planned lending and nonlending instruments appear generally well aligned with the Bank’s poverty diagnostics. That is, the objectives of the strategic plan are consistent with priorities identified by poverty diagnostics. This is clear from country examples. The 2008 Guatemala CPS proposed a project (Expanding Opportunities for Vulnerable Groups) directly mapped to the 2009 PA, and recommended strengthening the country’s CCT program. Extensive nonlending activities were proposed to support its poverty-related loan portfolio, anchored in periodic PAs and PA Updates. In Lao PDR, the 2012 CPS reflected recommendations of the 2006 PA to support lagging regions and districts with concentrated poverty. In Bangladesh, poverty diagnostics emphasized private sector–led growth, human development, and vulnerability with links to climate 50 CHAPTER 4 FRAMING THE STRUCTURE: FORMULATING COUNTRY STRATEGIES change, disaster preparedness, and management—all themes were reflected in country strategies. In Peru, the CPSs are consistent with the main priorities set in the diagnostic work, and the planned portfolio focuses on the main areas identified in the diagnostic, from connectivity and access for the rural poor to improved delivery of social services. Box 4-1. Poverty Focus of the Objectives of the Bank’s Country Strategies A review of the 66 countries that had at least two CPSCRRs during the evaluation period shows that the Bank seeks multiple objectives, with much variation in specifics. The number of countries pursuing at least one objective directly related to poverty increased over time (see figure, below).a Based on analysis of poverty-related pillars and objectives in country strategies―those defined as “poverty,” “education,” “health,” “social protection,” “social development,” “agriculture and rural development,” or “basic infrastructure” (for example, water and sanitation) ― education and health objectives tended to dominate, though social development and social protection objectives have become more important over time.b Percent of Countries Pursuing Objectives in the Indicated Areas CPS1 CPS2 100 Percent of Countries 80 60 40 20 0 Poverty Education Health Social Social Agriculture Basic Area: Reduction Protection Development and Rural Infrastructure Development Source: World Bank Business Warehouse (database), World Bank, Washington, DC (accessed October 2014). Note: CPS= Country Partnership Strategy a. The 66 countries used for this analysis each had two CASCRRs/CPSCRRs during the period of evaluation (FY 2004– 2012), the first period was covered by the first CASCRRs/CPSCRRs, and the second by the second CASCRRs/CPSCRRs. The first period and the second period are not the same for every country. b. There is no a priori judgment whether direct or indirect poverty focus interventions is more effective in poverty reduction. The optimal combination of the two depends on the specific context of each country. To assess the poverty focus of country strategies, a data base was assembled from the CASCRRs/CPSCRRs covering basic country data, with a list of CPS pillars and objectives was obtained from the Bank’s Business Warehouse database. There is no coding that allows the poverty categorization of CAS pillars and objectives. In an imperfect way, poverty-related pillars and objectives were broadly defined as those that directly focused on “poverty,” “education,” “health,” “social protection,” “social development,” “agriculture and rural development” and “basic infrastructure.” See appendix F for details. The Poverty Reduction Support Program (PRSP) process appeared to be a useful vehicle in aligning the poverty focus of the government’s development priorities and the poverty diagnostics. Strategies in International Development Association countries often directly linked to the country’s PRSP and the government’s 51 CHAPTER 4 FRAMING THE STRUCTURE: FORMULATING COUNTRY STRATEGIES development priorities. In the 2000s the CASs for both Bangladesh and Senegal were closely aligned with their respective PRSPs, incorporating relevant Bank and other donor-produced diagnostic work into strategy formulation (box 4.2). The alignment of the CAS with the country’s PRSP generally helped create a strong poverty focus. As discussed in box 1.1 in chapter 1, both direct and indirect interventions (growth and security, for example) are relevant to supporting poverty reduction. There is no a priori correct balance that the Bank can apply to individual countries since both are relevant to country strategy. However, where poverty levels are severe and growth is slow to reach poor populations, greater weight on direct interventions is likely needed. Box 4-2. PRSP and the Poverty Focus of the Bank’s Country Strategies World Bank country strategies respond to their authorizing environment and also reflect diagnostic work. Since 2000 the focus has been on linking Bank strategies with country strategies. For International Development Association countries, the strategies are PRSPs. During the period evaluated, the CASs for Bangladesh and Senegal, for example, were directly aligned with the preceding government PRSPs. In the 2006 and 2010 Bangladesh CASs, the corresponding PRSPs were consistent with the diagnostics of Bank-produced PAs, though government documents did not cite the assessments. Thus, by aligning the CAS with the PRSPs, there was a clear alignment of the CAS, the governments’ strategies, and the Bank’s PA. The 2006 and 2010 CASs supported all four pillars of the PRSPs (macroeconomic stability and sector growth strategies for infrastructure, safety nets, and human development) and focused on governance, which the PRSPs treated as a cross-cutting issue. PRSP I also thoroughly discussed non-income poverty diagnostics that broadened the Bank’s focus on poverty reduction. In Senegal the strategy and interventions were also based on government PRSPs. The 2003–2006 and 2007–2009 CASs were strongly aligned with the country’s first two PRSPs in creating wealth, building capacity, improving social service delivery, and protecting the vulnerable. The CASs justified the Bank’s strategies based on the PRSP and on the Bank’s comparative advantage relative to other financial and technical partners. The PRSPs were based on the available poverty diagnostics, so the Bank’s strategies were indirectly based on these diagnostics. Source: The Bangladesh and Senegal case studies. Note: A 2010 IEG report, Poverty Reduction Support Credits: An Evaluation of World Bank Support, found that Poverty Reduction Support Credits and the corresponding PRSPs aligned well with national development strategies. In some countries the Bank’s strategy focused mainly on aggregate economic growth. Growth is a vital driver of poverty reduction, but it has greater impact on poverty only when it is more inclusive and accompanied by direct poverty interventions. In Egypt, for example, the Bank’s diagnostic work found that the during the 2003–08 growth spurt, poverty declined, but extreme poverty increased. 3 The fruits of growth 52 CHAPTER 4 FRAMING THE STRUCTURE: FORMULATING COUNTRY STRATEGIES mostly bypassed rural areas, particularly in Upper Egypt, leading to the persistence of regional and zonal income disparities. 4 Role of Coordination and Consultation The Bank can better leverage its resources and have greater impact on poverty reduction in countries where it recognizes that it is a small player in the economy (in its financial support) and focuses on its comparative advantage relative to other development partners, including the government and the private sector. In the Philippines, the government had easy access to domestic and external financing on favorable terms, so the Bank had limited leverage on poverty issues through financing. During the 2000s the Bank’s team reduced the number of projects it implemented and focused on select areas where it could generate greater impact by selectively working with reform-minded champions. The Bank focused its support on piloting and scaling up a CCT program and on its community-driven development programs. Interventions to boost the productivity of agriculture and rural livelihoods were limited in scope, and efforts to push land reform (identified in the diagnostics as a major bottleneck for poverty reduction) were mostly absent. In Nigeria, where the Bank’s annual lending was only 2 percent of federal revenues, the Bank also relied on champions to have a greater impact on poverty reduction. The highest-impact interventions provided knowledge and technical support to motivated Nigerian teams within sectors and states, backing reform teams with solid analytics and practical support on issues ranging from debt relief to fiscal reform to primary health care services. Using analytic and advisory activities to inform debate and promote domestic dialogue is particularly important in resource- rich countries where the Bank has limited financial leverage. In Peru, when the government’s policies stimulated economic growth, the Bank rightly focused on the remaining pockets of poverty and tailored its services to their special needs, such as malnutrition, infant and maternal mortality, and early childhood education. The Bank’s health projects targeted the nine poorest regions and were designed to reduce maternal and infant mortality by improving both family care practices and health care during pregnancy, delivery, and breastfeeding. The Bank’s education projects, designed to improve the capacity of the Ministry of Education, focused on basic and preschool education. There is significant room for both the Bank and governments to encourage and help mobilize additional funding from other development partners, potentially including the private sector. Malawi and Lao PDR demonstrate such selectivity and donor coordination (box 4.3). 53 CHAPTER 4 FRAMING THE STRUCTURE: FORMULATING COUNTRY STRATEGIES Box 4-3. Selectivity and Donor Coordination It is critical for the Bank to be selective in its lending and nonlending activities when a large number of donors, relief agencies, and nongovernmental organizations are present. In Malawi, for example, informed by poverty diagnostics and largely based on evaluation of existing gaps and its comparative advantage, the Bank focused on specific areas of engagement in its 2007 CAS, including several with a more indirect impact on poverty. The Bank, for example, ceded leadership for health to the U.K. Department for International Development, and its leadership for intergovernmental fiscal finance to the German Agency for Technical Cooperation, but remained heavily engaged in nutrition-- an area where fewer partners were operating and successful models needed support for scaling up. Similarly, in Lao PDR, the 2012 CPS states that the Bank would not work in urbanization, water and sanitation, and agricultural inputs.a This strategy could be faulted for not adequately supporting agricultural productivity and social protection. However, the Bank’s decision to reduce its role in Lao PDR’s agricultural sector was understandable given the 2006 Vientiane Declaration’s division of labor, the need to reduce fragmentation of staff work programs, and other development partners’ strong support for the sector. The 2012 CPS argues that the exclusion of some areas of engagement, made possible through effective development partner coordination, has enabled the Bank to concentrate its resources in areas where it has the greatest strengths. Source: Malawi and Lao PDR country case studies. a. Although not specifically excluded in either the 2005 CAS or the 2012 CPS, those interviewed by the IEG team said that the World Bank Lao PDR country program did not cover conditional cash transfers for two reasons: to keep the country program more manageable, and because the government was not particularly interested in this agenda. To improve the poverty focus of country strategies, the Bank typically engages in formal consultations with stakeholders who are close to poverty-related issues, but questions arise over the consultations’ effectiveness. External stakeholders and many Bank staff see these consultations as more of an information exchange than a collaborative dialogue. In some cases, they are seen as procedural or “box checking” exercises. This finding is supported by the surveys, focus group meetings, and country cases (box 4.4). Stakeholders are at times seen as poorly informed about the Bank’s work or about strategy documents, which possibly suggests inappropriate targeting of the consultative process and weak dissemination. Staff suggested that for CAS/CPS preparation, stakeholder feedback should be incorporated on a more long-term, continual basis instead of in isolated sessions. Box 4-4. Perceptions of Stakeholders on Bank Consultation and Coordination The Bank frequently consults with donors and somewhat less frequently with other stakeholders when developing its country strategies. In the external stakeholder survey conducted for this evaluation, a majority of government respondents (71 percent) and other 54 CHAPTER 4 FRAMING THE STRUCTURE: FORMULATING COUNTRY STRATEGIES stakeholders (67 percent) stated that the Bank seeks feedback from donors in general. Also, 65 percent of government respondents and 64 percent of other stakeholders said that the Bank seeks feedback from civil society. Both government officials and other stakeholders agreed less frequently that the Bank seeks feedback from academia (56 percent of government officials and 52 percent of other stakeholders) and the private sector (54 percent of government officials and 53 percent of other stakeholders). On donor coordination, respondents from the external stakeholder survey were divided (see table, below). Over half of the government officials who responded agreed “strongly” or “somewhat” with the statement that the Bank coordinates priorities between donors and the government; however, a high percentage of government respondents also disagreed. When the same question was posed to the donors, 53 percent of respondents agreed with the statement, while 44 percent disagreed. It is natural that development partners and national agencies have their own agendas. The Bank could begin by better coordinating across its programs, demonstrating the synergies, collaboration, and leveraging that it advocates for the whole development program. External Stakeholder Opinion on Donor Coordination Response Options Government Officials Donors (percent) (percent) Agree strongly 20 42 Agree somewhat 37 11 Disagree somewhat 20 31 Disagree strongly 7 13 Don't know 14 4 Not applicable 1 0.0 Total respondents 300 55 Source: Survey of external stakeholders conducted for this evaluation. Note: 300 government officials and 55 donors responded to the survey. The data show their responses to the survey question, “When other donors provide direct support to reduce poverty in the country in which you work, the World Bank coordinates priorities between these donors and the government.” Stakeholder consultations are not weak in all cases. In the Philippines, the 2006 and 2010 CASs took a comprehensive consultative approach to CAS formulation and used the annual Philippines Development Forums to consult with other stakeholders. Through the forums, the Bank led and coordinated about 10 working groups on strategically important issues. The discussions surrounding these forums helped shape the Bank’s strategy and influence the strategies of both the government and other development partners. The Bank also continued to partner with the government, other development partners, nongovernmental organizations, civil society organizations, academia, and the private sector at different stages of country strategy formulation (though the effort and regularity of follow-up vary across groups). 55 CHAPTER 4 FRAMING THE STRUCTURE: FORMULATING COUNTRY STRATEGIES 1The review draws on the CASCRRs/CPSCRRs from 66 countries with two CASCRRs/CPSCRRs in the past decade. See appendix F for details. 2This came after the World Bank and the government had reached an agreement that put all procurement for the multibillion dollar project under effective World Bank control. The government, however, refused to take action against the particular official who figured in the allegations on the grounds that there was no evidence of any illegal activity. 3 The incidence of poverty and near-poverty in Egypt fell by about 20 percent during that high- growth period. At the same time, however, the incidence of extreme poverty (the inability to meet basic food needs) also increased by about 20 percent. See World Bank (2011), Arab Republic of Egypt Poverty in Egypt 2008–09: Withstanding the Global Economic Crisis. 4The poverty statistics of Egypt are highly sensitive to the selected poverty lines and the methodology used for their calculation. For example, using the national poverty line, poverty headcount increased from 19.6 percent in 2005 to 21.6 percent in 2009 and to 25.2 percent in 2011 (Source: WDI); using the $1.25 a day international poverty line, poverty headcount declined from 2.26percent in 2004 to 1.68 percent in 2008 (source: PovCalNet, poverty statistics for 2011 unavailable). 56 5. Building Out the House: Implementing Country Strategies Highlights  The Bank’s country strategies and the interventions supported by its lending and nonlending (advisory and technical assistance) portfolio broadly reflect the poverty-reduction strategies and development priorities of country clients.  The Bank portfolio often deviated from formulated strategy for good reason: in response to the changing external and internal environments. But it sometimes deviated because of a partner country’s weak commitment to poverty reduction, limited implementation capacity, or legislative constraints.  When the Bank’s lending and nonlending instruments complement each other, support tends to be more effective and better calibrated to country needs.  Government commitment to poverty reduction and capacity constraints are often main factors that keep budget execution in line with formulated country poverty strategy. The impact of strong analytic work and a robust strategy are lost if they do not translate into consistent strategy implementation. Hence, implementation is absolutely needed for impact on poverty reduction along the results chain. The Bank operates in a complex environment, and the choice of portfolio is conditioned by both the strategic focus of a government client and the Bank’s comparative advantage. Therefore, along the results chain, strategy implementation—which involves allocating resources and choosing beneficiaries—is inherently the most political stage. Given this, strategy implementation is highly context-dependent and complex, often less transparent, and challenging for an institution like the World Bank Group, which historically has avoided direct engagement in political culture. Due to these factors the discussion of strategy implementation relies primarily on the 10 country case studies. The chapter seeks to identify patterns and raise issues that the Bank confronts with efforts to tighten the link between strategy and implementation. This report’s limitations should also be kept in mind. While the case studies allow for a deeper investigation of portfolio choice and execution, the report does not address the question of impact and efficacy given the very difficult challenge of attribution—although individual project evaluations can address this at the micro level. 57 CHAPTER 5 BUILDING OUT THE HOUSE: IMPLEMENTING COUNTRY STRATEGIES Consistency between Bank Portfolio and Formulated Country Strategies Although the poverty focus of the strategy is important as a starting point, the implementation of the strategy and the use of Bank instruments determines the impact of Bank support for poverty reduction. In general, differences in implementation capacity, political commitment, political cycles, and the Bank’s own strategy and technical quality explain much of the cross-country differences in the Bank’s performance and the effectiveness of its support to the poverty reduction strategies, as reflected in the evaluative assessment of the 10 country case studies. 1 As indicated in the Operations Policy and Country Services guidelines, 2 the CAS should take as its starting point the country’s own vision of its development goals and its strategy for achieving them, as set out in a PRSP for IDA-eligible borrowers or a national development strategy for IBRD-eligible borrowers. It is natural to expect that the CAS and the country’s development strategy are well aligned, and if the Bank portfolio faithfully implements the CAS, it is also natural to expect strong alignment between the Bank portfolio and the CAS. However, challenges lie in the scenario in which the government’s priorities are not consistent with the Bank’s goal of poverty reduction. In this case, even if the poverty focus is well stated in the CAS documents, the alignment among the three and the poverty focus of the strategy implementation (or Bank portfolio) can often be reduced. IMPLEMENTATION UNDER CHANGING CIRCUMSTANCES The portfolio and its implementation may need to adjust to changes in internal and external factors. Changes in the strategy focus during implementation can be a sign of the Bank’s flexibility, and therefore are not always a bad thing. In some cases, deviation is strategic and necessary to match the lending portfolio to new circumstances. What is important is that the actual implementation of the Bank’s country strategy remain focused on poverty and its key income and non-income dimensions despite the volatile and complex environment in which the Bank operates. Deviations often occur because of changes in the external environment. After the 2008–2009 recession, for example, the Bank substantially increased its support for social safety nets as part of its response (box 5.1). In the Philippines, the Bank increased the value and coverage of conditional cash transfer programs (Pantawid Pamilyang Pilipino Program, for example) to support the vulnerable. 3 In Senegal, the Bank provided development policy loans (DPLs) to help sustain budgetary expenditures on health, education, and infrastructure and financed new interventions to provide social protection to the most vulnerable. These additional programs and activities were usually integrated into the ongoing Bank strategies 58 CHAPTER 5 BUILDING OUT THE HOUSE: IMPLEMENTING COUNTRY STRATEGIES and generally financed by delaying, reprogramming, or cancelling some previously envisioned activities and by making additional resources available. Box 5-1. The Bank’s Response to the Great Recession and Support for Social Safety Nets Recent IEG reports examined the World Bank Group’s response to the global crises of the late 2000s (IEG 2011c, 2012a, and 2011d). Chen and Ravallion (2009) estimate that an additional 53 million people worldwide fell into poverty in 2009 because of the financial crisis. As part of its response, the Bank increased its social protection lending and advisory services to four times the pre-crisis levels. Social protection services can include social safety nets (SSNs), active labor market programs, social investment, and pensions. Several of the country cases reviewed for this evaluation (including Bangladesh, Guatemala, Nigeria, Malawi, Romania, Senegal, and the Philippines) drew on social assistance initiatives to respond to the effects of the rapid rise in food and fuel prices and the recession of 2008– 2009. The majority of the Bank’s support to social protection programs went to SSNs (though lending to active and passive labor market programs also increased). Many countries had SSNs that were not fully prepared to respond to the impacts of the crisis. A survey of Bank staff conducted as part of an IEG review of the Bank’s support to SSN programs during 2000–2010 showed that only 16 percent of country SSNs were positioned to respond to the crises by identifying and reaching affected poor households. Weak country institutions and inadequate data were the constraints most commonly identified for Bank support to SSNs, particularly in lower-income countries (IEG 2011c). Data inadequacies included limited information on poverty and labor market outcomes and on the crisis-affected poor and vulnerable. The lack of data led many countries to focus their SSN programs broadly on the poor. This lack of data will make it difficult to assess the impact of the Bank’s support to households directly affected by the crisis. At the time the SSN evaluation was completed, the impacts of social protection interventions on households were still unknown, because many crisis-generated investment loans had not yet closed and their ex post evaluations had not been completed (IEG 2011b). Deviations often occur because of changes in the internal environment, too. When the Bank resumes policy dialogue with a new government administration in a country, there can be natural reasons for some changes to the strategy and implementation plans set before. In Peru, the formulated strategy (as written in CPS 2012) had a strong link to poverty issues and social inclusion, reflecting the findings and recommendations of the Bank’s diagnostic work. The new administration came into power and affirmed its commitment to social inclusion, and in recent years the implementation of the Bank strategy is even more strongly focused on poverty, with a clear set of interventions in the social sector to reach the poor.4 59 CHAPTER 5 BUILDING OUT THE HOUSE: IMPLEMENTING COUNTRY STRATEGIES IMPLEMENTATION WITH WEAK GOVERNMENT COMMITMENT The government commitment to poverty reduction is often a key factor in the fidelity between implementation and the formulated country strategy. As discussed in chapter 4, when commitment to poverty reduction is embraced by the government, the alignment between country strategy formulation and poverty diagnostics is often strong. Similarly, when commitment to poverty reduction is embraced by the government, the alignment between strategy implementation/choice of portfolio and strategy is often strong. For example, in Lao PDR, the country team achieved a high degree of alignment between what was planned and implemented for its 2005 CAS and 2012 CPS. 5 When there is weak government commitment or strong vested interests, implementation is much harder, and the results chain between analytic work and priorities breaks down. In Egypt, a middle-income country with strategic importance in the Middle East and North Africa Region, the relationship between the Bank’s formulated country strategy and its implementation during the review period (2004–12) offers a useful illustration of the tough choices that the Bank faced when an important client gave priority to areas other than poverty reduction. The choice is often between disengaging from significant lending or engaging in significant lending but in areas that may be only tangentially related to poverty reduction despite high national or regional poverty and the lack of shared prosperity (box 5.2). Another example is Romania. After accession to the European Union (EU) and a change in the governing coalition, interest in borrowing from the Bank waned as Romania rapidly adjusted its financing strategy toward increased use of market finance, investment loans from the European Investment Bank, and Structural and Cohesion grants from the EU. Because of declining political commitment, the 2006 Romania CPS period was marked by a halt in Bank lending: only 8 of 19 planned operations were implemented. 6 60 CHAPTER 5 BUILDING OUT THE HOUSE: IMPLEMENTING COUNTRY STRATEGIES Box 5-2. Country Assistance Strategy Formulation and Implementation: Airport Project in Egypt The statement of the poverty focus was clear and well-argued in both the Egypt Country Assistance Strategy (CAS) 2001 and CAS 2005. However, the choice of projects designed and financed by the Bank suggests that the relevance of the implemented portfolio to the Bank’s poverty-reduction strategy was weak. On the nonlending side, substantially less analytic and advisory activities were undertaken than proposed, and the core diagnostic of a Public Expenditure Review was notably absent. In order to meet its lending target and stay engaged with the authorities, the Bank agreed to a government request to finance projects that were not part of the original CASs, including the Cairo Airport in 2004.a Furthermore, the inclusion of such infrastructure projects brought the levels of lending close to what was planned (OED 2005b). The project was rated highly satisfactory by the project team and the Independent Evaluation Group, because it was generally well managed, performed well, and reflected a close working partnership with the government. But it is hard to make a strong direct linkage between such projects and reduction of poverty in areas with high and persistent poverty such as the rural areas or Upper Egypt (IEG 2009). The recent Country Partnership Framework indicated that the Bank Group strategy going forward is to focus on selectivity and for the Bank to refrain from lending in those areas that can attract private sector investment, such as airports. a. The 2001 CAS planned nine projects during 2002–2004, totaling $500 million. However, only three of the planned projects were approved, totaling $68 million through May 2004. To maintain lending levels, the Bank agreed to finance a number of projects that had not been part of the original CAS, including the Cairo Airport project, which totaled $335 million. The airport project was included in the subsequent 2005 CAS, which cited a close linkage with development of tourism. IMPLEMENTATION WITH CAPACITY AND LEGISLATION CONSTRAINTS Several country case studies show that weak implementation capacity and legislation against borrowing for recurrent expenditures were factors in the deviation of implementation from the formulated CAS. Box 5.3 discusses an example in Guatemala. In Peru, the government is not allowed to borrow externally for recurrent expenditures, although many of the most productive expenditures in the social sector may be recurrent. This limitation made it difficult to operationalize efforts to use sector-wide approaches in the social sectors. Box 5-3. Implementation Capacity and Legislation Constraints: the Case of Guatemala In Guatemala, the portfolio implemented for the 2008 CPS was quite different from the one proposed: new lending was $767 million, substantially lower than the planned $970 million. Besides the unforeseen external circumstances, legislation requirementsa and problems with project implementation weighed heavily on this lending decline. This led to long delays in project execution and a backlog of undisbursed funds that reduced both legislation and government interest in preparing investment loans. The immediate cause of the cancellation of the Enhancing Opportunities (conditional cash transfer) loan was the failure of the Guatemalan Congress to ratify the loan and government indifference to project 61 CHAPTER 5 BUILDING OUT THE HOUSE: IMPLEMENTING COUNTRY STRATEGIES implementation difficulties. Development Policy Lending (DPL) was favored over investment lending for its fast disbursement: more than 80 percent of the new loans during the 2009–2012 CPS period were DPLs.b Although the loans are widely seen as financing fiscal deficits and recurrent costs (appropriate to bridge temporary shortfalls), it is difficult to justify them if they are not accompanied by reforms to generate a strong economic framework. a. There was increasing opposition in the Guatemalan Congress to government borrowing, especially to finance recurrent expenditures. b. If the Emergency Social Services Project was also considered to be budget support, almost 100 percent of the total new lending in this CPS period would have been in the fast-disbursing category. Complementarities in Implementation TRENDS IN POVERTY-FOCUSED LENDING Although most of the Bank’s interventions contribute to poverty reduction either directly or indirectly, it is difficult to identify the right balance of “direct” and “indirect” between targeted investments and general growth promotion (targeted or not). The focus here is on Bank work with a direct poverty focus, recognizing the critical importance of growth for poverty reduction. The current coding and classification system is imprecise and inadequate to identify the poverty focus of the Bank’s interventions. 7 To address this methodological limitation and broadly gauge the poverty focus of the Bank’s interventions, this evaluation uses information from the Bank’s thematic and sector coding systems to create a weak proxy of direct poverty focus. It calculates the extent to which different instruments—specifically DPLs and investment lending—directly focus on poverty reduction as a share of total lending (annex G). 8 Using the rationale outlined in box 1.1 in chapter 1, the evaluation identified 31 themes (out of 84 total) that specifically and directly focus on the poorest or most vulnerable populations; these were then used as a proxy for “direct poverty focus.” Both direct and indirect interventions can support poverty reduction. This distinction is indicative, not judgmental, and the appropriate emphasis will vary by country. During FY2000–2012, aggregate trends suggest variation in the poverty focus of investment lending. The share of investment lending allocated to themes directly focused on poverty was about 50 percent for all Bank country clients (roughly twice the share of development policy lending allocated to themes directly focused on poverty), with a high of 58 percent in FY2003, and a low of about 35 percent in FY2008 (figure 5.1). The share of investment lending that went to areas directly focused on poverty was higher in IDA/Blend countries (about 60 percent) than in IBRD countries (about 25–30 percent). The share of investment lending to total 62 CHAPTER 5 BUILDING OUT THE HOUSE: IMPLEMENTING COUNTRY STRATEGIES lending in IBRD countries declined from nearly 30 percent in FY2000 to about 10 percent in 2008, before picking up to about 20 percent in 2012. In the same period, DPLs with themes focused directly on poverty increased slightly for all countries, with an average of nearly 23 percent, a high of more than 34 percent in FY2012, and a low of less than 11 percent in FY2008 (figure 5.2). 9 There are indications that since the 2008–2009 Great Recession, policy lending in IBRD countries has focused increasingly on areas more directly related to poverty, but not in IDA/Blend countries. For IBRD countries, DPLs with themes directly focused on poverty rose during FY2008–FY2012. Despite lower absolute numbers, the share of DPLs with themes focused directly on poverty for IDA/Blend countries was greater than for IBRD countries for the majority of the 2000s, though it declined sharply in 2010. Figure 5.1. Share of Investment Lending with Themes Directly Focused on Poverty 40 80 IL Commitemnts directly focused on poverty (percent 35 70 Total IL Commitments ($, billions) 30 60 25 50 20 40 15 30 of total) 10 20 5 10 0 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Approval FY IL commitments (not directly poverty-focused): all countries IL commitments (directly poverty focused): all countries IL commitments (directly poverty-focused), % of total IL commitments: all countries IL commitments (directly poverty-focused), % of total IL commitments: IBRD IL (directly poverty-focused), % of total IL commitments: IDA Source: World Bank Business Warehouse (database), World Bank, Washington, DC. (accessed October 2014). Note: FY= fiscal year; IBRD= International Bank for Reconstruction and Development; IDA- International Development Association; IL= investment lending. 63 CHAPTER 5 BUILDING OUT THE HOUSE: IMPLEMENTING COUNTRY STRATEGIES Figure 5.2. Share of Development Policy Lending with Themes Directly Focused on Poverty 25 40 DPL commitments directly focused on poverty (percent of 35 20 Total DPL commitments ($, billions) 30 15 25 20 10 15 total) 10 5 5 0 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Approval FY DPL Commitments (not directly poverty-focused): all countries DPL Commitments (directly poverty-focused): all countries DPL commitments (directly poverty-focused), % of total DPL commitments: all countries DPL commitments (directly poverty-focused), % of total DPL commitments: IBRD DPL commitments (directly poverty-focused), % of total DPL commitments: IDA Source: World Bank Business Warehouse (database), World Bank, Washington, DC (accessed October 2014). Note: DPL= development policy lending; FY=fiscal year; IBRD= International Bank for Reconstruction and Development; IDA= International Development Association The sharp increase in fast-disbursing policy lending in IBRD countries after the 2008 Great Recession, in level and as a share of total lending, is likely related to the efficiency of preparation and the fast-disbursing nature of policy lending. 10 Total commitments for IBRD and International Development Association (IDA) lending more than doubled from $25 billion in 2008 to $59 billion in 2010. 11 The composition of the Bank’s lending instruments also changed, with sharp differences between IBRD and IDA countries (figure 5.3). During FY2000–2012, the Bank deployed roughly one- third of its total lending to policy lending. 12 For IBRD countries, the share of policy lending was about 40 percent (except for a peak of 65 percent in 2002). The ratio for IBRD countries declined until 2008, after which there was a strong increase during the years immediately after the crisis. In FY2010, the share of policy lending in total lending climbed to nearly 50 percent, up from roughly 30 percent in 2008. But for IDA/Blend countries, the share of policy lending fluctuated around 20–30 percent until 2008, and then declined to just above 11 percent in FY2011–2012. The DPL was the Bank’s instrument of choice during the crisis because it “was generally efficient in providing for rapid increases in loan sizes and disbursement amounts” (IEG 2011c). 64 CHAPTER 5 BUILDING OUT THE HOUSE: IMPLEMENTING COUNTRY STRATEGIES During the crisis period, the preparation time for DPLs fell by roughly 30 percent to 5.9 months—3.2 months to appraisal and 2.7 months to Board approval (IEG 2012a). Figure 5.3. Share of Development Policy Lending in Total Lending 70 70 DPL Commitments, % of Total Lending commitments Total Lending commitments ($, billions) 60 60 50 50 40 40 30 30 20 20 10 10 0 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 IL commitments Approval FY DPL commitments DPL commitments, % of Total Lending commitments: IBRD DPL commitments, % of Total Lending commitments: IDA DPL commitments, % of Total Lending commitments: All Countries Source: World Bank Business Warehouse (database), World Bank, Washington, DC (accessed October 2014). Note: DPL=development policy lending; FY= fiscal year; IBRD= International Bank for Reconstruction and Development; IDA= International Development Association; IL= investment lending PORTFOLIO COMPLEMENTARITY When the Bank’s lending and nonlending instruments complement each other, support to client countries tends to be more effective and more calibrated to local needs. What is important is not which instruments are most effective for poverty reduction—DPLs, investment lending, nonlending services—but how effectively they combined with the other instruments to address particular poverty challenges. Experiences in Bangladesh, Peru, and Malawi are good examples of synergy between instruments (box 5.4). Box 5-4. Synergy of Analytical and Financial Support in Bangladesh, Peru, and Malawi In Bangladesh, the Bank’s work on safety nets—its analytic work, policy lending, investment lending, and capacity-building technical assistance—improved social protection outcomes through better geographic and beneficiary targeting and by shifting 65 CHAPTER 5 BUILDING OUT THE HOUSE: IMPLEMENTING COUNTRY STRATEGIES mechanisms from food distribution to cash transfer. The Bank began its support during the 2006 CAS period with analytic and programmatic technical assistance. Building on this, the Bank supported a pilot version of the Employment Guarantee Program for the Poorest, using a $150 million investment loan in the first year of the 2011 CAS. A scaled- up, follow-on operation began in FY2012 with a $500 million loan for the program and four other loans under the Social Safety Nets for the Poorest Project. Analytical work continued to support the implementation of the safety net programs, improving the targeting of education stipends for girls, linking the stipends to quality improvements, and promoting access. The program was modified to include boys when analytic work demonstrated that boys from poor households were being left behind. In Peru, there was strong synergy between the Accountability for Social Responsibility programmatic analytic and advisory activities and the Results and Accountability (REACT) Development Policy Loan (DPL) series, and between the REACT DPLs and the MIDIS (Ministerio de Desarrollo e Inclusión Social) DPL. The REACT and MIDIS DPLs supported the introduction of standards and monitoring stems to start strengthening beneficiaries’ “power” and hold providers more accountable. 13 The designs of investment projects in health, nutrition, and education were consistent with the diagnostic work on poverty. Technical assistance provided support to the institutional capacity of MIDIS monitoring and evaluation across several programs (such as Juntos, the school feeding programs) and programs aimed at early childhood development. In Malawi, the Bank tended to concentrate its development policy operations, including Poverty Reduction Support Credits, in sectors receiving lending and nonlending services, such as social protection and agriculture. The Community-Based Rural Land Development Project, which directly supported land reform, involved extensive nonlending services, including two impact evaluations essential to its success. Source: Country cases conducted for this evaluation. Several examples show how the Bank used its lending instruments to support poverty reduction in complementary ways. In Romania between 2009 and 2012, the Bank used investment loans and lending for budget support to provide needed resources to the health and education sectors, and to social protection. The complementarity between the two met the country’s need for reform and for resources to address the effects of the financial crisis. In Senegal, the Bank used DPLs and Poverty Reduction Support Credits to support a policy framework conducive to its poverty-related investment operations and projects. But there was less synergy between the budget support and the analytic work and technical assistance (IEG 2013). Scaling Up and Portfolio Implementation Bank-financed interventions that are explicitly focused on poverty are generally small in scale relative to the challenge of ending poverty and, in some cases, relative to the government’s own resources, particularly in middle-income countries. The impact of 66 CHAPTER 5 BUILDING OUT THE HOUSE: IMPLEMENTING COUNTRY STRATEGIES Bank support for poverty reduction depends on how well the Bank can crowd-in external resources (including those from the client government, development partners, and the private sector) to help scale up and sustain successful interventions after Bank financing has ceased. Use of pilots to lead by example and leverage other funding helps to amplify impact. Assisting in the scaling up of projects with resources of national and local government or other development partners should be an important component of the Bank strategy to reduce poverty. Some of the Bank’s highest-impact interventions go beyond financial assistance to provide knowledge and technical support to clients with solid analytics and practical advice. In many cases, particularly in middle-income countries, the balance of Bank instruments for reducing poverty and for capacity building and institutional sustainability in poverty-related areas would have been more appropriate with greater emphasis on nonlending services. In many cases, pilots are used to strengthen the design, implementation, and scaling up of projects, and to enhance the poverty focus of the Bank’s projects. Early intermediate outcomes can attract additional resources leading to scaling up (box 5.5). Box 5-5. Piloting and Scaling Up: Two Projects in Malawi Two projects in Malawi show how positive intermediate outcomes can lead to scaling up. The early success of the Malawi Social Action Fund project,a indicated in a tracer on a small version of the program, was the basis for scaling up the public works component in response to macroeconomic shocks, notably the 2005 drought. The public works component was scaled up again in response to Malawi’s 2010 and 2012 foreign exchange crises, mitigating the impact of the shocks on poor communities. The Irrigation, Rural Livelihoods, and Agricultural Development Project supported water management in poor rural areas—primarily through gravity schemes—to reduce over- dependence on rain-fed farming. The likely positive results of the intermediate outcome led to additional financing in both FY2012 and FY2013, including a scaled-up input-for- assets program to cushion the effect of the global and Malawian macroeconomic crises on the rural poor and enhance the developmental impact of small-scale irrigation. Source: Malawi country case produced for this report. a. The project has a public works component that uses community targeting and self-targeting to provide up to 12 days of wages in the lean August–September period, providing income for agricultural inputs. Scaling up requires not only clear planning and positive intermediate outcomes but also a deep understanding of the local context. In Nigeria, although the country strategies explicitly provided for scaling up, the Bank struggled to find financing modalities that actually produced service delivery results in the government structure. One such modality was community-driven development projects. These 67 CHAPTER 5 BUILDING OUT THE HOUSE: IMPLEMENTING COUNTRY STRATEGIES projects have an explicit poverty focus aimed at raising incomes in supported communities and have been scaled up to the extent possible across the country with Bank funding. However, they are not yet supported by the government’s sizable resources, and there is little indication that the program will be scaled up through effective local government linkages, despite project components aimed at doing so. Future Fidelity between Portfolio and Strategy As the Bank embraces the time-bound twin goals, challenges become more daunting. Two emerging cross-country issues have already been observed as countries implement their development and poverty-reduction strategies: changes in income distribution and changes in the degree of inclusiveness of the overall growth process. Both will affect the pace of progress against poverty and the sustainability of results. Although this evaluation does not address trends and drivers of inequality and exclusion, it is an inescapable aspect of the changing poverty profile in many countries and an important parameter embraced by the World Bank through its twin goals. The new emphasis on shared prosperity is based on recent evidence suggesting that growth alone is highly unlikely to eradicate extreme poverty by 2030 in many of the Bank’s country clients. Inequality: Sustained progress without shared prosperity is incompatible with long- term growth and stability: few countries have moved beyond middle-income status while maintaining high levels of inequality. Worsening distribution is a cause for concern not only on equity grounds but also because rising inequality can slow the pace of overall growth by creating or increasing political and social instability, and by reducing social mobility. Inclusion: Ending extreme poverty and spurring shared prosperity require mechanisms that ensure that the poor and the excluded are integral to the growth process. Social development lies at the heart of meeting the unique though heterogeneous demands of disadvantaged groups (including women and youth, ethnic minorities, and others). How a society provides opportunities will be dependent on context and time, critically linked to the country’s political economy and power-sharing arrangements and whether poor and disadvantaged people have a voice in determining national or regional economic policies. Inequality and inclusion are both very political. This points to even greater potential for a disconnect between strategy and implementation. Country commitment is critical to fidelity between the two. Ending extreme poverty and improving distribution will require mechanisms that ensure that the poor are integrated into the 68 CHAPTER 5 BUILDING OUT THE HOUSE: IMPLEMENTING COUNTRY STRATEGIES growth process. In many developing countries, the better-off have living standards that are similar to the average in the developed world; it is the poorer segments of the population that lag behind. Focusing on protecting the vulnerable from extreme deprivation, particularly irreversible human capital damage, is critical for equitable and sustainable poverty reduction. 1 Not surprisingly, the first year or two of the Country Assistance Strategy/Country Partnership Strategy (CAS/CPS) period often follows the program in the strategy document reasonably closely, because most activities in that program would have been already initiated, though they could still be modified. The later years of the program often diverge from the plan. 2 OPCS Guideline to Staff for CAS Products, April 2012. World Bank. 3The program was piloted in 2008 and reached just 6,000 households. Following the various crises of 2008–2009, it was scaled up to 376,000 households. By the end of 2010, it had been scaled up to 1 million households, and by 2013, 3 million. 4In 2012, the new government affirmed its commitment to social inclusion by creating a new Ministry of Development and Social Inclusion. The new strategy has a stronger link to poverty issues and social inclusion than the previous strategy, reflecting in part the findings and recommendations of the Bank’s diagnostic work, from connectivity and access for the rural poor to improved delivery of social services. 5Though, to adapt to the global crises, the Lao PDR country team made minor changes to adjust to government requests, changing macroeconomic or sectoral circumstances, the 2008 food crisis, and revised development partner arrangements, the volume, composition, and poverty focus of the CAS /CPS remained unchanged. 6Under the 2006 CPS, the Bank planned and approved Social Inclusion, Agriculture, Municipal Services, Knowledge Economy, Nutrient Pollution Control, Transport, Judicial Reform, and Avian Flu control projects. It dropped 11 of the 19 planned new projects, including 3 Human Development DPLs, a Rural and Regional Development project, two infrastructure projects (energy and transport), 3 Programmatic Policy Loans, and a Business Environment project. 7Currently, the interventions coded as poverty possibly reflects only a small portion of the Bank’s work that has a poverty focus. There is an indication that efforts are underway to improve the coding system. 8We use the thematic codes that are assigned to all Bank projects, which indicate approximately how much of the total loan amount is allocated to each of five possible thematic areas, to calculate a “thematically weighted” commitment amount for each project. This analysis includes all themes under Social Protection, Labor, and Risk Management; all themes under Social Development, Gender, and Inclusion; all themes under Human Development; all themes under Rural Development; and the Urban Services and Housing for the Poor theme under Urban Development—a total of 32 themes out of 82—as more directly poverty focus. The percentage of the poverty-focused theme is used as the weight to 69 CHAPTER 5 BUILDING OUT THE HOUSE: IMPLEMENTING COUNTRY STRATEGIES be multiplied by that project’s total commitments. It is important to note that this does not imply that the interventions of the 32 themes selected above are supposed to, or in reality, have larger or smaller impacts on poverty reduction than those of other themes. The purpose is mainly to group broadly the activities that are more or less directly related to poverty reduction. 9Development policy loans are designed to have impact at the national level via associated policy reform, even though they might not have a direct or immediate effect on poverty reduction per se. For this reason, it is inherently difficult to arrive at an objective assessment of a DPO’s “poverty focus.” In the figures, “poverty-themed” is used as shorthand for “directly poverty focused themed,” and “non-poverty themed” as shorthand for “indirectly poverty focused themed.” 10IEG’s report (IEG 2012a) found that during the crisis, the Bank reliance on projects that were relatively easy to prepare and negotiate (such as standalone DPOs, additional financing, and simple or repeater projects) was somewhat heavier than on other projects. The quick preparation and disbursement were consistent with the need to stabilize national economies and mitigate the impacts of the crisis. 11 This excludes MIGA and IFC commitments (IEG 2011c). 12 Except for a peak of some 50 percent in 2002. 13Specific institutional steps included in the DPLs involved enhancing the citizen identity registry to facilitate access to social services (including for children); performance budgeting for specific social program and regional entities; and reforms toward better articulating targeted programs to improve coverage and reduced leakages. 70 6. Opening the Doors and Windows: Poverty Focus of the Feedback Loops Highlights  While the Bank generates useful information on poverty reduction from its projects and programs, the feedback loops—from results to data analysis to diagnostics to strategy formulation and implementation—have generally been weak, with sizable variation across countries.  The design and implementation of country strategies need to be informed by monitoring and evaluation at the project and program levels, which also provides a basis for scaling up to better leverage resources.  Although most of the Bank’s country strategies were developed through participatory consultations with government and nongovernment stakeholders, such consultations rarely had meaningful effects on the design or implementation of Bank strategies. A strong mechanism for learning from results can help strengthen the design and implementation of the Bank’s projects and programs and improve the effectiveness of its limited resources on poverty reduction. The exploitation of information provided by M&E activities at the project, program, and country levels to feed back into data, diagnostics, and strategy formulation and implementation is therefore essential. This chapter assesses these feedback loops and their relationship to the Bank’s M&E mechanisms at the project and country strategy levels. It considers the strengths and weaknesses of the feedback loops, as well as stakeholder coordination and consultation. The analysis in this chapter is based on the country case studies, internal and external surveys, focus group meetings, and the review of CASCRRs/CPSCRRs in 66 countries. Project-Level M&E Deficiencies in the M&E design for projects are most frequently identified as shortcomings in Bank support at entry (IEG 2015). In general, collecting results information from projects has been weak. Between FY2007 and FY2013, among the 1,841 projects for which IEG assessed the M&E frameworks, 54 percent were rated “modest” and 15 percent were rated “negligible.” Only 31 percent had M&E frameworks that were rated “substantial” or “high.” There is no discernable improvement over time (figure 6.1), leaving two main shortcomings. First, outcomes are not measured as often as they should be, nor are the intended outcomes of Bank projects and programs the focus of measurement, particularly for the groups 71 CHAPTER 6 OPENING THE DOORS AND WINDOWS: POVERTY FOCUS OF THE FEEDBACK LOOPS targeted for poverty reduction. Second, when outcomes are measured they are seldom attributed to Bank-supported interventions to reduce poverty. Figure 6.1. IEG Project M&E Ratings Source: IEG Results and Performance Database. Note: The data are based on the number of rated projects that received ratings of negligible, modest, substantial, or high as a percentage of total projects rated (excluding those deemed not evaluable). Fewer than 55 percent of projects were rated in FY2000–2006, and so were excluded from the analysis. In FY2007–2013, 98 percent of projects had M&E ratings. The design, implementation, and use of M&E in Bank projects and programs can be strengthened considerably. Often what gets measured are the project milestones (mainly outputs or intermediate outcomes) but not the final outcome of poverty reduction, although practices vary. Good practice includes the Philippines ARMM Social Fund for Peace and Development (ASFP) project, which tracked improvements in income, food security, literacy rates, enrollment rates, infant mortality, malnutrition, water supply, and market access. In Guatemala, the latest series of fiscal DPLs includes project outputs that are directly targeted to poverty- relevant outcomes, 1 there are both poverty-related targets (child health and nutrition) and procedural targets. By contrast, most project-level indicators in Senegal examined outputs only implicitly linked to poverty or tracked the non- income Millennium Development Goals. 2 In Nigeria, the Bank did not produce substantial information on poverty reduction that would lead to M&E. Poverty was not included in the proposed outcomes and results to be monitored as part of assessment of the Bank’s program. In most of the country case studies, the results frameworks of projects were not linked in a substantive way to the CAS/CPS results framework or results chain. The focus of the monitored indicators on outputs or intermediate outcomes failed to provide sufficient information about the project’s impact on poverty reduction. If the objective is reducing poverty at the regional or national level, project-level M&E needs to pay more attention to explicit linkages to poverty reduction and the potential for scaling up interventions to achieve a wider effect. 72 CHAPTER 6 OPENING THE DOORS AND WINDOWS: POVERTY FOCUS OF THE FEEDBACK LOOPS A recent IEG evaluation, learning for Results in World Bank Operations: How the Bank Learns (2014), indicates that lessons from project experience are not systematically used or developed at either initiation or completion. Project completion reports, a main instrument for learning, also are generally weak in documenting and drawing lessons on poverty reduction. Lessons in the implementation completion reports (ICRs) were often too general to be useful and had limited external validity across countries (IEG 2015). While useful as information, the lessons were not specific about the implications for poverty reduction. The Senegal case study for this evaluation, for example, found that the lessons from project ICRs and ICRRs included the need for government ownership, dealing with institutional issues early in the project design, setting up an adequate M&E system with appropriate baselines and indicators, and moving to multidonor and harmonized budget support. Most staff, the IEG learning study found, viewed ICRs as an accountability tool focused on project ratings rather than learning (IEG 2015). And though reading the reports before designing projects could help prevent the repetition of mistakes, the lessons from ICRs were often only copied into appraisal documents without adjusting project design. Country-Level Results Monitoring and Learning The monitoring of CPS results covers a wide range of poverty-related areas, with most attention to education and health, but the monitoring of direct indicators for poverty is limited. Figure 6.2 shows the percentage of countries that monitored indicators in areas directly related to poverty reduction, drawing from the 66 CASCRRs reviewed. 3 Most countries focused poverty-related monitoring on three to seven areas. 4 The indicator most frequently monitored is basic infrastructure (such as water and sanitation), followed by schooling quality. 5 But only about 40 percent of CASCRs included the monitoring of a direct “poverty” indicator in the first period, which declined to some 20 percent in the second period. 6 The country case studies show varying experience. In Malawi, for example, the poverty data and feedback from the overall picture fed directly into the monitoring of the strategy and design of the subsequent CAS. There is constant reference to the dire poverty situation and the disappointment that the poverty context of the country is not changing more quickly. An explicit M&E system is used and referred to in subsequent CAS Progress Reports (CASPRs) and CAS Completion Reports (CASCRs), though poverty reduction was often only implicitly referred to in the lessons learned section of the CASCRs. However, in many countries, only a few poverty indicators are monitored in the CAS matrix. For example in Guatemala, the CAS/CPSs reviewed during the period 73 CHAPTER 6 OPENING THE DOORS AND WINDOWS: POVERTY FOCUS OF THE FEEDBACK LOOPS of evaluation present neither poverty indicators (such as headcount poverty ratios) in their results frameworks, nor changes in poverty indicators targeted as part of the M&E framework, though such changes are discussed in the background sections of the strategy documents. The inclusion of more explicit poverty-relevant indicators in CAS/CPS results matrices, and monitoring and tracking them regularly in the CAS updates, would help programs adapt to improve their effectiveness in reducing poverty. Figure 6.2. Country Monitoring of Poverty-Related Indicators Source: World Bank Database. Note: The figure shows the percentage of countries that monitored at least one indicator in areas directly related to poverty reduction from the 66 CASCRs reviewed. While there was consistency between CAS objectives and Bank interventions in some countries, it was not always clear how inputs were expected to produce certain outputs and poverty-reduction outcomes. In Nigeria, the 2010–2013 CAS updated the poverty context and generally observed the importance of non-oil growth for poverty reduction. But it was not clear how the individual strategic components related to the poverty challenges at the national or sector levels. In Guatemala, there was a thematic disconnect between poverty diagnostics and the choice of programs and projects, with an overreliance on development policy lending and limited links to poverty reduction. Most CASCRs focus on accountability rather than drawing lessons. Preparation of the CASCR focuses on the results matrix, with the lessons coming late in the process. Focus group discussions indicate that task teams seldom are well prepared to discuss sector results and lessons linked to the CAS pillars. Good practice in CAS monitoring includes Bank management commitment to the design of M&E to measure poverty-driven “indicators.” Focusing on a small number of monitorable targets, realistic indicators, and good baseline data is the key to success. In the Philippines, CAS Results Days offered a platform for country teams to break sector silos and improve the M&E system for both projects and country 74 CHAPTER 6 OPENING THE DOORS AND WINDOWS: POVERTY FOCUS OF THE FEEDBACK LOOPS programs. In Senegal, by contrast, excessive reliance on national reporting systems failed to provide clear and precise monitoring of the results chain (box 6.1). Box 6-1. CPS Results Monitoring Systems in the Philippines and Bangladesh Guided by lessons from the preceding CASCR, the Philippines country team committed to strengthening the design and monitoring system for the 2010–2012 CAS. The team created working groups mapped to each of the country strategy’s objectives. Within these groups, specific teams were assigned to track progress on their objectives. Five full-day workshops held throughout CAS implementation (CAS Results Days) allowed the teams to come together to share their progress. These workshops became part of the monitoring system for the CAS. As a result, the M&E system improved at both the country program level and the project level, though it continues to focus on the national monitoring system. The CAS included milestones to monitor progress, and supplemented them with specific and quantifiable indicators for each outcome, an improvement over the previous system. The main factor driving the difference between the two systems was the commitment of Bank staff and management to enhancing the results focus of their strategies. Similarly, in Bangladesh, a system in response to the lessons learned from the preceding CASCR was developed. In its 2011 CAS, each of the four strategic objectives was developed by a multisector results team, which identified the related outcomes and indicators, and also included elements that involved government participation. During the initial years of CAS implementation, these results teams were responsible for monitoring progress and reporting to the country team and government counterparts on a six-month/annual basis. However, though the country management effectively used the results monitoring process to modify and streamline the World Bank program in the 2013 CASPR, the regularity of this process declined over time, particularly the element which involved government participation. This was in part due to changes in Bank country management and also to shifts in the country program as well as the Bank’s relationship with the government. Source: Country cases for the Philippines and Bangladesh. Strengths and Weaknesses of the Feedback Loops across Countries Country strategies typically use just a small share of available poverty diagnostic work. Some issues are included in CASs/CPSs because they are current in the Bank (such as gender mainstreaming or shared prosperity), so the strategy is expected to refer to them. Focus group meetings suggest that country teams may not ignore poverty diagnostic work purposely, but they omit poverty analysis because it is not the “same language” they are used to, making it harder for them to understand and use a more nuanced approach to poverty. Where information is generated, it is not always used in relevant project and program design. Although this is improving, results feeding back into project implementation and design are typically limited (IEG 2012c). Impact evaluations can 75 CHAPTER 6 OPENING THE DOORS AND WINDOWS: POVERTY FOCUS OF THE FEEDBACK LOOPS have substantial knowledge spillovers to future projects and policies, especially ones that are similar to the ones evaluated. According to a recent IEG report, only 23 percent of PADs of World Bank-evaluated programs had their design or implementation influenced by previous IEs of other projects (IEG, 2012b) 7. Many project completion reports fail to mention impact evaluation evidence. 8 Only 37 percent of impact evaluations linked to a lending project were used as an input to the ICR (IEG 2012c), a missed opportunity, especially in areas with a critical mass of impact evaluations. A main weakness of the Bank’s feedback loops is that M&E of individual interventions often fail to provide broader insights to feed back into country strategies. Although many Bank country strategies articulate poverty links to individual components of the strategy, there is seldom a direct or explicit link between the activity-level outcomes and the CAS/CPS “higher-level” objectives. Piloting and scaling up of interventions are occurring in some countries. In Senegal, several pilot cases were designed for Bank operations across all three CAS/CPSs encouraging the scaling up of small or pilot projects. For access to land, a pilot activity was envisaged in the FY2003–2006 CAS. The aim was to introduce market- based mechanisms for land allocations and, once proven successful, to scale them up. The FY2007–2010 CAS supported pilot projects in education and health to provide performance-based incentives and to motivate key staff to relocate to underserved regions of the country. During the global financial crisis, the Bank scaled up its earlier pilot under the Nutrition Enhancement Project in the context of its Emergency Nutrition/Cash Transfer Project. In the Philippines, the Bank focused its support primarily on piloting and scaling up the conditional cash transfer program and the community-driven development programs. 9 The lessons from these pilots, and from the various other learning methods, helped strengthen the design, implementation, and scaling up of projects, enhancing the poverty focus of the Bank’s projects. Some CAS/CPS documents discussed in general terms the scaling up of earlier projects, either through Bank financing or through government ownership of the interventions subsequently financed through the budget or with donor resources. The Bangladesh CASs explicitly identified building on good performance during the previous period for the Reaching Out-of-School Children Project. Initially funded as a $50 million “pilot,” it received additional financing of $30 million and was followed by a second project of $130 million. The evidence is thin on whether successful pilots were used and whether they leveraged non-Bank resources, compared with a general expansion through 76 CHAPTER 6 OPENING THE DOORS AND WINDOWS: POVERTY FOCUS OF THE FEEDBACK LOOPS additional financing. From June 2005 to the end of FY2014, 14 percent of the Bank’s $233 billion in investment lending went to additional financing. 10 IEG is now evaluating additional financing for investment lending to examine whether additional financing was effectively used for scaling up or for covering cost overruns and financing gaps, project restructuring, or as simply a low-cost way to expand financing. The focus group meetings suggested that although most CAS/CPS documents stated that the lessons of the previous CASPRs or CASCRs were considered in developing the current CAS/CPS, there often was no direct evidence to that effect. Feedback does generally provide the basis for scaling up to better leverage national resources—both public and private. In several cases, the CASCR was done in the period leading up to, or in parallel with, CAS/CPS preparation. The CAS/CPS documents often only referred to the lessons in the main text and appended the CASCR as an annex to the strategy, but did not indicate how and where these lessons were taken on board—or how the lessons may have changed or affected Bank staff views. The lessons likely influenced Bank staff thinking, however. The Senegal FY2003–2006 CAS refers to two key lessons from the previous CASCR: more aid is a poor substitute for better aid, and the financial management and procurement faces serious problems in Senegal. However, it did not link these lessons to the measures it proposed. The focus groups indicated that, formally, the Bank takes steps toward learning lessons from the previous CAS/CPS, and in some cases even convenes action or lessons meetings with the country management units. The focus groups also indicated that usually the new CAS/CPS depends heavily on the existing project pipeline, the government’s national agenda, and at times the preferences or priorities of the country director. Continual learning and strong feedback loops can go a long way in strengthening the poverty focus of the country programs, as in Lao PDR. The Bank’s country team produced, or helped produce, sufficient evidence on poverty reduction and made good use of it in the design, implementation, and evolution of the poverty focus of the country program. Adequate poverty data, and good and extensive poverty diagnostics, provided strong analytical underpinning. The programmatic nature of the Poverty Reduction Support Operations (PRSOs) facilitated the evolution of the country program and its poverty focus. 11 The strength of feedback loops in a country varies along the results chain from data, to diagnostics, and to strategy formulation and implementation—and across sectors (box 6.2). In Nigeria, the feedback loops on poverty were incomplete and poverty was not explicitly included in the proposed outcomes and results matrix in the CAS to be monitored. In Romania, they were strong in response to findings from analytical work, 77 CHAPTER 6 OPENING THE DOORS AND WINDOWS: POVERTY FOCUS OF THE FEEDBACK LOOPS but less so from project M&E. Evidence from the PAs and other diagnostic work served as underpinning for support to disadvantaged groups, but it was not clear how the Bank used project monitoring data outside of follow-up projects. In Bangladesh, the feedback loops were strong in education and social protection, which supported program implementation and design through DP/government dialogue in the health and education sector-wide approach projects and in the development of follow-on operations in secondary education, skills training, and safety nets including employment programs. But the feedback loops were weak in infrastructure. Box 6-2. Variation in the Strength of Feedback Loops within a Country In Nigeria, the lack of champions demanding poverty data or diagnostics is a major reason for the incomplete feedback loop from results back to data and diagnostics. Bank- financed poverty-focused interventions were small in scale relative to the problems and the government’s own resources, although several demonstrated strong technical approaches to evaluation. Explicit attention to poverty-reduction objectives was either at a high level of generality (as in debt relief and non-oil sector growth programs) or detailed in a subset of the program (as in community-driven development programs). The Bank program did monitor the Millennium Development Goals, which include an indicator on income poverty. However, in general, the Bank did not produce substantial information on poverty reduction that would lead to effective M&E. In neither of the CASs reviewed was poverty explicitly included in the proposed outcomes and results matrix to be monitored as part of the assessment of the Bank’s program. In Romania, feedback loops from analytic and advisory activities worked primarily because the Bank has been a credible counterpart that built its reputation on poverty issues over the years with a strong record of work on data, poverty measurement, and poverty diagnostic issues. It also helped that the Bank had traction on policy advice thanks to its role as an impartial observer in a very fluid political environment. The Bank could perhaps learn more about poverty by strengthening the poverty M&E in projects and extracting more lessons from its project experiences. There is scope for strengthening the M&E of the poverty impacts from projects. In Bangladesh, the strength of feedback loops varied across sectors. In some cases they are strong, as in the social protection program in which a history of good analytic work (feeding into the PAs) contributed to a well-prioritized set of operational recommendations (focusing on efficiency, efficacy, and targeting), which then formed the basis for dialogue with the key implementing agencies and project interventions when circumstances were ripe (as described in the preceding section). Feedback is also relatively strong in education beginning with a focus on improved targeting of education stipends to increase girls’ school attendance leading to the realization that attendance rates for boys were falling and the adjustment of stipend programs to tackle this problem. Poverty-focused feedback loops have been weaker in infrastructure lending although there is an increasing interest in these sectors (notably rural roads and rural electrification) in drawing on impact evaluations to improve project design and ultimate service utilization. 78 CHAPTER 6 OPENING THE DOORS AND WINDOWS: POVERTY FOCUS OF THE FEEDBACK LOOPS Source: Country cases for Nigeria, Romania, and Bangladesh. The effectiveness of feedback loops depended on such internal factors as the priorities of Bank management and country teams and such external factors as the political commitment and administrative and technical capacity of counterparts. M&E in project implementation also varies considerably with the interests and skills of staff, though there is a general effort to improve awareness of impact evaluations and relevant staff skills. Feedback loops linking institutional strategy, budgeting, and policymaking have improved in recent years, but remained weak in the majority of countries. In some cases, the Bank’s strategic and portfolio response has lagged considerably. The enabling factors for strong feedback loops included government commitment, staff incentives, and Bank management support for measuring results and increasing technical expertise through staff training and technical assistance. Continuity in a program—which offered assurance that feedback would be used to scale up engagements and incorporate lessons in follow-on operations—also helped. Stakeholder Consultation and Coordination Most Bank country strategies include some participatory consultations with both government and nongovernment stakeholders. In Guatemala, all three country strategies involved extensive consultations and resulted in shifting the emphasis in the proposed programs. Changes to the 2005 CAS increased emphasis on environment across Bank-supported activities. They also broadened a proposed Local and Rural Development Project to include basic infrastructure services, such as water, sanitation, electrification, and information and communication technologies. And they sharpened the focus on human development, infrastructure needs, and access to finance in indigenous communities. In Senegal, the preparation of the PRSP and the CASs/CPSs included extensive consultations with government, development partners, the private sector, and civil society to ensure that the poverty policies and priorities would be well thought out and broadly supported. The FY2003–2006 CAS asserted that the Bank would seek to ensure better country ownership of the poverty-reduction agenda through closer policy dialogue, especially with civil society. Along with extensive public consultation on the development of the CAS, several Bank operations used participatory and community development methods as inputs to define local priorities. 12 The focus group discussions noted that in many cases there was no clear evidence that consultations had a significant effect on either the design or the implementation of Bank strategies. They were often treated as an opportunity for Bank staff to 79 CHAPTER 6 OPENING THE DOORS AND WINDOWS: POVERTY FOCUS OF THE FEEDBACK LOOPS inform the government and other stakeholders of the proposed strategy and interventions. The limited evidence available from the strategies and other documentation indicates that suggestions from stakeholders were not fully taken on board. The consultations usually followed a formal process, often late in the CAS/CPS process, and partly due to the tight timetable. There were limited evidence that showed the discussions had strong direct impact on the development of the strategy. In Peru, the topics discussed in the consultations differed from the major directions of the CPS in areas of relevance to reduce poverty. Multiple consultations were carried out with civil society for the 2007 CPS, in addition to consultations with the incoming and outgoing authorities and with the private sector. In different parts of the country, they focused on good development practices and income generation for the poor. But the topics discussed were not related to public spending at the local level, to access to services, or to a new social contract in health and education, a major area in the CPS. Preparing meaningful and well-informed consultations in the short timeframe for preparing a CAS or a CPS is difficult, according to staff at focus group discussions. Identifying stakeholders who know the Bank’s well for consultation can be a challenge. In some countries, the Bank relies on the government to invite stakeholders to consult on the country strategies, possibly clouding the transparency of the selection process. Finding the right stakeholders is even more difficult in fragile and conflict-affected states, where inviting stakeholders from across the country may preclude safety and security issues. Effective coordination of development partners can help the Bank concentrate its resources where it enjoys its greatest strengths, as in the selectivity of the Bank’s Lao PDR country program (see box 4.3). But both the governments and the donors have long pointed out the need to improve coordination and engagement with stakeholders. 1In addition to specific targets for the Tax/GDP ratio, which are ultimately meant to place public social spending on a more sustainable basis, the operation also has as outcome targets: (i) the percent of children under 1 year in 83 municipalities receiving the basic health/nutrition package, and (ii) the number of Zero Hunger Plan (Plan Hambre Cero) offices that have been established in the country to coordinate nutrition initiatives. 2Of the projects reviewed, the subset of lending interventions for which poverty-specific data were collected included the Community and Social Development Project and the Community Poverty Reduction Project. 3The review of the CAEs in 14 countries that covered the period of the evaluation shows the same pattern as the CASCRRs in 66 countries. 80 CHAPTER 6 OPENING THE DOORS AND WINDOWS: POVERTY FOCUS OF THE FEEDBACK LOOPS 4Out of the 66 countries reviewed, which have two CASCRRs completed during the period of evaluation, 48 monitored three to seven areas with direct poverty focus in the first CAS period reviewed, while 58 monitored three to seven areas during the second CAS period reviewed. 5 See detail in appendix F and discussions in the previous chapters. 6The 66 countries used for this analysis each had two CASCRRs/CPSCRRs during the period of evaluation (FY2004–2012), the first period was covered by the first CASCRRs/CPSCRRs, and the second by the second CASCRRs/CPSCRRs. The first period and the second period are not the same for every country, so the analysis compares countries only with themselves and not with other countries. 7 The result is drawn from a review of PADs of 117 World Bank-evaluated programs (corresponding to 142 World Bank IEs), see IEG (2012b). 8In a review of ICRs that had World Bank impact evaluations completed between 2000 and 2012, only 47 percent of the completed impact evaluations were mentioned in project completion documents (IEG 2012c). 9The 4Ps CCT program was initially piloted with 6,000 households in 2008. By 2009 it had been scaled up to 376,000 households, and it is currently being expanded nationally. To complement the initiative, the Bank also supported the development of the National Household Targeting System for Poverty Reduction (NHTS-PR), which has become the main system of identifying the poor, providing objective information for the CCT, community-driven development, and national health (PhilHealth) projects. 10The Board has approved 608 additional financing projects with a total volume of $32.7 billion, representing 14 percent of a total of $233 billion in investment lending during June 2005–June 2014. Data source: World Bank Business Warehouse (database). 11For example, the government and the Bank felt that sectoral coverage in PRSOs 1–3 was too broad, and designed the PRSOs 4–7 series so that sectoral coverage was more concentrated to improve the delivery of basic education and health services to the rural poor. The operations included the Urban Development and Decentralization Program (Cr. 3006– 12 SN), the National Rural Infrastructure Program (Cr. 3315–SN), and the Social Development Fund (Cr. 3446–SN). 81 7. Conclusions and Recommendations This evaluation examined the World Bank’s support for poverty reduction in its country programs over the period of FY2004–2012. It unpacked the main line of inquiry—”How, and how well, has the World Bank focused its programs on reducing poverty?”—into five interrelated elements in the evaluative chain—data, diagnostics, strategy formulation, strategy implementation, and feedback loops—to assess the poverty focus of the Bank’s country programs. It reviewed the adequacy of the information base and usefulness of the analytical underpinnings in support of country strategy formulation and implementation, and evaluated the consistency of the poverty focus throughout the evaluative chain and the strengths and weaknesses of the feedback loops. The evaluation finds that although the Bank has focused its support for poverty reduction in most of its country programs using a combination of lending and nonlending instruments, there is significant room for improvement. While the Bank’s work on data and diagnostics were generally good, there is significant variation in coverage and quality across countries. Moreover, for many countries, strategy formulation and implementation can be improved considerably. Areas that require attention include better reflecting the findings and recommendations of the poverty diagnostics in the country strategy formulation, enhancing the consistency of poverty focus in strategy implementation and portfolio design, and strengthening results monitoring and feedback loops. Summary of Evaluative Findings Poverty Data. The Bank provides countries with support to build capacity to collect poverty data, a global public good that serves as a foundation for poverty-reduction efforts. Programs like the Living Standards Measurement Study contribute to the collection of robust and credible data. But data availability and quality are uneven and the gaps are widest where the poverty challenge is most serious―in low-income and fragile economies. High-quality poverty data are a product of statistical systems capable of producing consistent estimates of income and non-income poverty using censuses, surveys, and administrative data. Many countries require assistance to produce statistics that meet the standards required for international reporting. Sustainability without external support is a challenge for many countries, compromising the timeliness and quality of data. Lack of political will, particularly in countries where poverty is a sensitive political topic, weak technical capacity, and costs remain major obstacles to improvements in data quality and its accessibility, 82 CHAPTER 6 OPENING THE DOORS AND WINDOWS: POVERTY FOCUS OF THE FEEDBACK LOOPS which hinders the understanding of the magnitude, geography, and nature of poverty. The ability of the World Bank to meet these challenges would be enhanced by partnering with the countries and with other development agencies to fill persisting gaps in coverage and generate good-quality data more uniformly across countries, by investing in sustainable data collection by government agencies, and by reinforcing the Bank’s responsibility to help collect, analyze, and make public data on poverty and development. There is broad scope for strengthening the Bank’s role as a global provider of high-quality and consistent poverty data. Poverty Diagnostics. The Bank produces poverty diagnostics of high technical quality, often in partnership with countries and other agencies. Poverty Assessments generally make good use of available quantitative data to derive poverty incidence indicators, identify the key drivers of poverty, and develop a poverty profile. The robustness of diagnostics is constrained mainly by weak poverty data. The best diagnostic work has been done in countries in which good-quality data were available in a timely manner. However, the country case studies suggest that the analysis in the poverty assessments often does not fully take into account the social and political economy framework for and obstacles to poverty reduction or tailor the recommendations on poverty reduction to the specific country context. The resulting absence of specific and actionable policy recommendations, along with the lack of alignment between the timing of analytic work with political cycles and weak dissemination, has in many cases weakened significantly the impact on strategy formulation of otherwise solid poverty diagnostics. Country Strategy Formulation. The Bank’s country strategy documents demonstrate awareness of and a focus on poverty reduction. With variations across countries, they are largely consistent with the poverty diagnostics, though the linkage between the analytics and the strategies often is not discussed directly or explicitly. As might be expected, the Bank’s strategy is more poverty-focused when the client government has a clear commitment to poverty reduction and its own development strategy is poverty-focused. When serious poverty challenges combine with weak government commitment to poverty reduction, the Bank faces a dilemma: to reduce engagement or continue support in areas that are tangential to poverty reduction priorities. Political factors and uncertain opportunity then complicate the Bank’s country assistance strategy and its overall effectiveness. In low-income countries, the PRSP process has been important in focusing strategies on poverty (both income and non-income), increasing cohesion in sectoral strategies and the overall macroeconomic framework, and improving coordination among donors and recipient countries. 83 CHAPTER 6 OPENING THE DOORS AND WINDOWS: POVERTY FOCUS OF THE FEEDBACK LOOPS Country Strategy Implementation. The Bank’s country strategies and the interventions supported by its lending and nonlending portfolio broadly reflect the client countries’ poverty reduction strategy and development priorities. Deviations of the Bank portfolio from formulated strategy often occurred in response to a changing external or internal environment. This is understandable and necessary when driven by external shocks or major political changes that lead to changing national strategies. Deviation may also result from weak government commitment to a poverty reduction agenda. Among the instruments, analytic and advisory activities (AAA) are usually well deployed to impact on policy and on lending related to poverty reduction. The Bank’s lending instruments, on the other hand, may not have been used enough to maximize complementarities and synergies to strengthen their collective impact on poverty reduction. In particular, project lending is often viewed narrowly and on its own terms rather than as a means of leveraging far greater non-Bank resources and having a broader and more sustained impact. Complementarity of the lending portfolio with AAA, and the complementarity between policy lending and investment lending are important and experience is varied. Feedback Loops. The Bank’s mechanisms for learning from project experience, from results to data analysis to diagnostics to strategy formulation and implementation, have generally been weak, though with significant variation across countries. The Bank does generate information and learning about poverty reduction from its programs, and in most cases, the formal processes of M&E are required but inconsistently implemented. At the project level, the processes tend to focus narrowly on outputs or immediate outcomes and often fail to reflect the broader impact of an intervention in the medium or long term. Deficiencies in the M&E design for projects are most frequently identified as shortcomings in Bank support at entry. At the strategy level, they tend to focus on the process itself (i.e., “checking the box”) without an assessment of whether a real difference is being made to poverty. In both cases, the processes are not systematically integrated in the strategy or in individual projects. Project and program-level M&E are most effective when they measure outcomes and inform the design and implementation of country strategies, providing a basis for scaling up support to better leverage resources. Improving the use of poverty data in project M&E to improve planning and implementation is crucial. Finally, the formal process for preparing a country strategy often includes some form of participatory consultation. However, there is limited evidence that such consultations have had strong effects on either the design or implementation of Bank strategies. This evaluation has focused on recent World Bank experience supporting poverty reduction in countries at a time when major changes in the Bank’s strategy and 84 CHAPTER 6 OPENING THE DOORS AND WINDOWS: POVERTY FOCUS OF THE FEEDBACK LOOPS structure have just taken place. It addresses a topic central to the Bank’s new twin goals, without evaluating major aspects of these goals, including attention to the distribution of services, income, and opportunity (shared prosperity); the importance of social inclusion (inclusive growth); and sustainability over time. While not specifically addressing the topic, this evaluation does point to the critical importance of social and political dimensions of poverty which are often less central to the Bank’s analytic work than its economic and more measureable determinants. Poverty diagnostics that factor in the structures of political authority, institutional capacity, and excluded communities and their social characteristics were found to be better prepared to provide actionable policy recommendations for reducing poverty. In many countries, a large segment of the extreme poor are isolated culturally or geographically and excluded socially and economically with little voice. Much of the public services and interventions need to fit their cultural features. Country strategies need be candid in recognizing these challenges. The social contract must include mechanisms to raise adequate resources to support these policies. Ensuring inclusion of the poor in the growth process requires investments that improve opportunities and provide safety nets to protect the vulnerable against extreme deprivation and shocks. The issues identified in this report will become increasingly important for achieving the twin goals. Strong data and diagnostics must underpin the new Systematic Country Diagnostic work to present a comprehensive view of priorities for eliminating extreme poverty and improving distribution in all countries. Strategies need to be brought into closer alignment with diagnostics and address social and political constraints, or at a minimum demonstrate awareness of them. With limited financial resources, the Bank can still support transformative change through using its diagnostic work, technical assistance, and lending instruments as pilots for scaling up and as catalysts to leverage resources from other stakeholders. Enhancing the relevance, efficiency, and effectiveness of Bank support must draw on constant learning from experience and deploying this knowledge to deliver better results. The effectiveness of the Bank’s interventions in poverty reduction will increasingly depend on how the Bank uses its diagnostic work, technical assistance, and lending instruments as pilots for scaling up and as catalysts to leverage resources from other stakeholders. Focusing on mechanisms that facilitate the deployment of analytic and advisory activities to influence policymaking, as well as on using its policy lending and investment lending in a complementary manner to crowd-in resources from other stakeholders, including particularly the partner countries, is essential for the Bank to maximize its impact on poverty reduction. 85 CHAPTER 6 OPENING THE DOORS AND WINDOWS: POVERTY FOCUS OF THE FEEDBACK LOOPS Recommendations The Bank is relatively well positioned in the early links in the results chain—data and diagnostics—capable of providing high-quality inputs for policy and decision making, but requires more consistent delivery across client countries. There is greater scope for improving Bank performance in the later links (strategy, portfolio implementation, and learning from feedback) if it is to deliver on the institutional commitment to the twin goals. The findings support recommendations in all five areas to guide improvements of the Bank’s future work on poverty reduction, in particular in the process of the design and implementation of the Systematic Country Diagnostics (SCDs) and Country Partnership Frameworks (CPFs), calibrated to individual country circumstances. POVERTY DATA • Ensure that poverty data development and reporting needs are comprehensively addressed in the SCD and country policy dialogue to identify gaps, steps to fill them, and requisite financing arrangements. • Advocate and organize support to sustainably improve the capability of national statistical agencies, both internal operational support and in partnership with external agencies. • The Bank Group should take a stronger lead in strengthening mechanisms for quality and transparency on poverty data, motivate country compliance, and regularly disseminate data. POVERTY DIAGNOSTICS • Strengthen the Bank Group’s poverty diagnostic work to ensure that it incorporates relevant social and political dimensions of poverty analysis. • Focus poverty analysis on actionable priorities for policy interventions to accelerate poverty reduction and develop the SCD discussion of linkages between recommended actions and their expected impact on poverty reduction. STRATEGY FORMULATION • Pursue the recommended actions on poverty from the SCD through CPF country strategies. • In the country strategy address the mix between indirect poverty interventions (inclusive growth) and direct poverty interventions (social safety nets, access to basic services) with attention to their sequencing to achieve the Bank’s twin goals. 86 CHAPTER 6 OPENING THE DOORS AND WINDOWS: POVERTY FOCUS OF THE FEEDBACK LOOPS STRATEGY IMPLEMENTATION AND FEEDBACK LOOPS • Develop and adopt explicit evaluation protocols for piloted interventions to capture lessons from experience on poverty reduction, with a view towards opportunities for scaling up successful interventions. • Ensure attention at project inception to evaluability through (1) developing standards for baseline measurement, (2) explicit linking of the baseline to indicators relevant to project objectives, including any that refer to poverty or inclusion impacts, and (3) robust planning for monitoring data required for ex-post evaluation. 87 References Alkire, Sabina. 2010. “The Co-creator of the UN’s New Multidimensional Poverty Index Defends her New Baby.” Oxfam-From Poverty to Power (blog), July 29. http://oxfamblogs.org/fp2p/the- creator-of-the-uns-new-multidimensional-poverty-index-defends-her-new-baby/ Benjamin, Dwayne, Loren Brandt, and John Giles. 2011. “Did Higher Inequality Impede Growth in Rural China?” Economic Journal 121 (557): 1281–1309. Berg, Andrew, Jonathan D. Ostry, and Jeromin Zettelmeyer. 2012. “What Makes Growth Sustained?” Journal of Development Economics 98 (2): 149–166. Bourguignon, François. 2004. “The Poverty-Growth-Inequality Triangle.” Working Paper 125, Indian Council for Research on International Economic Relations, New Delhi. Christiaensen, Luc, and Lionel Demery. 2007. Down to Earth: Agriculture and Poverty Reduction in Africa. Washington, DC: World Bank. Dang, Hai-Anh, Peter F. Lanjouw, and Umar Serajuddin. 2014. “Updating Poverty Estimates at Frequent Intervals in the Absence of Consumption Data: Methods and Illustration with Reference to a Middle-Income Country.” Policy Research Working Paper 7043, World Bank, Washington, DC. Dang, Hai-Anh, Peter Lanjouw, Jill Luoto, and David McKenzie. 2014. “Using Repeated Cross- Sections to Explore Movements in and out of Poverty.” Journal of Development Economics 107: 112–128. Davies, Tim. 2014. “Open Data in Developing Countries—Emerging Insights from Phase I.” Open Data in Developing Countries Working Papers 2, World Wide Web Foundation, Berlin. Demombynes, G., and J. Sandefur. 2014. “Costing a Data Revolution,” Data for Development Viewpoint, Copenhagen Consensus Center, Copenhagen. Dollar, David, Tatjana Kleinberg, and Aart Kraay. 2014. “Growth, Inequality, and Social Welfare: Cross-Country Evidence.” Policy Research Working Paper 6842, World Bank, Washington, DC. Ferreira, Francisco H. G. and Nora Lustig, eds. Forthcoming. “Appraising Cross-National Income Inequality Databases.” Journal of Economic Inequality, Special Issue. Ferreira, Francisco H. G., Phillippe G. Leite, and Martin Ravallion. 2010. “Poverty Reduction without Economic Growth? Explaining Brazil’s Poverty Dynamics, 1985–2004.” Journal of Development Economics 93: 20–36. Filmer, Deon, and Norbert Schady. 2006. Getting Girls into School: Evidence from a Scholarship Program in Cambodia. Washington, DC: World Bank. 88 REFERENCES Fosu, Augustin Kwasi. 2010. “Income Distribution and Growth’s Ability to Reduce Poverty: Evidence from Rural and Urban African Economies,” UNU–WIDER Working Paper 2010/92, United Nations University, Helinski. Fox, Louise. 2008. Beating the Odds: Sustaining Inclusion in Mozambique's Growing Economy. Washington, DC: World Bank. Gasparini, Leonardo, Federico Gutiérrez, and Leopoldo Tornarolli. 2005. “Growth and Income Poverty in Latin America and the Caribbean: Evidence from Household Surveys,” CEDLAS, Working Paper 30, CEDLAS, Universidad Nacional de La Plata, Argentina. Green, Duncan. 2010. “How Can We Improve the Way We Measure Poverty? The UN’s New Poverty Index (and Groovy Graphics).” Oxfam-From Poverty to Power (blog), July 27. http://oxfamblogs.org/fp2p/can-we-improve-the-way-we-measure-poverty-the-uns-new- poverty-index/ Herzer, Dierk, and Sebastian Vollmer. 2012. “Inequality and Growth: Evidence from Panel Cointegration,” Journal of Economic Inequality 10: 489–503. IEAG (Independent Expert Advisory Group on a Data Revolution for Sustainable Development). 2014. A World That Counts: Mobilising The Data Revolution for Sustainable Development. New York, NY: Independent Advisory Group Secretariat. IEG (Independent Evaluation Group). 2007. Philippines—Client Perspectives on Elements of World Bank Support. Washington, DC: World Bank. ―――. 2009. Egypt: Positive Results from Knowledge Sharing and Modest Lending. An IEG Country Assistance Evaluation 1999–2007. Washington, DC: World Bank. ―――. 2010. Poverty Reduction Support Credits: An Evaluation of World Bank Support. Washington, DC: World Bank. ―――. 2011a. Assessing IFC’s Poverty Focus and Results. Washington, DC: World Bank. ―――. 2011b. Social Safety Nets: An Evaluation of World Bank Support, 2000–2010. Washington, DC: World Bank. ―――. 2011c. The World Bank Group's Response to the Global Economic Crisis: Phase I. Washington, DC: World Bank. ―――. 2011d. The World Bank’s Involvement in Global and Regional Partnership Programs: An Independent Assessment. Washington, DC: World Bank. _____ 2012a. Liberia Country Program Evaluation: 2004–2011. Washington, DC: World Bank. ―――.2012b. World Bank Group Impact Evaluations: Relevance and Effectiveness. Washington, DC: World Bank. ―――.2012c. The World Bank Group’s Response to the Global Economic Crisis—Phase II. Washington, DC: World Bank. 89 REFERENCES ―――.2013. “Senegal-Country Assistance Strategy for the period FY07–FY10.” Country Assistance Strategy Completion Report Review, World Bank, Washington, DC. ―――.2014. Learning and Results in World Bank Operations: How the Bank Learns. Washington, DC: World Bank. ―――. 2015. Results and Performance of the World Bank Group 2014. Washington, DC: World Bank. IMF (International Monetary Fund) 2012. “Data Quality Assessment Framework—Generic Framework.” IMF, Washington, DC. Inchauste, Gabriela, Sergio Olivieri, Jaime Saavedra, and Hernan Winkler. 2012. “What is Behind the Decline in Poverty Since 2000,” Policy Research Working Paper 6199, World Bank, Washington, DC. IYCN (Infant and Young Child Nutrition) Project. 2011. Consulting with Caregivers: Formative Research to Determine the Barriers and Facilitators to Optimal Infant and Young Child Feeding in Three Regions of Malawi. Washington, DC: IYCN. Knowles, Stephan. 2005. “Inequality and Economic Growth: The Empirical Relationship Reconsidered in the Light of Comparable Data.” Journal of Development Studies 41 (1): 135–159. Kraay, Aart. 2006. “When Is Growth Pro-Poor? Evidence from a Panel of Countries.” Journal of Development Economics 80: 198–227. Loayza, Norman V., and Claudio Raddatz. 2006. “The Composition of Growth Matters for Poverty Alleviation.” Policy Research Working Paper 4077, World Bank, Washington, DC. Lustig, Nora, Omar Arias, and Jamele Rigolini. 2002. “Poverty Reduction and Economic Growth: A Two-Way Causality,” IDB Publications 53919, Inter-American Development Bank, Washington, DC. McKinsey Global Institute. 2011. Big data: The next frontier for innovation, competition, and productivity. Washington, DC: McKinsey Global Institute. OECD (Organisation for Economic Co-Operation and Development). 2007. “Promoting Pro-Poor Growth: Private Sector Development.” In Promoting Pro-Poor Growth: Policy Guidance for Donors, 57–131. Paris: OECD. ―――. 2005b. “Egypt CAS Completion Report Review.” CAS Completion Report Review 89073, World Bank, Washington, DC. QAG (Quality Assurance Group). 2003. “Quality of ESW in FY02, A QAG Assessment.” Quality Assurance Group Assessment, World Bank, Washington, DC. Ravallion, Martin. 1997. “Can High Inequality Development Countries Escape Absolute Poverty?” Economics Letters 56: 51–57. ―――. 2001. “Growth, Inequality and Poverty: Looking Beyond Averages.” World Development 29(11): 1803–1815. 90 REFERENCES ―――. 2007. “Inequality is Bad for the Poor.” In Inequality and Poverty Re-Examined, edited by J. Micklewright and S. Jenkins, 37–61. New York: Oxford University Press ―――. 2010. “Guest Blog: World Bank Research Director Critiques the New UN Poverty Index.” Oxfam-From Poverty to Power (blog), July 28. http://oxfamblogs.org/fp2p/guest-blog-world- bank-research-director-critiques-the-new-un-poverty-index/ Ravallion, Martin, and Gaurav Datt. 1996. “How Important to India’s Poor is the Sectoral Composition of Economic Growth?” World Bank Economic Review 10 (1): 1–25. ―――. 2002. “Is India's Economic Growth Leaving the Poor Behind?” Journal of Economic Perspectives 16 (3): 89–108. Ravallion, Martin, and Shaohua Chen. 1997. “What Can New Survey Data Tell Us About Recent Changes in Distribution and Poverty?” World Bank Economic Review 11(2): 357–82. ―――. 2009. “The Impact of the Global Financial Crisis on the World’s Poorest.” Vox CEPR’s Policy Portal (blog), April 30. http://www.voxeu.org/article/impact-global-financial-crisis-world-s- poorest. Thomson, Anne, Graham Eele, Felix Schmieding. 2013. Independent Evaluation of the International Household Survey Network (IHSN) and Accelerated Data Program (ADP). Final Report. Oxford: Oxford Policy Management. UNDP (United Nations Development Programme). 2004. Unleashing Entrepreneurship: Making Business Work for the Poor. New York: UNDP. Voitchovsky, Sarah. 2005. “Does the Profile of Income Inequality Matter for Economic Growth?: Distinguishing Between the Effects of Inequality in Different Parts of the Income Distribution.” Journal of Economic Growth 10: 273–296. World Bank. 1990. World Development Report 1990: Poverty. New York: Oxford University Press. ―――. 2001a. “Country Assistance Strategy for the Arab Republic of Egypt.” Country Assistance Strategy Document 22163, World Bank, Washington, DC. ―――. 2001b. “Country Assistance Strategy for the People’s Republic of Bangladesh.” Country Assistance Strategy Document 21326, World Bank, Washington, DC. ―――. 2001c. “Country Assistance Strategy for Romania.” Country Assistance Strategy Document 22180, World Bank, Washington, DC. ―――. 2001d. World Development Report 2000/2001: Attacking Poverty. Washington, DC: World Bank. ―――. 2002a. “Country Assistance Strategy for the Republic of Peru.” Country Assistance Strategy Document 24205, World Bank, Washington, DC. ―――. 2002b. “Poverty in Bangladesh: Building on Progress.” Sector Report 24299, World Bank, Washington, DC. 91 REFERENCES ―――. 2003a. Country Assistance Strategy for the Republic of Senegal.” Country Assistance Strategy Document 25498, World Bank, Washington, DC. ―――. 2003b.” Kyrgyz Republic: Enhancing Pro-poor Growth.” Poverty Assessment 24638, World Bank, Washington, DC. ―――. 2003c. “Poverty in Guatemala.” Poverty Assessment 24221, World Bank, Washington, DC. ―――. 2004a. “Guidance Note on Poverty Assessments.” Guidance Note, World Bank, Washington, DC. ―――. 2004b. “Papua New Guinea: Poverty Assessment.” Poverty Assessment 66746, World Bank, Washington, DC. ―――. 2004c. “Recession, Recovery and Poverty in Moldova.” Report 28024, World Bank, Washington, DC. ―――. 2005a. “2004 Annual Review of Development Effectiveness (ARDE): The World Bank's Contributions to Poverty Reduction.” Annual Report 32136, World Bank, Washington, DC. ―――. 2005b. “Country Assistance Strategy for the Arab Republic of Egypt for the Period FY06–FY09.” Country Assistance Strategy Document 32190, World Bank, Washington, DC. ―――. 2005c. “Country Assistance Strategy for the Lao People’s Democratic Republic (Lao PDR).” Country Assistance Strategy Document 31758, World Bank, Washington, DC. ―――. 2005d. “Country Assistance for the Republic of Guatemala.” Country Assistance Strategy Document 31776, World Bank, Washington, DC. ―――. 2005e. “Country Assistance Strategy for the Republic of the Philippines.” Country Assistance Strategy Document 32141, World Bank, Washington, DC. ―――. 2005f. “Country Partnership Strategy for the Federal Republic of Nigeria (2005–2009).” Country Partnership Strategy Document 32412, World Bank, Washington, DC. ―――. 2005g. “Ethiopia—Well-Being and Poverty in Ethiopia: The Role of Agriculture and Agency.” Poverty Assessment 29468, World Bank, Washington, DC. ―――. 2005h. “Peru Opportunities for All: Peru Poverty Assessment.” Poverty Assessment 29825, World Bank, Washington, DC. ―――. 2005i. World Development Report 2006: Equity and Development. Washington, DC: World Bank. ―――. 2006a. “Country Assistance Strategy for the People’s Republic of Bangladesh for the Period FY06–09.” Country Assistance Strategy Document 35193, World Bank, Washington, DC. ―――. 2006b. “Country Partnership Strategy for the Republic of Peru.” Country Partnership Strategy Document 37913, World Bank, Washington, DC. 92 REFERENCES ―――. 2006c. “Country Partnership Strategy for Romania for the Period FY06–FY09.” Country Partnership Strategy Document 36147, World Bank, Washington, DC. ―――. 2006d. “Indonesia—Making the New Indonesia Work for the Poor.” Poverty Assessment 37349, World Bank, Washington, DC. ―――. 2006e. “Lao PDR Poverty Assessment Report: From Valley's to Hilltops—15 Years of Poverty Reduction.” Poverty Assessment 38083, World Bank, Washington, DC. ―――. 2007a. “Arab Republic of Egypt: Poverty Assessment Update, Volumes 1 and 2.” Poverty Assessment, World Bank, Washington, DC. ―――. 2007b. “Brazil—Measuring Poverty Using Household Consumption.” Poverty Assessment 36358, World Bank, Washington, DC. ―――. 2007c. “Country Assistance Strategy for the Republic of Senegal for the Period FY07–FY10.” Country Assistance Strategy Document 36608, World Bank, Washington, DC. ―――. 2007d. “Country Assistance Strategy of the World Bank for the Republic of Malawi.” Country Assistance Strategy Document 38326, World Bank, Washington, DC. ―――. 2007e. “Malawi Poverty and Vulnerability Assessment: Investing in Our Future.” Poverty Assessment 36546, World Bank, Washington, DC. ―――. 2007f. “Republic of Yemen Poverty Assessment, Volumes 1–4.” Poverty Assessment 53076, World Bank, Washington, DC. ―――. 2008a. “Country Partnership Strategy for the Republic of Guatemala.” Country Partnership Strategy Document 44772, World Bank, Washington, DC. ―――. 2008b. “Poverty Assessment for Bangladesh: Creating Opportunities and Bridging the East- West Divide.” Poverty Assessment 44321, World Bank, Washington, DC. ―――. 2008c. “Sénégal—Diagnostic de la Pauvreté.” Poverty Assessment 44455, World Bank, Washington, DC. ―――. 2009a. “Country Assistance Strategy for the Republic of the Philippines for the Period FY2010– 2012.” Country Assistance Strategy Document 47916, World Bank, Washington, DC. ―――. 2009b. “Country Partnership Strategy for the Federal Republic of Nigeria (2010–2013).” Country Partnership Strategy Document 46816, World Bank, Washington, DC. ―――. 2009c. “Country Partnership Strategy for Romania for the period July 2009–June 2013.” Country Partnership Strategy Document 48665, World Bank, Washington, DC. ―――. 2009d. “Guatemala Poverty Assessment: Good Performance at Low Levels.” Poverty Assessment 43920, World Bank, Washington, DC. ―――. 2009e. “Nigeria—Employment and Growth Study.” Public Sector Study 51564, World Bank, Washington, DC. 93 REFERENCES ―――. 2010a. “Country Assistance Strategy for the People’s Republic of Bangladesh for the Period FY11–14.” Country Assistance Strategy Document 54615, World Bank, Washington, DC. ―――. 2010b. Poverty Status in Afghanistan—A Profile Based on the National Risk and Vulnerability Assessment (NRVA) 2007–08. Washington, DC: World Bank. ―――. 2010c. “Philippines—Fostering More Inclusive Growth.” Policy Reduction Strategy Paper 49482, World Bank, Washington, DC. ―――. 2010d. “Poverty in Lao PDR 2008: Lao Expenditure and Consumption Survey 1992/03– 2007/08.” Poverty Study 70411 (Laotian), World Bank, Washington, DC. ―――. 2010e. World Development Report 2010: Development and Climate Change. Washington, DC: World Bank. ―――. 2011a. “2011 Philippines Development Report: Generating Inclusive Growth to Uplift the Poor.” Poverty Study 66640, World Bank, Washington, DC. ―――. 2011b. Perspectives on Poverty in India: Stylized Facts from Survey Data Volumes I and II. Washington, DC: World Bank. ―――. 2011c. “World Bank Corporate Scorecard 2011: Integrated Results and Performance Framework.” Working Paper 80234, World Bank, Washington, DC. ―――. 2011d. World Bank for Results 2011. Washington, DC: World Bank. ―――. 2012a. “Country Assistance Strategy for the Republic of Malawi for the Period FY13–FY16.” Country Assistance Strategy Document 74159, World Bank, Washington, DC. ―――. 2012b. “Country Partnership Strategy for Lao People’s Democratic Republic for the Period FY12–FY16.” Country Partnership Strategy Document 66692, World Bank, Washington, DC. ―――. 2012c. “Country Partnership Strategy for the Republic of Guatemala for the Period 2013–2016.” Country Partnership Strategy Document 69229, World Bank, Washington, DC. ―――. 2012d. “Interim Strategy Note for the Arab Republic of Egypt.” Interim Strategy Note 66443, World Bank, Washington, DC. ―――. 2012e. “Country Partnership Strategy for the Republic of Peru for the Period FY12–FY16.” Country Partnership Strategy Document 66187, World Bank, Washington, DC. ―――. 2012f. World Development Report 2013: Jobs. Washington, DC: World Bank. ―――. 2013a. “Bangladesh Poverty Assessment: Assessing a Decade of Progress in Reducing Poverty 2000–2010.” Working Paper (Numbered Series) 78559, World Bank, Washington, DC. ―――. 2013b. “Country Partnership Strategy for the Federal Republic of Nigeria for the Period FY 2014–FY 2017.” Country Partnership Strategy Document 82501, World Bank, Washington, DC. ―――. 2013c. “Country Partnership Strategy (FY2013–2017) for the Republic of Senegal.” Country Partnership Strategy Document 73478, World Bank, Washington, DC. 94 REFERENCES ―――. 2013d. “Philippine Development Report: Creating More and Better Jobs.” Development Policy Review ACS5842, World Bank, Washington, DC. ―――. 2014a. “Country Partnership Strategy (CPS) for Romania for the Period 2014–2017.” Country Partnership Strategy Document 84830, World Bank, Washington, DC. ―――. 2014b. A Measured Approach to Ending Poverty and Boosting Shared Prosperity: Concepts, Data, and the Twin Goals. Washington, DC: World Bank. ―――. 2014c. Prosperity for All/Ending Extreme Poverty: A Note for the World Bank Group Spring Meetings 2014. Washington, DC: World Bank. ―――. 2014d. “Trust Fund for Statistical Capacity Building, Annual Progress Report: April 1, 2013 – March 31.” Annual Report 88510, World Bank, Washington, DC. ―――. 2014e. The World Bank Group Strategy. Washington, DC: World Bank. ―――. 2015a. A Measured Approach to Ending Poverty and Boosting Shared Prosperity: Concepts, Data, and the Twin Goals. Washington, DC: World Bank. ―――. 2015b. Purchasing Power Parities and the Real Size of World Economies: A Comprehensive Report of the 2011 International Comparison Program. Washington, DC: World Bank. World Bank and IMF (International Monetary Fund). 2004. Global Monitoring Report 2004: Policies and Actions for Achieving the Millennium Development Goals and Related Outcomes. Washington, DC: World Bank. ―――. 2015. Global Monitoring Report 2014/2015: Ending Poverty and Sharing Prosperity. Washington, DC: World Bank. World Bank and UNICEF (United Nation’s Children’s Fund). 2009. Romania—Rapid Assessment of the Impact of the Economic Crisis on Poverty. Washington, DC: World Bank. 95 Appendix A. Summaries of 10 Country Case Studies Independent Evaluation Group (IEG) evaluators prepared case studies of 10 countries in the first half of 2014. The countries were selected from a population of 144 countries comprising all International Development Association (IDA), International Bank for Reconstruction and Development (IBRD), and blend countries. 1 The selection is purposive and does not aim to fully represent the various categories of countries. It tries to cover a range of countries at different income levels to reflect the differing approaches and challenges to poverty reduction in countries at different levels of development. The case study countries were selected to roughly reflect regional balance. During the selection process, countries were first grouped according to (i) regions; (ii) income level; and (ii) whether or not they are classified as fragile and conflict affected states (FCS). An emphasis was placed on countries with significant Bank engagement (lending and nonlending activities). To provide lessons that reflect a wide range of operational experience, the final selection of countries also took into consideration variations in the number of poor people and in poverty rates, and potential lessons for learning in consultation with external experts and senior Bank staff. The 10 countries selected for study are Bangladesh, Egypt, Guatemala, Lao People’s Democratic Republic (Lao PDR), Malawi, Nigeria, Peru, the Philippines, Romania, and Senegal. Focusing on FY2004– 12, each case study consisted of desk reviews, structured interviews with Bank staff, and in-country consultations with stakeholders. The overarching question they sought to answer was: “How, and how well, does the World Bank focus its programs on reducing poverty in partner countries?” To address this issue, they asked the following questions: • Did the Bank have the appropriate data to understand the nature of poverty and provide an information base for robust analytical work on poverty? • Did the Bank’s diagnostics guide development programs to effectively reduce poverty? • Have Bank country strategies adopted the findings of analytical work on poverty to help prioritize and guide policy dialogue and lending? • Have interventions—operations, technical assistance, and capacity building— reflected the strategic priorities for poverty reduction? • Has the Bank collected and drawn lessons from poverty reduction interventions to strengthen feedback loops and improve the effectiveness of its country strategies and programs? If so, how did it do so? 96 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES These questions, detailed in the Approach Paper, formed the basis of the evaluation exercise. During their field visits, each of the 10 evaluation teams closely followed the country case study protocol. The countries were selected to reflect all regions as well as different income levels, numbers of poor people, and poverty rates. Countries of special interest (e.g., International Bank for Reconstruction and Development [IBRD], International Development Association [IDA], fragile states, small states) are represented. The case studies illustrate variations in country program work in poverty reduction, given different institutional and political contexts. Comparative tables are presented at the end of this appendix. All country case studies were prepared after a thorough review by IEG evaluation teams of related Bank documents as well as reports or documents prepared by country authorities. These reviews were then followed by field visits and in-country consultations with stakeholders as well as Bank managers and staff working on the 10 countries. The evaluative evidence provided by the 10 country case studies yields a rich and complex picture of the World Bank’s support and its dynamics in the areas of poverty data, poverty diagnostics, and formulation and implementation of poverty reduction strategies through the Bank’s lending and non- lending portfolio during the evaluation period of FY2004–12. Differences in implementation capacity, political commitment, political cycles, and the Bank’s own strategy and technical quality explain much of the cross-country differences in the Bank’s performance and the effectiveness of its support to the poverty reduction strategies of the 10 countries. External shocks also seem to have played an important role in determining the quality and nature of Bank support, as shown by the impact of the 2008–09 global financial crisis on the Bank’s portfolio in each of the following summaries of the country case studies for this evaluation. Summary of Country-Specific Findings The evidence provided by the 10 country case studies yields a rich and complex picture of the World Bank’s support and its dynamics in the areas of poverty data, poverty diagnostics, the formulation and implementation of poverty reduction strategies, and the related feedback loops, through the Bank’s lending and nonlending portfolio during the evaluation period of FY2004–12. Differences in implementation capacity, political commitment, political cycles, and the Bank’s own strategy and technical quality explain much of the cross-country differences in the Bank’s performance and the effectiveness of its support to the poverty reduction strategies of the 10 countries. External shocks also seem to have played an important 97 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES role in determining the quality and nature of Bank support. The 2008–09 global financial crisis resulted in a marked change in the Bank’s planned portfolios for its clients (as envisaged in Country Assistance Strategies [CASs] and Country Partnership Strategies [CPSs]), including an increase in the ratio of fast-disbursing operations to all commitments as well as some changes in project lending and analytical and advisory assistance (AAA) in response to clients’ needs. Seven of the 10 case study countries—Bangladesh, Guatemala, the Lao People’s Democratic Republic, Malawi, Peru, Romania, and Senegal—received more calibrated and sustained support from the Bank for their poverty measurement efforts. In these seven countries plus the Philippines, the Bank’s poverty-related diagnostic work, which included recommendations, helped create strong results chains with respect to the formulation and implementation of poverty reduction strategies. In all of these countries (albeit with significant variance in terms of the quality and timeliness of outcomes), Bank support helped countries produce and maintain good-quality poverty data and prepare relevant, timely, and sound diagnostic work on a more or less sustained basis. In Bangladesh, Lao PDR, Malawi, and Senegal, where local capacity was weak, the Poverty Reduction Strategy (PRS) process seems to have played an important role by focusing the attention of policy makers and the donor community and directing their financial and technical assistance toward meaningful measurement and analysis of poverty efforts, with key targets and instruments that underpinned an actionable overall poverty reduction strategy over time. In Bangladesh, Lao PDR, Malawi, and Peru, strong data and diagnostic work enabled the Bank to formulate poverty reduction strategies that were calibrated to the realities on the ground and reflected the key findings and recommendations of poverty diagnostic work. Bank support of strategy implementation and feedback loops (from results to data analysis to diagnostics to strategy formulation and so forth) was more nuanced in these countries. Its effectiveness depended on a variety of internal and external factors, including the election cycle, political commitment, the administrative and technical capacity of counterparts, and external shocks. The quality and effectiveness of the Bank’s overall support for poverty reduction appears to have been somewhat stronger in Lao PDR and Peru, particularly recently, though support to Bangladesh, Malawi, and Romania was also significant, relevant, timely, of good quality, and appreciated by counterparts. In contrast, the overall quality and effectiveness of the Bank’s support for poverty reduction in Guatemala, Nigeria, the Philippines, and Senegal was moderate/fair. In Guatemala and Senegal, the quality of poverty-related data support and diagnostic 98 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES work was strong, but in-country capacity limitations and other constraints, such as budgetary pressures, limited the Bank’s effectiveness. Bangladesh, Lao PDR, and Malawi had reasonably strong feedback loops from data to diagnostics to strategy formulation and implementation of Bank country strategies. Feedback loops in the other cases study countries were weaker, often as a result of inadequate monitoring and evaluation (M&E) systems at the project and program levels. In Nigeria—home to the world’s third-largest population of people in extreme poverty—no local champions demanded poverty data and diagnostics. As a result, data quality issues persisted, and Bank-financed, poverty-focused interventions were small relative to the size and complexity of the problems and the government’s own resources. The Bank’s poverty diagnostics generally made good use of available data, which expanded over time, and included disaggregated poverty profiles. But the Bank’s poverty diagnostics did not drive the strategies, because governance issues became an institutional priority. In the Arab Republic of Egypt, the country with the weakest and most limited poverty reduction strategy among the 10 country case studies. The government gave priority to areas other than poverty reduction during the evaluation period. 2 It adversely affected the availability of poverty data. It was not possible to assess the quality of the data due to the limited data accessibility. Weak data made it difficult for the Bank to conduct good-quality diagnostics in a timely fashion, although it did end up producing a few high-quality diagnostic reports during the review period. In both Egypt and Nigeria, the political economy context appears to have constricted the strategy space in which the Bank operated. Because of the nature of the client, the Bank’s country strategies for Egypt and Nigeria did not include clear road maps or integral and consistent visions of the sustained interventions needed to reduce poverty, and implementation deviated widely from the plans set forth in the Bank’s strategy documents. This comparative summary of the findings of the country case studies yields very broad conclusions. The 10 case studies suggest that Bank support for poverty reduction had a positive impact on various components, such as poverty data, poverty diagnostics, and strategy formulation and implementation, even though results chains and feedback loops were weak in a number of cases. Most countries in the sample responded positively to the Bank’s efforts, taking advantage of the Bank’s technical expertise to improve the quality and timeliness of their poverty data, aligning their development strategy with the strategy implied by the Bank’s diagnostics, and defining (in some cases redefining) their national poverty goals. Most of these countries also allowed some recalibration of the Bank’s project and program portfolio to refocus on areas where poverty was deeper or where a 99 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES significant portion of the population did not share the prosperity generated by the economic growth process at the national level. Cross-Cutting Themes The country case studies address a number of cross-cutting issues that have direct bearing on the nature of Bank support for poverty reduction and strategy formulation. This section provides a brief summary of two of these issues—income inequality and social inclusion—which are likely to become increasingly important to the implementation of the Bank’s strategy going forward. INEQUALITY In many economies, growth in the incomes of the poor is accompanied by falling inequality, as witnessed in many Latin American countries during the past decade. Some rise in inequality may be necessary to generate growth—by creating incentives that reward innovation and risk-taking, and inducing firms and individuals to invest in human and physical capital. This phenomenon occurred in a number of East Asian countries a few decades ago, when income growth of the bottom 40 percent was rapid but still lagged the growth of average income. Income growth of the bottom 40 percent that is consistently lower than average should be a cause for concern—not only on sheer equity grounds, but also because the resulting rise in inequality could eventually slow the pace of overall growth, by affecting the quality of institutions, creating or increasing political and social instability, and reducing mobility in society. Sustained progress in shared prosperity is incompatible with a long-term increase in inequality: no country has transited beyond middle-income status while maintaining high levels of inequality. One could, therefore, argue that where inequality is high, boosting shared prosperity is likely to require that the income of the bottom 40 percent grow faster than average income. Between the early to mid-1990s and 2012, poverty headcounts declined significantly, albeit to varying degrees, in all 10 countries studied. In contrast, income inequality, as measured by the Gini coefficient, fell in only six countries (Egypt, Guatemala, Malawi, Nigeria, the Philippines, and Senegal). Inequality rose in Bangladesh, Lao PDR, Peru, and Romania. A somewhat different picture emerges from comparison of inequality in the early 2000s and the latest level available, which overlaps with the period of this evaluation. Over this period, inequality as measured by Gini rose in Lao PDR, Malawi, Nigeria, and Senegal. The rise in inequality in the three Sub-Saharan countries is particularly disconcerting, as it was already high before the recent rise. 100 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES The trend reflects both slow growth and slow poverty reduction between 2009 and 2013, in the aftermath of the global crisis. As measured by the share of income or consumption of the top 10 percent of the population relative to the share of income or consumption of the bottom 40 percent, inequality rose in Lao PDR, Malawi, Nigeria, Romania, and Senegal during the same period. Among the 10 country cases, inequality has been highest (with Gini in the range of 40 to 52) in Guatemala, Malawi, Peru, the Philippines, and Senegal, where issues of exclusion, ethnicity, and lagging regions, have not been addressed effectively, at least until recently. In Egypt the apparent increase in inequality and vulnerability of the population, particularly among domestic migrants and informal workers may not have been adequately captured by the official poverty data prior to the 2010–11 revolution, which, until recently, had indicated a significant decline in both extreme poverty and inequality between 2000 and 2008. Distribution of the fruits of growth is a major issue in many countries, although the poor quality of data in some cases casts doubt on the evidence. When the initial level of income or non-income inequality is high, the growth rate is lower and has a more subdued effect on poverty. INCLUSION Spurring shared prosperity requires mechanisms that ensure that the poor and the dispossessed are integral to the growth process. No single institutional arrangement for ensuring such inclusion will be optimal for all societies; how a society provides these opportunities will be context- and time-dependent and critically linked to the country’s political economy and power-sharing arrangements, and whether poor and disadvantaged people have a voice in determining national or regional economic policies. The notion of shared prosperity requires that progress be sustainable over time and across generations, in terms of the environment, social inclusion, and fiscal prudence. However, ensuring inclusion of the poor in the growth process requires investments that improve opportunities for all citizens, including women and youth, and provide safety nets to protect the vulnerable against extreme deprivation and shocks. The social contract must also include adequate mechanisms to raise resources to support these policies, including a tax system that creates incentives for economic growth and promotes fairness. Regional income disparity is a major issue in Egypt, where wide disparities persist between Upper Egypt and Lower Egypt, and between rural and urban populations. The highest poverty rates are in Upper Egypt, where 44 percent of the rural 101 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES population was living in poverty in 2009, compared with the national average of about 25 percent. Egypt’s growth spurt in 2003–08, which helped reduce poverty at the national level, by and large bypassed rural areas and the south. The strategies of both the Bank and the government in dealing with this important issue, which was identified in the Bank’s diagnostics work, proved ineffective. Similar issues weakened the link between economic growth and poverty reduction in certain regions of Guatemala, Nigeria, Peru, and Romania, particularly remote places with inadequate connectivity and lack of access to social services. In Guatemala, geographic, ethnic, gender, and rural-urban disparities in social indicators are large. The incidence of child malnutrition—particularly stunting (height-for-age)—is the highest in Latin America and among the highest in the world. It is especially high among rural and indigenous groups, with a concentration in the northern and northwestern regions. The Bank and the government did not pay adequate attention to ethnic issues or the empowerment of indigenous communities during the evaluation period. In all case study countries, the Bank’s strategy formulation, which benefited from good-quality diagnostics, emphasized social inclusion and inclusive growth. Implementation of the strategy and the lending portfolio were slow to be fully calibrated to deal with the lack of inclusiveness of growth in an effective manner, however. Ending extreme poverty and spurring shared prosperity requires mechanisms that ensure that the poor are integrated into the growth process. Social development lies at the heart of meeting the unique, though heterogeneous demands of disadvantaged groups (including women, youth, ethnic minorities, and others). In many developing countries, the better-off have living standards that are similar to the average in the developed world; it is the poorer segments of the population that lag behind. Focusing on protecting the vulnerable from extreme deprivation, particularly irreversible human capital damage, is critical for equitable and sustainable poverty reduction. Bangladesh CONTEXT Bangladesh’s location and high population density results in high vulnerability to natural disasters, particularly floods and cyclones. Limited and poorly performing infrastructure, and a confrontational political environment, with associated 102 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES weaknesses in government capacity, exacerbated by high levels of corruption, have contributed to a challenging development environment. Given these challenges and the depth of poverty and famine Bangladesh faced in the 1970s, performance has far exceeded expectations. Economic growth has accelerated by about one percentage point in each decade since independence in 1971, averaging about six percent a year during the past 10 years. Growth has also been relatively stable, with lower fluctuations than many other rapidly growing low-income countries. Over the period 2003–13, gross domestic product (GDP) per capita grew by an average of 4.9 percent. Gross national income (GNI) per capita (in purchasing power parity) reached $1,940 per capita in 2011. Despite these achievements, Bangladesh remains home to about five percent of the world’s poor. Bangladesh dramatically lowered the share of people living on less than $1.25 per day, from 58.6 percent in 2000 to 43.3 percent in 2010—a rate that was 60 percent faster than the rate of poverty reduction in the rest of the developing world, excluding China. As the share of people living on less than $1.25 per day is about 40 percent, a focus on reducing extreme poverty is at the same time a focus on promoting shared prosperity for the bottom 40 percent of the population. The two World Bank Group goals thus come together. Bangladesh has sharply improved its social indicators. It is expected to partially achieve the Millennium Development Goals (MDGs) for poverty reduction and primary school enrollment, and it is on track to achieve the MDGs for gender parity in education; child mortality; maternal health; and HIV/AIDS, malaria, and tuberculosis. The contributors to these achievements include Bangladesh’s attention to health outcomes, elementary education, family planning, and gender equality (especially in education and workforce participation), all supported by (mostly female) grassroots workers and organizers mobilized by the government and leading world nongovernmental organizations (NGOs), such as BRAC, and a vast network of microfinance institutions led by the Grameen Bank. The World Bank is Bangladesh’s single-largest development partner, allocating 23 percent of all IDA resources disbursed between 1971 and 2012 to Bangladesh. During the period 2006–14, the Bank committed $8.8 billion in IDA funds to 62 projects in Bangladesh. POVERTY DATA Poverty data in Bangladesh are generally of high quality and have improved over time. The Household Income Expenditure Survey (HIES) data provide robust estimates of both extreme and general poverty. Although it is conducted only every 103 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES five years, it appears to be timely for capturing the impact of structural changes, though not for documenting the impact of short-term crises or immediate responses to policies or programs. The HIES is the main data set used for poverty headcount estimates and for linking poverty incidence with other characteristics. Because it has used a comparable set of questions and sample frame in the last three rounds (2000, 2005, and 2010), it has facilitated a robust framework for analyzing both the incidence and the drivers of poverty over this period. Nevertheless several World Bank staff, researchers, and policy analysts noted that the five-year gap between surveys was a long period in relation to monitoring the impact on poverty of programs and short-term interventions. Although past efforts to design and implement a reliable but more frequent poverty monitoring instruments were not successful, there is strong demand for greater frequency in poverty headcount estimates. Therefore, there is strong demand for interim poverty data. The Bangladesh Bureau of Statistics (BBS) is pursuing one approach to linking policy targeting and monitoring with the implementation of a poverty database project. It will develop a proxy means test score that can be applied to every household and used for more effective policy targeting. The Bank is supporting this activity with a $40 million component to its Social Safety Nets (SSNs) Project. The BBS also conducts a number of sample surveys every year, either as a regular activity or on an ad hoc basis. These surveys include the Agriculture Crop Production Survey and the Survey of Current Industrial Production. At different intervals it conducts important sample surveys regularly, such as the HIES and the Labor Force Survey. Interviews and other sources of information suggest that stakeholders do not regard many of these other surveys as having the same robustness as the HIES. Overall, the Bank has been involved with the development of the HIES since the 1990s. It has played a significant role in building capacity in Bangladesh to carry out the surveys, disseminate results, and prepare poverty maps. Since 2000, improvements in survey questionnaires have allowed extensive longitudinal analysis, and the data and other basic survey results have been made widely available. Nevertheless, the timing of the survey is not well coordinated with the country’s planning cycle, and important poverty-related data (Labor Force Surveys) receive less attention and are regarded as less reliable. DIAGNOSTICS The Bank produced comprehensive poverty assessments following the HIES rounds in 2000, 2005, and 2010. Each of these assessments exploited the available data and provided a rich profile of the poor and of key drivers of poverty and poverty 104 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES reduction in Bangladesh. Because of the comparability of survey design over the decade, successive poverty assessments were able to look in detail at the dynamics of the incidence and determinants of poverty and provide a rigorous analysis of interactions between them. Although the focus of each poverty assessment varied, they all examined the major programs affecting poverty dynamics, with a focus on improving targeting of government programs and strengthening governance to reduce leakages in safety nets. As the Bank had long supported health and education through multidonor, sector-wide approaches (SWAps), the poverty assessments focused more on the institutional framework governing social protection and support for the private sector–led growth process in Bangladesh. The diagnostic work of the poverty assessments included close collaboration with BBS and benefited from the support of other development partners, such as the UK Department for International Development (DFID) for specialized funding or detailed analytic inputs. Poverty analysis and diagnostics are a major focus of all stakeholders in Bangladesh, many of whom, such as the Center for Policy Dialogue, have produced significant independent pieces of analysis. United Nations (UN) agencies are also involved in poverty diagnostics through such work as the Bangladesh Human Development Report and the Child Equity Atlas of the United Nation’s Children Fund (UNICEF). BRAC, the world’s largest NGO, with an extensive network of grassroots workers in Bangladesh, is also a major source of situational, qualitative, and quantitative poverty diagnostic work. Although the recommendations of these poverty diagnostic reports have tended to become less specific over the past decade, their policy impact has probably increased over time, thanks mainly to the consistency of policy messages across the poverty assessments and increased activity by the Bank in social protection. These diagnostic reports have tended to coincide with the Poverty Reduction Strategy Paper (PRSP) process since 2002 and have facilitated considerable interactions between the government, the Bank, other donors, and local think tanks. Although the Bank has continued to support Bangladesh with diagnostics and analytical reports related to human development, safety nets, disaster preparedness, and poverty-linked rural investments, it did not prepare a Public Expenditure Review (PER) because of the lack of demand by the government (since the Bank’s policy-based lending halted as a result of major policy differences between the government and the Bank). Key stakeholders and development partners regretted the absence of a PER by the Bank. STRATEGY FORMULATION, IMPLEMENTATION, AND FEEDBACK LOOPS Overall, the Bank’s role has been central to the analysis of income poverty in Bangladesh. Stakeholders defer to the Bank’s diagnostic capacity even though perspectives on underlying causes and interrelationships vary. The links between 105 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES data, diagnostics, and strategy formulation in Bangladesh have grown stronger over time. In the case of the FY2006–09 and FY2011–14 CASs, the relevant PRSPs were consistent with the data and diagnostics of the relevant poverty assessments. These strategies emphasized private sector led growth, human development, and, more recently, vulnerability with links to climate change and disaster preparedness and management. The formulated strategies were consistent with the relevant poverty diagnostics. Regarding CAS implementation, the poverty focus of the Bank program in Bangladesh benefited from a diverse and strong analytic base, especially before and during the initial years of the FY2006–09 CAS. However, the analytic base weakened somewhat as linkages with policy-based lending and government dialogue weakened. By the time of the FY2011–14 CAS, a complicated policy dialogue with a new government put policy-based lending on the back burner, although the strategy did include a broad sectoral mix of investments, such as assistance to the transformative $1.2 billion Padma Bridge Project, which clearly addressed the issues of regional integration that the 2008 poverty assessment had highlighted. However, because of allegations of intended corruption involving senior government officials, the Bank cancelled the bridge project. It also withdrew from all other infrastructure and energy projects, with the potential Poverty Reduction Support Credits (PRSCs) an additional casualty. The Bangladesh program is characterized by a large number of follow-on and additional financing operations. Over the period FY2006–13, almost half (44 percent) of all IDA operations were follow-on projects or additional financing. These operations included sequences of Development Support Credits and support to the health and education SWAps. Another follow-on project that the CASs identified as building on good performance during the previous period was the Reaching out of School Children Project, which was initially funded as a $50 million pilot. It received additional financing of $30 million and was followed by a second $130 million project. A notable example of building on a successful intervention is the support to employment generation safety nets, which began as technical support to the 100–day Employment Guarantee Program of the government late in the FY2006–09 CAS period, followed by $150 million project support early in the FY2011–14 CAS period. The FY2011–14 CAS indicated that successful experience with that project would trigger a scaled-up, follow-on operation, which materialized in FY2012, with the $500 million SSNs Project for the poorest. This sequence is a good example of the effective interaction of analytic, technical, and financial support. There is little direct evidence on the sustainability of Bank-supported activities implemented during the FY2006–09 and FY2011–14 CAS periods, as most activities 106 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES linked to poverty reduction received continuing support over multiple periods. Bangladesh’s low revenue generation suggests a lack of fiscal space. Most observers do not regard finance as the binding constraint, however, they regard capacity and institutional issues as more significant, and pointing out that in spite of low revenues, Bangladesh’s investment program traditionally underspends. However, Bank staff cite sustainability concerns as a major program risk, and the frequency of additional financing and follow-on operations suggest that the lack of fiscal space is an important constraint. The Bank’s country strategies in the FY2006–09 and FY2011–14 CASs were well aligned with Bangladesh’s national poverty reduction strategy, as articulated in the relevant PRSPs. The CASs supported all four pillars of the PRSPs (i.e., macroeconomic stability; sector growth strategies, including infrastructure; safety nets; and human development). They also focused on governance, which the PRSPs treated as a cross-cutting issue. PRSP I (2005) had a particularly thorough discussion of poverty evidence and diagnostics, which put the Bank focus on supporting the reduction of income poverty in a broader context. Of course, ownership is the key to actually implementing a national strategy. Many observers question the extent of PRSP ownership in Bangladesh. Ownership certainly varied over the three regimes holding power during the past 10 years, with the technocratic caretaker government most committed to PRSP policies and programs, particularly on the cross-cutting governance issues. The current government has reverted to national five-year plans as the mechanism to articulate its national vision and strategy. The Sixth Five-Year Plan (2011–15) is a much broader document than the PRSP, but it still retains a strong focus on poverty reduction, including an analysis of poverty diagnostics from the 2010 HIES. Noting that a new five-year plan was under preparation to take effect from 2016, the CAS Progress Report (CASPR) extended the Bank program by one year to coincide better with the Bangladesh planning cycle. However the disconnect between the next round of the HIES and the government and Bank’s planning cycles will continue, as the 2015 HIES results will not be available during the plan preparation period. The sequence from data to diagnostics to strategy formulation and implementation varies across sectors. In some cases, feedback loops are strong. In the Social Protection Program, for example, a history of good analytic work feeding into the poverty assessments contributed to a well-prioritized set of operational recommendations focusing on efficiency, efficacy, and targeting. This then formed the basis for dialogue with the key implementing agencies and project interventions when circumstances were ripe. Feedback loops are also relatively strong in education. A focus on improved targeting of education stipends to increase girls’ school attendance led to the realization that attendance rates for boys were falling, 107 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES and stipend programs were adjusted to tackle the problem. Poverty-focused feedback loops have been weaker in infrastructure lending, although there is growing interest in these sectors (notably rural roads and rural electrification) in drawing on impact evaluations to improve project design and ultimate service use. Several elements worked reasonably well in Bangladesh but could have worked better. Bank support to the various survey rounds (BBS and HIES) over a 20–year period created a strong database and effective partnership, which could have expanded earlier to include other key surveys (such as labor force surveys) and more general strategy formulation and capacity building at BBS. The Bank could have supported broader and easier access to BBS data and devoted more attention to the challenge of poverty monitoring between the five-year HIES rounds. Poverty assessments closely followed the HIES rounds with strong analytic and research links and consistent messages; a programmatic approach to them might have provided more timely inputs into strategy formulation. The alignment of the CASs with the Bangladesh PRSPs facilitated a strong poverty focus, but the results chain could have been more explicit in sectors having a more indirect influence on poverty reduction. If the Bangladesh CASs had presented and analyzed poverty reduction more rigorously, the poverty focus would likely have been more clearly linked to the growth objective, with the probability of stronger results. Country management and the country team succeeded in preserving a poverty focus to the Bank program following the cancellation of the Padma Bridge Project, but this episode suggests that the Bank’s reputational risk assessment could have been more effectively managed so as not to sacrifice impact on key development outcomes and poverty reduction. SUMMARY ASSESSMENT In terms of the Bank’s contributions to Bangladesh’s goal of poverty reduction, many aspects of the Bank’s program over the period of this evaluation worked reasonably well, but they could have worked better. Bank support to the various survey rounds and the HIES over a 20–year period created a strong database and effective partnership that could have expanded earlier to include other key surveys (such as labor force surveys) and more general strategy formulation and capacity building at the BBS. The Bank could have supported broader and easier access to BBS data more strongly and devoted more attention to the challenge of poverty monitoring in between the five-year HIES rounds. Poverty assessments closely followed the HIES rounds, with strong analytic and research links and consistent messages. A programmatic approach to poverty assessments could have provided more timely inputs for strategy formulation. The alignment of the CASs with the Bangladesh PRSPs facilitated a strong poverty focus, but the results chain could 108 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES have been more explicit in sectors with a more indirect influence on poverty reduction. If the Bangladesh CASs had applied a poverty reduction more rigorously, the poverty focus would likely have been more clearly linked to the growth objective, with the probability of stronger results. Country management and the country team succeeded in preserving a poverty focus to the Bank program following the cancellation of the Padma Bridge project, but this episode suggests that the Bank’s reputational risk assessment could have been more effectively managed so as not to reduce the impact on key development outcomes and poverty reduction. Arab Republic of Egypt CONTEXT Before the political changes began taking shape in 2009–10, Egypt’s economy was strong in many respects. The economy was growing at a fairly rapid pace, averaging five percent a year between 2000 and 2010. Starting in 2004, the government pursued wide-ranging structural reforms—including tariff reduction, privatization of state- owned enterprises, and reduction in regulation of the private sector—that aimed to improve the business environment and make Egypt’s economy more competitive. During 2003–13, GDP per capita growth of Egypt averaged about 1.7 percent a year, and its GNI per capita reached $6,120 in 2011. However, behind strong growth and significant capital inflows during the first decade of the 2000s, Egypt’s economy faced a number of vulnerabilities. Average Egyptians saw little immediate benefit from economic reforms. Growth had done little to reduce persistently high unemployment, which averaged about 10 percent during the 2000s (with youth unemployment approaching 40 percent in more recent years). Poverty headcount, as measured by the national poverty line, increased from 19.6 percent in 2005 to 21.6 percent in 2009 and to 25.2 percent in 2011 (Source: WDI). The revolution in Egypt unraveled the conditions underpinning growth in the 2000s and brought potential vulnerabilities to the forefront. Economic growth has fallen to about two percent a year, and unemployment has continued to rise. Egypt made good progress toward achieving several non-income MDGs since 2000. In the past decade (between 2000 and the latest date for which data is available), the average primary school enrollment rate rose from about 85 percent to about 93.8 percent, the ratio of girls to boys in primary schools rose from 92 percent to 97 percent; the under-five mortality rate fell from 45 percent to 21 percent; and the 109 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES maternal mortality ratio fell from 75 to 50 per 100,000 live births. There were also improvements in access to safe water and to sanitation facilities. However, according to diagnostic reports by the Bank, these improvements in averages mask substantial regional and income disparity in progress toward the MDGs. Wide disparities persist between Upper Egypt (the south) and Lower Egypt (the north) and between rural and urban populations. The highest poverty is concentrated in Upper Egypt, where 43.7 percent of the rural population was living in poverty in 2009 and where 95 percent of Egypt’s poorest villages are located. Although only a little more than half of the population lives in rural areas, more than 78 percent of the poor and 80 percent of the extreme poor live there. These income disparities are reinforced by gaps in social indicators: virtually all health indicators and literacy rates are worse in Upper Egypt than in Lower Egypt and worse in rural areas than in urban areas. Illiteracy rates among young women in Upper Egypt are twice the rates of their male counterparts. Despite the rise in poverty in recent years and large disparities across regions and income groups, the Gini coefficient for Egypt (0.307) is only among the lowest in developing countries. Moreover, it shows a decline over the past decade, which contradicts with the perceived level of inequality by Egyptian experts. This likely points to underestimation and possible data quality issues. Some recent estimates indicate that if the data corrected for under-reported or unreported top incomes, the estimated Gini coefficient may rise significantly. POVERTY DATA The poverty and inequality issues in Egypt are tied to the political discourse used by successive governments. The Central Agency for Public Mobilization and Statistics (CAPMAS) is not an independent statistical agency and this is an important issue. International organizations, including the World Bank, and the public do not have access to the full dataset. Full access to the raw data is controlled through “trusted” consultants, and only relatively small samples (typically 20 percent) are made public. This lack of access compounds the observation that there is a concentration of households at relatively low levels of consumption, and numbers are likely to vary considerably with slight variations in the poverty lines. The HIES estimates released by CAPMAS and repeated in the 2011 World Bank poverty assessment suggest that, during the later years of the evaluation period, there was no substantial increase in poverty or inequality (poverty remained virtually unchanged between 2004–05 and 2008–09 in urban areas and increased slightly in rural areas, and Gini coefficients in both urban and rural areas declined). However, data from nutritional surveys suggest a significant increase in child 110 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES wasting and stunting since 2003. Moreover, recent estimates released by CAPMAS show a steady increase in poverty, with absolute numbers doubling between 2004 and 2014. The HIES data most likely miss the tails of the size distribution and are less able to capture nonwage income. Consequently, conclusions concerning inequality using the HIES should generally be treated with caution, as several researchers have pointed out. The troubling independent anthropometric data 3—especially the trends on the deterioration of standards since the early 2000s—should have been available to Bank staff. These data—and the clear increase in the number of the extreme poor—should have tempered the story on no increase in poverty or inequality. As pointed out in the discussion during the University of Cairo launch of the recent Bank monograph on poverty and inequality, the CAPMAS data likely underestimate both tails of the size distribution and are unable to capture the extent of nonwage income, particularly from financial assets. There is clearly a distrust by some academics and economic experts of the poverty and income inequality estimates presented in the Bank report, especially given the perception that disparities are clearly visible but not picked up by the incomplete data. More troubling from a policy perspective is the Bank’s observation that there is limited mobility among Egyptian households. This observation is not consistent with the perception that mobility is high wherever there is an expectation of employment opportunities, both in urban areas in Egypt and abroad. Indeed, some of the informality in housing as well as the labor market reflects mobility of some but not all members of extended- family households. Anthropological work in Egypt suggests that women seldom leave their family homes, as it is difficult to establish property rights in new locations. Moreover, some family members are needed to work on small farms and maintain their homes. Consequently, the structure of formal households remains remarkably stable, leading some observers to believe that there is relatively little mobility in Egypt. Yet men typically migrate to find employment, either seasonal or longer term. In addition to the HIES, specific village-level studies have established that many women and children in Upper Egypt live in abject conditions. These vulnerable groups may also be affected by a shock in urban areas that affects informal employment, which may also affect the income of the extended household, particularly the build-up of assets and remittances sent to rural family members. The Bank’s recent work on the labor market, however, suggests that there is considerable informality and vulnerability in Egypt, including in the “large-scale or established sector,” as is seen in other middle-income countries, such as Mexico. 111 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES Much of this type of informal labor force is urban-based and reflects firms’ concealment of labor use, outputs, and profits. Informal workers are also likely to be most affected by relative price changes, including changes in the prices of food and energy. Reducing inequality is important to ensure that the benefits of growth are not captured by the elites, fanning the flames of resentment in the aftermath of the Arab Spring. Given the extent of informality and likely concealment by firms in the formal sectors, it may be necessary to splice the HIES data, reflecting the middle ranges of the size distribution, together with information from other data sources. Thus, a possible reconciliation of the phenomenon of evident poverty in Upper Egypt with the new appreciation of informality and vulnerability is likely to be important, as it will also guide the design of policy responses. Further work on diagnostics is needed, including a careful juxtaposition of the data from the HIES and the labor force and other surveys, and sophisticated econometric assessments. DIAGNOSTICS Substantially less AAA were undertaken in Egypt than proposed in the various strategy documents. Notably absent were the core diagnostics of a Public Expenditure Review and a social and structural review. Consequently, it was convenient that poverty assessments suggested the incidence was low and stable and concentrated in the rural areas of Upper Egypt, and only modest increases had occurred in poverty based on national poverty line or inequality during the period of 2005–2009. Other indicators of poverty, basic needs, and informality suggested that the situation was not so positive. By missing out on the increasing vulnerability of informal migrants in urban areas, the CAPMAS estimates—although correctly identifying poor households in Upper Egypt—may not have provided the government the needed information for policy making. The recent release of new data for 2012–13 and previous years by CAPMAS cast further doubts on the Bank’s 2011 poverty assessment and raises further questions concerning a recent Bank report on income inequality. It suggests a continuous increase in poverty since the early 2000s, with absolute numbers doubling between 2004 and 2012–13, and the proportion of the population below the poverty threshold increasing from 16.7 percent in 1999–2000 to 26.3 percent in 2012–13. The big question remains whether the numbers accurately reflect the number of poor inhabitants in metropolitan area slums. A conundrum in the CAPMAS data is that the percentage of people living in extreme poverty—people spending less than Egyptian Pound LE 214 a month (or LE 2,570 annually)—declined after 2008. The proportion fell from 6.1 percent in 2008–09 112 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES to 4.8 percent in 2010–11 and 4.4 percent in 2012–13. Although these figures are higher than the 1999–2000 figure of 2.9 percent, the implications need to be thought through and investigated further. The recent jobs and informality report by the Bank is the second serious analytical report that breaks new ground in an innovative manner. However, like the equally innovative and interesting economic geography report by the Bank in 2012, it relied on an unsatisfactory overall assessment of the governance framework that basically assumes that the modernization of the public finance management (PFM) and tax agendas are well in hand. But perhaps the most serious drawback of both reports is the assumption that mobility in Egypt is low, which is not the case, particularly given the large outflow of labor to the other parts of the Middle East and North Africa Region. One of the most important contributions of the Bank’s report on jobs and informality (2014) is that it introduces realism on labor market conditions. As in many other middle-income developing countries, like Mexico, informality covers more than just people selling cigarettes on street corners. Large, formal sector firms often hire informal workers: the 51 percent share of informal workers in formal firms is remarkable. The report points to the complex nature of need and vulnerability—and the prospect that a broader approach to long-term poverty reduction may be needed than providing cash transfers to the population in Upper Egypt. The economic geography report (2012) provides an analytical framework that could be recalibrated by tweaking the assumption of limited internal migration. It reveals a more complex and diverse nature of vulnerability and need than the assumed concentration in Upper Egypt and a much more difficult PFM and governance environment than assumed in the report. STRATEGY FORMULATION, IMPLEMENTATION, AND FEEDBACK LOOPS The Bank’s Board discussed the last CAS for Egypt in May 2005. In July 2008 the Board discussed Egypt’s CASPR, and the then-current CAS was extended to the end of FY2009. An Interim Strategy Note for Egypt, covering the period June 2012– December 2013, was presented to the Board in May 2012. The work on a new CPS, which started in 2010, was suspended in 2011 because of the political upheaval in Egypt. The new CPS will cover FY2015–18. The CAS Completion Report (CASCR) for the FY2002–04 CAS noted that although the overarching objective was to reduce poverty and unemployment, poverty may have increased. Substantially less AAA was undertaken than proposed, and notably absent were the core diagnostics of a PER and a social and structural review. It also noted that with IBRD terms, the government was able to borrow more cheaply, 113 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES including from the U.S. Agency for International Development (USAID) for an information infrastructure project. Budgets were not consolidated, and megaprojects, wages, and subsidies led to a spiraling deficit. Consequently, the Bank delivered a small portion of the economic and sector work (ESW) program and a large loan (Cairo airport) to meet its lending target. A similar pattern was repeated with the FY2006–09 CAS. Despite the clear statement of the poverty focus in the well-argued FY2006–09 CAS, the subsequent choice of projects suggests that any linkage with poverty is purely coincidental. Attempts to design and implement development policy loans (DPLs) have foundered on the issue of overall macroeconomic imbalances and the absence of an agreement between the Egyptian authorities and the International Monetary Fund (IMF). The spiraling fiscal deficit is caused in large part by the system of administered prices, especially for energy products. As shown in work by both the IMF and the Bank, the bulk of the benefits of such policies now accrue to the relatively well-off and the middle classes, especially in urban areas. This evidence lends support to the public perception that inequalities have increased. Until the recent election, governments lacked either the political legitimacy or adequate appreciation of the situation to take corrective measures. Typically, inaction is justified on the grounds that price adjustments would hurt the poor—but the reality is that powerful middle-class interests are likely to be affected. A question that has been asked in Egypt and within the Bank is whether conditional cash transfers (CCTs) could be used as part of a strategy of energy price rationalization, as often recommended by both the IMF and the Bank. Clearly, CCTs have a role to play in empowering rural women and encouraging girls to go to school, as part of a permanent social safety net for the poorest groups in society. There has been considerable research in Egypt on this issue. However, these groups are distinct from those who have to adjust as a result of the relative price changes. People affected by price changes are not just the poorest; CCTs could cause labor market distortions and become an entitlement that cannot easily be removed; diminish the intended reduction in incentives to adjust spending in response to higher prices; and with weak PFM mechanisms are subject to leakage and abuse. In the absence of a PER, the likelihood of the government endorsing the tax and PFM infrastructure (a Treasury Single Account [TSA]) appeared unlikely, although the overall fiscal fragility was recognized. “Silos” among donors do not help. The IMF seems to have principal responsibility for advice on tax policy, including property tax, and budget and treasury design (the Government Finance Statistics Manual framework and TSA). However, USAID has been assisting with implementation (of the Government Financial Management Information System 114 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES [GFMIS] and tax administration). The apparent failure of the GFMIS and the inability to implement a TSA could have serious consequences for the Bank’s support for a poverty reduction strategy. Similarly, it is becoming increasingly clear that an effective system of local property taxation has to be linked to local service delivery and access to credit for local infrastructure. These are not issues with which the IMF typically deals. Regulatory costs and bureaucratic hurdles are high in Egypt. As described in a 2012 World Bank report, 120 industrial zones and numerous cities have been developed in several phases. All were established on public land, with relatively poor linkages and connectivity and considerable regulation. Few have been successful, and the incentives for different players (firms and workers) are poorly aligned. Any new initiative to create growth hubs—say, along the Suez corridor—would have to simplify regulations, establish an incentive compatible tax regime, and provide workers with adequate housing and services as well as the infrastructure needed for stepped-up investments. The quality of poverty data in Egypt during the period of this evaluation is questionable. The full sample is made available only to select groups of consultants and researchers favored by the government, and it is not possible to assess the quality of the full set of data. The data that are publicly available do not seem to capture the new type of poverty that has been building up around the major metropolitan areas and in major shanty towns, thus leading to a serious underestimation of the extent of poverty in Egypt. Although the quality of the Bank diagnostics has been generally good and the topics covered relevant, they may have been based on questionable data. In addition, the government has not allowed the Bank to prepare important analytical pieces, such as PERs or Public Investment Reviews, which are highly relevant to any government’s poverty reduction strategy formulation and related policy design. The strategy formulated by the Bank over the years attempted to include some focus on poverty, but the government in power during most of the evaluation period (2004 to February 2011) did not wish to borrow from the Bank for supporting its development programs in the social sectors. Moreover, implementation of the strategy (in terms of the Bank’s portfolio of projects and program) deviated from the agreed plans in the various strategy documents, resulting in an even weaker focus on poverty during the implementation phase. During much of the period covered by this evaluation, the Bank tried to engage and maintain a working relationship with a reluctant but strategically important client who was not committed to the goal of poverty reduction and did not want to borrow from the Bank for the purpose of poverty reduction or developing its social 115 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES sectors. Despite these difficulties, the Bank made a few successful attempts (mainly through analytical reports) to highlight the daunting challenge of poverty, including its strong regional dimension. Its efforts did not result in a strategic shift toward a greater focus on poverty and inclusion, however, until after the revolution of 2010– 11. It is expected that the newly elected government of Egypt will pay greater attention to poverty and inclusion issues than previous governments ones did. SUMMARY ASSESSMENT The quality of poverty data in Egypt during the period of this evaluation is questionable. The full sample was made available only to select groups of consultants and researchers favored by the government, and the data that were publicly available do not seem to accurately capture the new type of poverty that has been building up around major metropolitan areas and in major shanty towns, thus leading to possible serious underestimation of the extent of poverty in Egypt. Although the quality of Bank diagnostics has been generally good and the topics covered have been relevant, the Bank may have been using questionable data. The government has not allowed the Bank to prepare important analytical pieces, such as PERs, which are highly relevant to any government’s poverty reduction strategy formulation and related policy design. Although the strategy formulated by the Bank over the years has attempted to include some focus on poverty, previous governments in Egypt during the evaluation period had generally different priorities than focusing on poverty. In general, the governmentthat was in power during most of the evaluation period (2004 to February 2011), did not seem to focus on poverty and related issues. Moreover, implementation of the strategy underpinning the Bank’s portfolio of projects and program deviated from the agreed plans in the various strategy documents, resulting in an even weaker focus on poverty during the implementation phase of strategy. During much of the period covered by this evaluation, the Bank was trying to maintain a working relationship and be responsive to the requests of a reluctant but strategically important client who had different priorities than poverty reduction and did not want to borrow from the Bank for the purpose of reducing poverty or developing its social sectors. Despite these difficulties, the Bank made a few successful attempts (mainly through its analytical reports) to highlight the daunting challenge of poverty, including its strong regional dimension, facing Egypt. Its efforts never resulted in a strategic shift toward a greater focus on poverty and inclusion until after the revolution of 2010–11, however. 116 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES Guatemala CONTEXT Guatemala grew by about 4.5 percent a year until it was hit by the global financial crisis in 2008–09 and a serious of natural disasters in 2010–11. Average annual growth in 2003–13 was about 3.5 percent (1 percent in per capita terms). GNI per capita was about $4,760 in 2011. Although Guatemala’s economy is the largest in Central America and its per capita GDP is close to the regional average, it has one of the highest levels of poverty (especially in rural and indigenous areas) and income inequality in the region. Moreover, poverty indicators have changed very little over the past decade in spite of positive average economic growth. Although some modest gains appear to have been made in reducing extreme poverty, the lot of the nonindigent poor has not improved and may have even worsened. The overall national poverty headcount rate declined from 56 percent in 2000 to 51 percent in 2006, but it rose to 54 percent in 2011, yielding an overall reduction of only 2.5 percentage points over the decade. Moreover, the relation between poverty indicators and economic growth appears to be very weak. Guatemala’s social indicators exhibit a somewhat steadier picture of gradual progress over time, although huge gaps remain across ethnic, gender, geographic, and rural and urban divides. Most worrisome is the persistence of child malnutrition indicators: with about 50 percent of its children malnourished, Guatemala continues to lag behind some of the poorest countries in world in this indicator. On the positive side, the national household surveys suggest that there has been a substantial decline in inequality between 2003 and 2011. However, questions remain about the accuracy of the figures on distribution based on these household surveys. POVERTY DATA Starting in the late 1990s, the Guatemalan authorities, with Bank and other donor support through the Mejoramiento de las Encuestas de Hogares y la Medición de Condiciones de Vida program, pushed to strengthen the main public institutions in charge of carrying out living standards surveys and generating reliable poverty- relevant data. Living Standard Measurement Surveys (LSMSs) were carried out in 2000, 2006, and 2011. During this period, Guatemala also carried out two National Maternal and Child Health Surveys in 2002 and 2008/09, and annual labor surveys that complement the findings from the LSMSs. This capacity-building effort helped create a critical mass of technical expertise in the National Statistics Institute and the government’s planning secretariat, resulting in a vast improvement in the quality of data, greater transparency in the management of data, a broad consensus on how to measure poverty, and widespread consciousness of the importance of having reliable, objective poverty data. In addition to the technical and financial support 117 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES provided, the Bank played a critical role in helping achieve public consensus on the measurement of poverty, based on objective technical criteria instead of political ones. The advances made in strengthening Guatemala’s capacity to produce high-quality poverty statistics may not be sustainable. Despite their tremendous success, the technical assistance programs that supported the production of good poverty statistics expired at the beginning of this decade, and the prospects of future support with resources from the national budget remain uncertain. The Bank may have underestimated the need for continued institution-building support to guarantee sustainability, as there are already signs that the institutional capacity that was built up over the previous decade has begun to decline, compromising the ability to continue generating good poverty statistics in the future. Although providing solid analytical work, the poverty assessments for Guatemala were not uniformly consistent in proposing actionable recommendations and assessing institutional capacity-building needs. The 2003 poverty assessment provided a number of detailed recommendations for addressing poverty in Guatemala; the subsequent two assessments limited themselves to broad strategic recommendations. More detailed recommendations for addressing labor and product market rigidities might have been useful. Even so, the poverty maps and proxy-means testing formulas developed as part of the Bank’s diagnostic work have played an important role in the government’s strategic planning processes and in the design of public programs. These instruments either did not exist in Guatemala or existed only in very rudimentary form until the 2003 poverty assessment was prepared. The Bank’s poverty diagnostic work played an important role in bringing the topic of poverty reduction into the national political discourse from a technical perspective. Before this work, this topic was considered too sensitive to discuss in view of the country’s historical and political circumstances. DIAGNOSTICS The Bank’s poverty diagnostic work has improved the government’s and donors’ understanding of the main drivers of poverty by identifying the contribution made by different sources of income on the reduction of poverty. It also points toward structural rigidities that may have prevented faster poverty reduction. The 2005 Country Economic Memorandum (CEM) found that two of the most important factors responsible for this relatively slow growth performance are low education attainment levels and poor public infrastructure, which in turn reflect low public sector spending levels. The Bank’s PERs have raised questions about the quality of public expenditures, particularly in the social sectors, noting problems related to geographic targeting and technical efficiency. In particular, the 2013 PER found no 118 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES clear relationship between the level of public spending and student achievement in education. Health outcomes also did not appear to be strongly linked to health expenditures. The general absence of quantitative links between recommended measures and poverty reduction is a significant limitation of the Bank’s diagnostic work. Although existing diagnostic work makes a convincing case that certain interventions, such as measures leading to improvements in social indicators, contribute to poverty reduction, it provides little help in estimating the degree to which poverty rates may decline in response to certain sector outcomes and thus offers little guidance in choosing the appropriate balance between interventions that promote growth, improve social indicators, and redistribute incomes to reduce poverty in the most efficient manner. STRATEGY FORMULATION, IMPLEMENTATION, AND FEEDBACK LOOPS All three country strategies produced by the Bank since 2004 refer to poverty reduction as the underlying objective and draw on the findings from the latest poverty assessments, CEMs, and PERs to propose interventions designed to reach that objective. The Bank prepared three country strategies during this period: the FY05–08 CAS, the FY09–12 CPS, and the FY13–16 CPS. All recognize the importance of economic growth as a necessary condition for achieving faster poverty reduction and point toward low education attainment and poor productive infrastructure as key obstacles to faster growth. In addition, all three strategy documents recognize that unequal access to social and productive public services has been a major obstacle to poverty reduction, independent of the poverty-growth nexus. Accordingly, all three strategies propose measures to promote faster sustained economic growth and new operations designed to increase access to social services and public infrastructure services, with particular attention on groups that have traditionally been underserved. Although sharing these common elements, the three strategies exhibit differences in emphasis that respond to new developments and new insights gained from successive poverty diagnoses. The FY05–08 CAS focused mainly on DPL-supported actions meant to promote growth and competitiveness, coupled with investment loans to strengthen the education, health, and rural infrastructure sectors to provide better structural underpinnings for faster economic growth. In contrast, the FY09–12 CPS supported a program of conditional cash transfers, with the aim of reaching the pockets of poverty that the 2006 LSMS had revealed to be unresponsive to economic growth. The FY13–16 CPS focused on the need to address crime and violence as increasing threats to economic growth and poverty reduction in Guatemala and the need to strengthen institutional capabilities to improve the quality of public spending, in parallel with efforts to raise fiscal revenues further (as recommended in 119 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES the 2013 PER). The Bank strategies also adapted to different priorities across government administrations. The greater focus on growth-promoting reforms in the FY05–08 CAS and FY13–16 CPS also reflect the greater priority given to improved governance and competitiveness by the administrations in power, while the focus on conditional cash transfers in the FY09–12 CPS coincides with the administration in office at that time and its greater emphasis on social inclusiveness and reducing inequality. The FY09–12 and FY13–16 CPSs called for a larger share of lending in the form of fast-disbursing support, for reasons in part unrelated to poverty reduction. In the case of the FY09–12 CPS, more lending in the form of DPLs was called for to address fears of illiquidity triggered by the then emerging global financial crisis. The FY13– 16 CPS proposed to maintain a very DPL-intensive lending portfolio, less because of crisis-related fears and more because of increasing operational obstacles to the implementation of investment projects in Guatemala. A preference for DPLs over investment lending on account of the latter’s greater implementation difficulties, however, is difficult to justify on poverty grounds and ultimately becomes counterproductive if it threatens to undermine fiscal sustainability. Bank interventions proposed for Guatemala reflect the Bank’s priorities, as expressed in the results framework of each country strategy. The new lending operations proposed in each strategy document are closely aligned with the main recommendations that emerged from previous diagnostic work. In addition, the Bank proposed an extensive program of nonlending AAA activities to complement its poverty-related loan portfolio, anchored in periodic poverty assessments and poverty assessment updates that were included in the CAS and CPSs. The 2003 poverty assessment recommended actions in the areas of economic growth, education, health and nutrition, rural infrastructure, public sector management, and the targeting of interventions on the poor. The 2009 poverty assessment added the recommendation to strengthen the CCT program. Even though not all the proposed interventions have a direct poverty-reducing objective, each has a proposed loan or loan component that can be mapped to each of the main recommendations highlighted in the poverty assessments. Programmatic series of DPLs played a major role in all the CASs/CPSs prepared since 2004, accounting for at least half of the total proposed lending. The DPLs have been broadly justified as vehicles for supporting structural reforms to promote faster growth as the key underpinning for poverty reduction. Beyond that indirect link, only very small portions of the DPL-supported programs address poverty reduction. The only policy objective that is remotely linked to poverty reduction in the first series of DPLs (2005–07) was the raising of total social spending to at least 6 percent 120 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES of gross domestic product (GDP). Concerns about sustainability have influenced the operations included in the Bank’s country strategy to reduce poverty. Two developments in the composition of lending operations implemented between 2005 and 2012 are the hugely increased proportion of lending through fast- disbursing DPLs and the decline in lending devoted to poverty-related investment projects. The share of DPLs increased from 55 percent of total lending during the FY05–08 CAS period to 83 percent of total lending during the subsequent CPS period. The FY05–08 CAS was implemented during a period of favorable macroeconomic conditions, marked by modest economic growth and a stable external environment. As a result, the actual lending program turned out to be fairly close to the planned base-case program envisioned in the FY05–08 CAS. Although most of the lending and nonlending targets in the FY05–08 CAS were achieved, project implementation problems led to uneven achievement of the poverty-related development outcomes. Most lending and AAA outputs proposed in the FY05–08 CAS were delivered, and the bulk of the expected outcomes from Bank-supported interventions were achieved. As a result, the Independent Evaluation Group (IEG) rated achievement of development outcomes of the CAS overall, as well as the Bank’s contribution to those outcomes, as moderately satisfactory. The weakest performance was in the development outcomes associated with the most poverty- relevant projects. IEG rated these outcome as moderately unsatisfactory in its CASCR review of 2008. The FY09–12 CPS was implemented in much more difficult economic and political circumstances, resulting in significant program deviations. In contrast to the FY05– 08 CAS period, the CPS period of FY09–12 saw the global financial crisis, a food price crisis, rising drug-related violence, and several Guatemala-specific natural disasters in 2010. As a result, much of the proposed lending program was redirected toward other priorities. A major setback to Bank efforts to support poverty reduction was the cancelation of a project to support the government’s flagship CCT program. Despite considerable differences between the originally planned and implemented portfolio during this CPS period, however, the Bank’s performance was considered moderately satisfactory. The CPS Completion Report (CPSCR) rated both Bank performance and the achievement of results as moderately satisfactory. The CPS outcomes as redefined by the progress report were, for the most part, achieved or partially achieved. The project implementation problems that led to the cancelation of the CCT project appear to have worsened over time. These problems are largely attributable to an increased reluctance to borrow by a very conservative Congress, as well as to increasing institutional obstacles that have led to long delays in project 121 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES implementation. After the Bank approves a loan, it takes Guatemala’s Congress an extremely long time to ratify it. As a result, many projects need to be restructured by the time they become effective. Once they become effective, most projects face further implementation delays, brought on by a combination of weak institutions, overly centralized public management, complex procurement rules, inflexible administrative procedures and high staff turnover in government. The Bank has produced considerable evidence on poverty reduction in its programs and adjusted its programs in response to that evidence, but it does not target poverty indicators directly. Neither the CAS nor the two CPSs prepared since 2004 present poverty indicators in their results frameworks, and changes in poverty indicators are not targeted as part of the monitoring and evaluation (M&E) framework. Most Bank-supported operations reviewed for this task include M&E systems to assess project impact or contain provisions for building capacity to put such systems in place. Nevertheless, it is widely perceived that a culture of systematic evaluations still remains to be created across the public sector. The feedback loops characterizing World Bank operations in Guatemala—from data production to poverty diagnostics to country strategy and implementation—have not been strong in all areas. The Bank seems to have done an excellent job in supporting the generation of poverty data and in preparing poverty diagnoses over the past decade, so that lack of adequate data and diagnostics do not stand out as major obstacles in addressing poverty in Guatemala. Concerns have subsequently arisen about the sustainability of past capacity-building efforts in Guatemala’s statistics-producing institutions. These concerns were not pertinent for most of the period covered by this analysis; rather, they may become a problem for future operations. The CAS/CPSs and choice of planned lending and non- Lending operations appear generally well aligned with the poverty diagnostics in seeking to address the key bottlenecks to poverty reduction. A potential thematic disconnect between poverty diagnostics and the choice of programs and projects does not stand out as a major problem. The area in the feedback loop where the Bank’s support for antipoverty efforts appears to have been least successful during the past decade is country strategy implementation. The result has been overreliance on DPLs with limited links to poverty reduction. Although this lack of success was partly caused by unforeseen external circumstances, it also reflects both institutional impediments that paralyze the public sector’s capacity to function and an inadequate fiscal framework that undermines sustainability. 122 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES SUMMARY ASSESSMENT The feedback loops characterizing Bank operations in Guatemala—from data production to poverty diagnostics to country strategy formulation and implementation—have not been evenly strong in all areas. The Bank seems to have done an excellent job in supporting the generation of poverty data and in preparing poverty diagnoses over the past decade, so that lack of adequate data and diagnostics do not stand out as major obstacles to addressing poverty in Guatemala. Concerns have arisen about the sustainability of past capacity-building efforts in Guatemala’s statistics-producing institutions. These concerns were not pertinent for most of the period covered by this analysis but rather may become a problem for future operations. The CAS and CPSs and the choice of lending and nonlending operations appear generally well aligned with the poverty diagnostics; a thematic disconnect between poverty diagnostics and the choice of programs and projects does not stand out as a major problem. The one area in the feedback loop where the Bank’s support for anti-poverty efforts appears to have been least successful during the past decade is country strategy implementation. This difficult in project implementation has resulted in an overreliance on development policy loans (DPLs) with limited links to poverty reduction. Although this lack of success was caused partly by unforeseen external circumstances, it also reflects both institutional impediments that paralyze the public sector’s capacity to function and an inadequate fiscal framework, which undermines sustainability. Lao People’s Democratic Republic CONTEXT During the review period, the Lao People’s Democratic Republic enjoyed annual GDP growth of about 8 percent (5.7 percent in per capita terms), and the poverty headcount rate continued its long-term decline, falling from 33 percent in 2002/03 to 28 percent in 2007/08 and 22 percent in 2012/13. GNI per capita reached $2,580 in 2011. Progress was also achieved on primary education and basic health care (life expectancy, for example, increased from 46 years in 1970 to 67 years in 2011). The share of rural households with access to electricity expanded from 16 percent in 1995 to 45 percent in 2004 and about 80 percent more recently. However, there are real threats to maintaining macroeconomic stability and growth and therefore to achieving further reductions in poverty. Although primary school enrollment rose from 75 percent in 2000 to 96 percent in 2012 and the under-five mortality rate fell from 120 to 72 (per1000 live births), progress on achieving some of the other MDGs is inadequate. Growing fiscal deficits threaten macroeconomic stability in the near term, and the overly rapid expansion of foreign direct 123 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES investment in hydropower, mining, transport, and agricultural projects may produce negative social and environmental impacts. Significant disparities in economic growth and poverty reduction persist across the country’s diverse ethnic groups and geographical regions, and it appears likely that some of the nutrition, universal primary education, child mortality, maternal health, and environmental sustainability MDGs will not be met. POVERTY DATA The Bank country team has had adequate survey and administrative data of sufficient quality to carry out detailed poverty diagnostics and support the poverty reduction focus of the country program for the Lao People’s Democratic Republic. The five Lao PDR Expenditure and Consumption Surveys (LECS) (undertaken in 1992/93, 1997/98, 2002/03, 2007/08, and 2012/13) are nationally representative household surveys of consumption expenditure and a wide range of socioeconomic information. The Lao PDR Statistical Bureau has conducted all of the LECS surveys. The Bank and other development partners have provided extensive technical and financial support to the LECS surveys since 1991. Although there are some concerns about the quality, accessibility, and timeliness of the LECS survey data, the data have been sufficient and played a core role in the preparation of the Bank’s poverty diagnostics during the review period. The Bank and other development partners provided extensive technical and financial support over the review period to improve the quality, availability, and timeliness of administrative data on educational attainment, health outcomes, nutritional status, and other indicators. These additional data have allowed for deeper and more multidimensional poverty diagnostics work. DIAGNOSTICS The Bank’s poverty diagnostics work was of good quality and demonstrated good practice in most, but not all, areas. The good practice elements of the poverty diagnostics work include (i) collaborating and sharing with the government and development partners all aspects of the poverty diagnostic work; (ii) building Lao PDR capacity for gathering and analyzing poverty data; (iii) making full use of the available survey and administrative poverty data to provide a comprehensive and multidimensional poverty profile that assessed poverty trends over time, across Regions, and across social groups; (iv) assessing the key drivers of poverty reduction including growth, access to social services and basic infrastructure, agricultural productivity and rural development; (v) examining the poverty reduction impact of relevant government poverty reduction programs, funding, and sectoral policies; and (vi) encouraging deep country ownership of the poverty analytics and seeking full consistency with the government’s poverty reduction strategy. 124 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES The Bank team’s poverty diagnostic work was closely tailored to country specifics, such as the inclusion of an examination of the special concern of unexploded ordinance. It provided a good understanding of extreme poverty and the special concerns of poor women and upland ethnic minority groups, and it consolidated list of priority poverty reduction measures. The shortcomings of the work include the limited analysis of the government’s poverty reduction institutions and the limited integration of the results of available participatory and qualitative poverty assessments. Although the consolidated list of priority poverty reduction measures was not as specific, actionable, time-bound, or costed as it might have been, these weaknesses did not in any way thwart or limit the poverty reduction focus of the Bank’s country program. STRATEGY FORMULATION, IMPLEMENTATION, AND FEEDBACK LOOPS Poverty reduction is a central focus of the Bank’s country program in Lao PDR. The FY05–08 CAS and FY12–16 CPS fully reflect the Bank’s poverty diagnostics and have a sharp focus on poverty reduction. There is a near exact fit between the recommendations of the 2006 poverty assessment and the FY12–16 CPS; the one important exception is fully explained in the FY12–16 CPS and by the division of labor agreed to under the 2006 Vientiane Declaration on Aid Effectiveness. There is a favorable balance of analytical work, technical assistance, and project work, and the results chain is evident. The FY12–16 CPS does an excellent job of mapping all aspects of the country program into the government’s poverty reduction strategy and is fully consistent with the development partners harmonization explicated in the 2006 Vientiane Declaration. The key elements of the poverty focus of the Bank’s country program include (i) the lynchpin Nam Theun 2 Project (NT2) and the closely associated Poverty Reduction Support Operation (PRSO) projects; (ii) the Poverty Reduction Fund and Khammouane Development Projects; (iii) education and health projects; (iv) rural electrification, transport, and agricultural development projects; and (v) the extensive analytical work and technical assistance programs associated with these activities. The core poverty reduction results chain of the programs of AAA, technical assistance, and project activities in the FY05–08 CAS and FY12–16 CAS was clear and strong throughout the review period and very closely mapped to the government’s poverty reduction strategy. The poverty reduction results chain is sustaining strong economic growth through improved economic management, regional integration, private sector development and competitiveness, and natural resource management and development. It is improving social outcomes and reducing vulnerability through strengthened public financial management and public service delivery, improved infrastructure services in transport and energy, 125 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES and targeted poverty reduction programs. Adequate quality poverty data and extensive and good-quality poverty diagnostics underpinned this results chain throughout the review period. The poverty diagnostics work showed that extreme poverty, ethnic minorities, and other disadvantaged groups are concentrated largely in first-priority districts—those districts with high poverty headcount and receiving high priority for antipoverty programs. As nearly all of the Bank’s support for basic infrastructure, education, health, and other social services; poor area development programs; and natural resource development and management activities is concentrated in these districts and some other lagging regions, much of the country program has disproportionately benefitted the extreme poor, ethnic minorities, and other disadvantaged groups. The FY05–08 CAS takes explicit note of the major risks of possible loss of government commitment to continued policy reform and weak implementation capacity as well as poor performance of the NT2 program. The FY12–16 CPS goes a step farther by stating that the FY05–08 CAS “included some bold and potentially risky activities, most obviously the NT2 project, but also commitments to budget support operations, community driven development, and policy reforms in a low capacity environment.” The Bank’s country program did indeed embark on a high- risk and high-reward poverty reduction path during the review period, and the FY12–16 CPS concluded that “performance was impressive.” This favorable performance in very large part reflected the fact that the country program was designed to mitigate these risks through a strong focus on capacity building and the concentration of much of the Bank’s resources on the core NT2 program. The FY12– 16 CPS correctly concluded that “this success in turn demands raising the level of engagement” during FY12–16. The risks to the country program are now at least as great as they were at the outset of the FY05–08 CAS. Looking forward, there are signs that the growing fiscal deficit and weak management of the rapidly expanding investment in, and land concessions for, hydropower, mining, and agricultural activities have the potential to undermine the achievements made during the review period. Interviews and the review of documents by the IEG team provide additional details on these risks to the poverty reduction focus of the country program. First, the growing inflow of foreign direct investment (FDI) for hydropower development is part of the rapidly growing FDI for mining, transport, and agricultural projects, which is contributing to the surge in government expenditure and the fiscal deficit. This process appears to be overwhelming the government’s economic management capacity, contributing to threats of inflation, an unsustainable debt overhang, and 126 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES the potential for Dutch disease. Second, as stated in the FY12–16 CPS and the 2010 Lao PDR Development Report, the good practice standards established by NT2 for large-scale infrastructure projects appear to be weakening under the onslaught of escalating FDI. Third, during the two years since the FY12–16 CPS, the government was not in full compliance with the reporting and auditing requirements of the NT2 Revenue Management Agreements. However, the most recent IEG team interview suggests that this problem may have been corrected by mid-2014. Overall, the interviews for this case study confirmed that the Lao PDR country team and government are well aware of these issues and are now actively collaborating on measures to control these risks. IEG identified two additional concerns. First, although the Bank country team’s decision to forgo more active engagement in agriculture is understandable, a strong case is to be made for engaging in efforts to better mitigate the negative impacts on the well-being of the rural population of land concessions for hydropower, mining, transport, and agricultural projects. Although some of these land concessions may have benefits to the rural population that exceed their costs, in many and perhaps most cases this has not been true. Several interviewees suggested that rural inhabitants adversely affected by such land concessions are now becoming the poorest of the poor. Consistent with the Bank’s core mandate of poverty reduction, it seems essential that its country team fully engage on this key issue. A promising option is the approach being developed under the Northern Uplands Development Program. This approach reportedly includes a politically appropriate program to better protect the welfare of the rural population giving local governments several politically safe options for better containing or perhaps even reversing the potentially negative impact of land concessions on the rural population. A second potential weakness in the current approach appears to be related to the fact that the Bank’s country team has not actively examined labor mobility as a priority poverty reduction measure. International experience has shown that labor mobility has been a powerful poverty reduction mechanism. (The outflow of more than 200 million rural migrant laborers in China, for example, played perhaps the lead role in that country’s extraordinary success in poverty reduction over the past 30 years.) Some work on skills training for labor mobility is being undertaken as part of the upcoming Lao PDR Development Report, but it would appear that more ambitious analytical work on this topic is merited. The feedback loop of the core elements of the poverty reduction focus in the Bank’s country program is strong and clearly evident. The country team produced, or helped produce, sufficient evidence on poverty reduction and made good use of this evidence in the design, implementation, and evolution of the poverty reduction 127 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES focus of the country program. Given the programmatic nature of most of the core elements of the poverty reduction focus, there has been a continual learning process and feedback loop. This programmatic nature has facilitated the evolution of the country program and its poverty reduction focus. For example, the government and the Bank felt that the sectoral coverage of PRSOs 1–3 was too broad and that PRSOs 4–7 concentrated their sectoral coverage on improving the delivery of basic education and health services to the rural poor. This narrowed focus on just education and health services for the rural poor also included a strong emphasis on improving provincial reporting of basic education and health sector performance indicators, including improved management information systems for tracking education and health outcomes. SUMMARY ASSESSMENT The feedback loop of the core elements of the poverty reduction focus in the Bank’s country program is strong and clearly evident. The country team produced, or helped produce, sufficient evidence on poverty reduction and made good use of this evidence in the design, implementation, and evolution of the poverty reduction focus of the country program. Given the programmatic nature of most of the core elements of the poverty reduction focus of the country program, there has been a continual learning process and feedback loop. This programmatic nature has facilitated the evolution of the country program and its poverty reduction focus. For example, the government and the Bank felt that sectoral coverage of the PRSOs 1–3 was too broad and that PRSOs 4–7 concentrated their sectoral coverage on improving the delivery of basic education and health services to the rural poor. This narrowed focus on just education and health services for the rural poor also included a strong emphasis on improving provincial reporting of basic education and health sector performance indicators, including improved management information systems for tracking education and health outcomes. The Bank and other development partners have worked intensively to improve the quality and accessibility of statistical capacity in Lao PDR. The main objective of the 2013 Strengthening the National Statistical System Project is to “improve the capacity of the Lao PDR national statistical system to produce and disseminate reliable and timely macroeconomic and poverty statistics in accordance with international standards and in response to user needs.” Strong Bank and development partner support has also played an important role in improved monitoring and evaluation (M&E) capability in at least several Bank-supported projects and activities in Lao PDR. For example, interviews by the Independent Evaluation Group confirmed the vital role that an AusAid trust fund plays in the M&E system in the Poverty Reduction Fund (PRF) 2 Project. Similarly, the Policy 128 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES and Human Resources Development Fund financed a socioeconomic survey of electrified and un-electrified villages and households undertaken during preparation. Malawi CONTEXT Malawi is a poor, landlocked, primarily agricultural, aid-dependent country of 16 million people in southern Africa. It experienced significant economic growth in most of the past decade, averaging 7 percent over 2006–10. Most farms in Malawi are less than 1 hectare and focus on maize for food security (given past famines). There has been slow progress on diversification including away from burley tobacco (which had helped move many smallholders out of poverty). The economy is highly vulnerable to climatic, price, and political shocks, compounded by high HIV incidence. Between 2005/06 and 2010/11, the national poverty headcount fell from 52.4 percent to 50.7 percent. It declined from 25.4 percent to 17.3 percent in urban areas and rose from 56.2 percent to 56.7 percent in rural areas. The rural poverty results in particular disappointed the development community, given several pro- poor elements in the strategy. The Gini coefficient increased from 0.39 to 0.45. The subsequent period was marked by macroeconomic imbalances, falling donor inflows, and little growth, which are likely to have worsened poverty. Progress on the other MDGs during the assessment period was mixed. Between 2000 and 2012, the primary completion rate rose from 65 percent to 74 percent, the ratio of girls to boys in primary and secondary schools rose from about 96.6 percent to 103.7 percent, and the under-five mortality rate declined from 174 to 71 per (per 1000 of live births). Child mortality, malnutrition, and HIV/malaria targets are likely to be met by 2015. In contrast, maternal mortality and primary enrollment progress has been slow, and their targets will be missed. POVERTY DATA The data effort in Malawi is well above average for Sub-Saharan Africa, especially for small poor countries. Household survey data is solid, collected roughly every five/six years by National Statistical Office (NSO). Results between 2005/06 and 2010/11 (the latest round) are fully comparable and representative at the national and three sub-regional levels as well as urban/rural. Data are disaggregated, including by the gender of the household head. The surveys are well documented, accessible on both the Bank and NSO websites, and made available after formal release roughly within two years of data collection. Access has steadily improved. A 129 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES panel survey is currently underway for a subsample from 2010/11 repeated for 2013. Poverty statistics were calculated from consumption basket of food and nonfood components to the poverty line. There is a national poverty line as well as an ultra- poor poverty line based on the food poverty line. The national line is generally slightly below the international extreme poverty line, although the methods are not strictly comparable. The data include both income and non-income poverty indicators, including nutrition, infant mortality, and access to safe water, disaggregated by region, rural or urban area, and gender. Modules are comprehensive. There is little divergence between income and non-income indictors, with the possible exception of nutrition. Although there are different ethnic groups across regions, there is not an ethnic diversity dimension to poverty. The government has been keen to track poverty, especially rural poverty, as has its vast development partner community. The Bank has maintained an active research LSMS technical team throughout the period in its Development Economics (DEC). Malawi has received substantial financing and technical support from donors—in particular the government of Norway (lead donor), the Bank, DFID, Irish Aid, and Millennium Challenge Corporation—for poverty data. The donors have good mechanisms for coordination and technical discussions. On technical assistance from the Bank, the Bank LSMS Measurement Study Team in DEC has been collaborating with NSO on a multiyear program, with the objective of designing and implementing the Integrated Household Survey (IHS) in 2013. NSO and the broader stakeholder community is appreciative of the technical support, which is steady and collegial. However, capacity remains thin, as the capacity-building effort has not resulted in sustained domestic capacity, with high staff turnover and limited domestic skills. Nonetheless, Malawi is a model of how to ensure high-quality poverty data despite capacity constraints, through a combination of coordination within the Bank and with other donors. Despite attention in the context of the Millennium Development Goals (MDGs), the PRSP, and the Common Approach to Budget Support and effective coordination by Economic Planning and National Development, parts of the M&E system remain weak. For example, administrative data on education inputs and outcomes remains elusive, and agricultural production data are notoriously weak. The Bank is now stepping up its broader support efforts, including to NSO. DIAGNOSTICS The poverty diagnostic work for Malawi has been of good quality, with the Bank carrying out major exercises upon availability of the IHSs. These surveys were 130 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES available for the CAS FY07–10 (extended to FY12). A programmatic poverty work program was initiated in 2014, drawing on IHS3 2010/11. For the CPS FY13–16, a poverty update was carried out using preliminary results of IHD3. These poverty analytics link data and diagnostics, noting the high quality of the data, including poverty profiles and maps. The diagnostics made use of available data and supported enhanced data collection efforts in order to be able to not only to monitor poverty trends and guide development programs but also to evaluate at a more granular level. The Bank diagnostic reports discuss the key drivers of income and non-income poverty at the national level and for different segments of the population. They draw on sources such as a five-year rural panel data set from the International Food Policy Research Institute and the Center for Social Research. They discuss poverty trends over time and provide explanations of changes in the incidence and characteristics of the poor at the individual and regional levels. The Bank diagnostics are particularly strong on issues of poverty mobility and vulnerability. The diagnostics discuss interlinkages such as the growth-inequality-poverty nexus, especially in rural areas. Analysis of the incidence of growth shows that real incomes of the rural poor are falling and only better-off households are experiencing some growth in real expenditures. This result raises concerns about the impact on poor rural households of the Farm Input Subsidy Program, introduced in 2005/06, on poor rural households and specifically the beneficiary targeting methodology. The diagnostics are rich in gender-specific content. The Bank’s diagnostic work provides concrete recommendations to address obstacles to reducing poverty, organized by themes, taking into consideration the Malawian context and drawing on a broader body of analysis conducted outside the Bank. The diagnostics reveal breadth of coverage, with a strong team leader facilitating collaboration among a cross-sectoral team, although inevitably some parts of the country team engaged more than others. Poverty diagnostics have generally flagged important gaps in the understanding of poverty reduction, guiding the AAA program. The 2006 poverty assessment also flagged the institutional developments needed to improve M&E. Additional notes continued to update the Bank’s understanding of what works and what doesn’t in selected areas. Close collaboration with DEC helped ensure that some of its research was policy and program relevant. Other integrative pieces touched on poverty. From example, the 2010 Growth CEM included in-depth work on smallholder agriculture, and the 2007 and 2013 PERs included benefit incidence analysis in their treatment of the health, education, social protection, roads, and agriculture sectors. 131 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES STRATEGY FORMULATION, IMPLEMENTATION, AND FEEDBACK LOOPS The Bank’s strategy as laid out in the FY07–10 CAS and FY13–16 CPS reflected the poverty diagnostics and made good use of the recommendations of the poverty diagnostics, which were timely and perceived to be of good quality. The strategies were consistent with the direction and recommendations, of the poverty diagnostic work. The CASs and CPSs clearly lay out components that directly and explicitly address poverty, namely, nutrition; agriculture, including smallholder support, land, irrigation; and education quality and access. A specific example of the influence of the diagnostics was the inclusion of nutrition in the FY07–10 CAS. The 2006 poverty assessment sharply highlighted the extent of stunting in Malawi and its links to achieving progress on a broad range of fronts. Other components had been the staple of the Bank’s program over a longer period of time. For a small poor country such as Malawi, poverty considerations are inevitably central for the Bank. With two-thirds of the population below the international extreme poverty line and roughly half the population below the country’s own definition of poverty line, the broad based issues of growth and governance have a huge indirect poverty dimension. Another reason for the consistency of the poverty strategy with the diagnostics is that the government’s explicit strategy has tended to focus on its rural and poor population. Although actual practice may deviate and have different consequences, at least on paper, it is more straightforward to align than in a country that is heavily urban or in which explicit poverty reduction is absent in the country’s stated objectives. The road map for poverty reduction laid out in the country strategy had to be in the context of a large presence of donors, relief agencies, and humanitarian NGOs and strong commitment by the Bank to working within a broader development partner framework. Hence, at the same time as the Bank was linking its strategy to the diagnostics, it had to be increasingly selective and partial in the scope of its programs and nonlending activities. These choices were informed by poverty diagnostics but made largely based on evaluation of comparative advantage and gaps and encouragement by the broader community (government, other donors, civil society) for the Bank to focus on specific areas, including several with a more indirect impact on poverty. The CAS had an integral and consistent vision of the interventions needed within this broader partnership framework. The Bank’s strategies are explicit in explaining why other issues important for poverty reduction are excluded in the Bank program. The FY07–10 CAS reduced the Bank’s role in health to a junior one, concentrating on fiduciary capacity for the SWAp and malaria booster program. The Bank’s financial 132 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES support to the sector was not renewed once the two health projects expired (at the end of FY08). For intergovernmental fiscal finance and strengthening of local governments, at the coal face of service delivery to the poor, the Bank ceded leadership to the German Agency for Technical Cooperation, later replaced by the German Agency for International Cooperation. It kept the links to the broader budgetary and expenditure management intervention as well as design dialogue for the Malawi Social Action Fund (MASAF, later the Local Development Fund). These strategic selectivity choices have not been amiss, as progress in health MDGs is more or less on track and decentralization seems to be proceeding better in Malawi than elsewhere. The Bank remains heavily engaged in nutrition, where there are few partners and successful models need support in scaling up. However, the Bank continues to be spread too thinly; more tough choices will need to be made if the Bank is to ensure effective support in key areas. The results chain laid out in the CAS and CPS was convincing in relation to the projected activities and the supported outcomes based on the best knowledge at the time of formulation. In both the FY07– 10 CAS and the FY13–16 CPS, the critical areas and drivers of poverty, as developed in the rich diagnostic program, are identified and explored at length. There was a good balance between lending (both budget support and investment lending) and AAA. For vulnerability, for example, a block of policy dialogue under the development policy operations (DPO) series was devoted to effective targeting issues. The investment lending vehicle MASAF3–4, HIV/AIDS MAP (to which the Bank contributed), and the Irrigation and Rural Livelihoods Project (IRLDP) supported programs of public works and social cash transfers. The AAA program included work on specific thorny issues, such as effectiveness and allocation across programs (including relief), the targeting of Farm Input Subsidy Program, and options for exploiting data for a more accurate common targeting system. Bank support blended policy dialogue under DPOs (such as land tax) with specific poverty-targeted investments (such as Lilongwe Rural Development Project) with technical assistance and other convening functions (such as the Land Governance Assessment Framework). In many indirect areas, the CASs and CPSs identify the links to poverty reduction. For example, macroeconomic stability is explicitly linked to budget allocations and expenditure execution critical to primary service delivery. There have been some synergies across programs. The Bank has tended to concentrate its policy dialogue underpinning DPOs, including PRSCs, in areas where sectoral teams also are active in lending and nonlending services, as 133 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES suggested above for social protection and agriculture. SWAps have been effective in coordinating implementation of interventions but less so in integrating policy dialogue. Nonetheless, cross-sectoral interactions within the Bank team have been more episodic than systematic. The CASs and CPSs made explicit provisions for scaling up by maintaining presence in a subsector over an extended period. (Examples within the Bank’s portfolio, such as small-scale irrigation IRLDP, public works MASAF, and nutrition, are described below under feedback loops.) The Bank’s strategy was well aligned with the Malawi’s own strategy. The government’s priorities for poverty reduction are officially and technically stated in the Malawi Growth and Development Strategy for 2006/07–2010/11, which underpins the FY07–10 CAS, and the Malawi Poverty Reduction Strategy II, which underpins the FY13–16 CPS; the congruence and alignment of the Bank’s strategy with the government’s stated priorities are complete. However, there are important deviations, especially with what might be deemed to be actual government priorities. The 2010 CASPR determined that despite the changes in the external environment resulting from the global downturn, the strategy remained appropriate. Results have been achieved in several areas, leading to scaling up. One example is IRLDP. Intermediate outcome indicators pointed to likely positive results, which led to additional financing in both FY12 and FY13, including a scaled-up input-for- assets program to cushion the effect of the global and Malawi macroeconomic crisis on the rural poor and enhance the developmental impact of small-scale irrigation. Another example is the MASAF public works program. On the capacity-building side, the Bank has been involved in strengthening the broader M&E system of the government, of which poverty dimensions are one part. This complex undertaking links poverty outcomes with program inputs and outputs and further scales it up to cross-sectoral and cross-cutting decision making. During the assessment period, CASPRs needed to reflect both external shocks (the 2008 global food and fuel price shocks and the 2009 global economic crisis) and changes in internal circumstances during the assessment period (inappropriate responses to external shocks under the 2005–11 government and a new government in 2012 upon the president’s death). In particular, the Bank’s strategy had to be modified in a significant way in 2010–12. Although there was not a perceived need to change the strategic focus on poverty, there were adjustments in DPOs and Bank- financed programs targeting poor households. On the AAA side, the Malawi program intensified dialogue on the poverty impacts of macroeconomic developments and alternative response measures, drawing on both recent Malawi 134 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES household-level data and other country experience, which were well communicated at senior levels. Feedback loops in the Malawi country strategy from poverty data to poverty diagnostics to strategy formulation and implementation have been strong. It is not clear what the answers are, however, in a country with such severe challenges. There has been progress in certain areas: maternal and child health and nutrition indicators have improved; moderately poor households have moved out of poverty through farm and nonfarm diversification and improved access to irrigation and other agricultural services; and there has been no repeat of famine in the country. Yet the poverty challenges have only been exacerbated over time: population growth is putting further pressure on limited land resources in a situation of land threshold effects; the mix of competitive politics and continuing patrimonial attitudes of the elite is not having the hoped for impact on government institutions and accountability; and a crowded aid support network has its own unintended adverse consequences on institutional development and norms. The increase in rural inequality and the large share of ultra-poor is disappointing but perhaps not surprising. Larger country programs and the evolution of global practices can learn from the connectedness of the program and attention to learning that has been taking place in Malawi. One example is the usefulness of the DPO instrument, including PRSCs in a poverty reduction context to round out sectoral programs on the policy and budgetary resources front. In addition to cross-cutting public expenditure and financial management issues, the Bank has concentrated its policy dialogue underpinning DPOs in areas where sectoral teams are active in lending and nonlending services—namely, agriculture and vulnerability. SWAps have helped on the coordination front for implementation of interventions; they have been less successful in integrating a policy dialogue. The PRSC series has been assessed as satisfactory in its development effectiveness. Another example of good practice is the sharing of successful delivery mechanisms for targeted programs across the various sectoral teams working on Malawi. Being a smaller program may have facilitated this knowledge exchange. SUMMARY ASSESSMENT Feedback loops in the Malawi country strategy from poverty data to poverty diagnostics to strategy formulation and implementation have been strong. It is not clear, however, what the answers are in a country facing such severe challenges. There has been progress in certain areas: maternal and child health and nutrition indicators have improved; moderately poor households have moved out of poverty through farm and nonfarm diversification and improved access to irrigation and 135 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES other agricultural services; there has been no repeat of famine in the country. Yet the poverty challenges have only been exacerbated over time. Population growth is putting further pressure on limited land resources in a situation of land threshold effects; a mix of competitive politics and continuing patrimonial attitudes of the elite are not having the hoped for impact on government institutions and accountability; and a crowded aid support network has its own unintended adverse consequences on institutional development and norms. The increase in rural inequality and share of ultra-poor is disappointing but perhaps not surprising. Nigeria CONTEXT Nigerian statistics reveal a puzzling contrast between rapid economic growth and minimal welfare improvements for much of the population. Annual GDP growth rates that average more than 7 percent in official data during the last decade place Nigeria among the fastest-growing economies in the world. GNI per capita reached $2,580 in 2011. This rapid growth has been concentrated particularly in trade and agriculture, which would suggest substantial welfare benefits for many Nigerians. Nevertheless, improvements in social welfare indicators have been much slower than would be expected in the context of this growth. Poverty reduction and job creation have not kept pace with population growth, implying social distress for an increasing number of Nigerians. Progress toward many of the MDGs been slow, and the country ranked 153 out of 186 countries in the 2013 United Nations Human Development Index. Nigeria weathered the 2008–09 crisis, the sharp fall in petroleum prices, and the virtual collapse of parts of the banking sector well. The government drew on the special stabilization account established in 2004 from oil revenues to maintain public expenditures and finance a stimulus package. Although nonoil growth continued at a robust pace and real GDP grew at 7 percent between 2009 and 2012, with average per capita income growth reaching 4.3 percent, well above the average of about 2.2 percent per year in 2003–13, it had a limited impact on poverty, which remained very high (54 percent of the population at a poverty line of $1.25 a day in 2011). Nigeria is unlikely to meet most of the MDGs, with the primary school enrollment rate stuck below 65 percent, an under-five mortality rate of about 124 per 1,000 live births, and a maternal mortality ratio of 610 per 100,000 live births, one of the highest ratios in the world. Access to improved sanitation and safe water remain appallingly low at 28 percent and 64 percent, respectively. 136 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES Comparison of poverty estimates using adult-equivalent welfare measures and absolute poverty lines for 2004 and 2010, with only a modest decline in national poverty headcount from 48 to 46 percent, along with modest declines in the poverty gap and severity of poverty, raises questions about the quality of the consumption data. Both biases in the surveys themselves and changes that introduced comparability issues have undermined the quality and reliability of poverty data. Poverty may have been slightly overestimated in 2010 compared with 2004. Even if the figures are overestimated, however, poverty is still high. Poverty is much higher in rural areas and in northern states, and inequality has been rising. There is also a large North–South divide in terms of education, health care, and the condition of women. POVERTY DATA During the period of this review, the Bank began to obtain survey data in Nigeria to carry out poverty diagnostics and guide development programs on poverty reduction. The comprehensive living standard survey (Harmonized Nigeria Living Standard Survey, HNLSS) is well designed being representative, timely, and distinguishing different degrees or levels of poverty, including extreme (food) poverty. Nonetheless, official statistics in general and the results of the large 2009/10 HNLSS survey in particular lack credibility. Data are considered insufficient for strategic purposes. It appears increasingly likely that the 2009/10 survey underestimated consumption and hence overestimated poverty rates in Nigeria. Inadequate data reflect a combination of the change in the consumption estimation method and the way in which the survey was carried out (including nonpayment of enumerators in the later phase of the survey, given the shortfall in the government’s budgetary contribution), as acknowledged by both the National Bureau of Statistics and the Bank. Technical problems with the first 2003/04 survey were discovered only recently. To reinforce the principle of cost sharing, as agreed to before the survey, the DFID, which financed the survey, chose not to fill the financing gap. Data concerns extend to other components relevant for understanding the drivers of poverty. At least for the HNLSS, the Bank teams have access to the raw data, something they lack for other critical core statistics (including the recent GDP rebasing exercise). Population numbers are very outdated. For non-income poverty indicators, there are multiple sources and conflicting trends. Agricultural data are weak, with the recent GDP rebasing exercise resulting in much lower agricultural growth rates. The Bank is working with multiple partners to provide technical assistance on several aspects of these data series. 137 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES DIAGNOSTICS The poverty assessment and other ESW made good use of available data, which expanded over time and include disaggregated poverty profiles. The availability of comprehensive household survey data was a major contributing factor. Notes and assessments were completed two to three years after the surveys, about a year after the release of major poverty data sets, which were available at the time of strategy formulation. The 2004 poverty assessment, conducted in the context of country reengagement, drew on limited qualitative and quantitative surveys. A relatively comprehensive poverty assessment was carried out in 2006/07, taking advantage of the 2004 household survey. It was a more traditional and comprehensive, including on policy recommendations, drawing on sectoral reports produced jointly with development partners. The 2013 Poverty Note used analysis of the comprehensive household survey of 2009/10. The chapter on poverty in the 2014 Nigeria Economic Report draws on the panel surveys of 2010/11 and 2012/13. Partners look to the Bank’s diagnostics for guidance and leadership on the poverty agenda. The poverty diagnostics’ assessment of the adequacy of the Nigeria’s institutions, programs, funding, and M&E arrangements was partial and stronger in some areas than in others. A major contributing factor was the fact that the Bank’s diagnostics have generally narrowed explicit poverty considerations to the human development sector and to a lesser extent agriculture, as in the major 2007 poverty assessment. This poverty assessment was of high quality in its analysis and actionable recommendations. The 2004 and 2013 poverty assessments made broader links (namely, macroeconomic stability as well as growth and employment, respectively) but at a high level of generality. The 2013 poverty assessment also used the data to confirm that poverty correlates by and large had not changed. In a country as large and complex as Nigeria, other diagnostic work would have been expected to drill down on key poverty reduction issues and inform both the client and the Bank on how to operationalize their strategies. Notable in this regard was the public expenditure and monitoring work on basic health (for example, the 2005 Health Status Report) and education services (for example, the 2008 Review of Costs and Financing of Public Education). Other reports included the 2006 Getting Agriculture Going report, the 2009 Employment and Growth study, and the 2013 Social Safety Net Stocktaking. Although not explicitly included in the poverty assessment narratives, the Bank continued to work on fiscal management of oil wealth at both the federal and state level, with a focus on macroeconomic stability and sustainability for indirect poverty impact. Overall, the Bank’s analytics are held in high regard. 138 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES Even if the totality of the poverty diagnostics had represented a more integrated package, the major weakness of the Bank’s poverty diagnostics would still have been the effectiveness of those diagnostics. The Bank focused more heavily on the technical quality of the work than on following up and communicating with stakeholders, resulting in much more limited awareness of the Bank’s work than necessary to have impact in a large and complex country like Nigeria. There is a relatively low level of understanding of poverty analysis and only limited Bank efforts at outreach. No interviewee outside the Bank referenced the Bank’s 2004, 2007, or 2013 poverty assessments. Weaknesses in the data (i.e., no comprehensive survey available in 2004, the absence of multiple surveys in 2007 to analyze drivers of poverty, and the underestimation of consumption in 2013) limited the ability to draw strong conclusions or make credible recommendations on several issues. The poverty diagnostic work has not been used to derive action plans for future poverty reduction work and the country’s poverty strategy, although elements are contained in select sectoral components. The diagnostics did not explain how and when such an action plan or updated strategy will be developed. The character and extent of poverty in Nigeria did not change much during the decade. The Bank’s strategy partly reflected the poverty diagnostics. Each of the strategies focused on similar issues, including areas highlighted by the poverty evidence and diagnostics as key to poverty reduction. The country strategy tended especially to emphasize the human development component of the poverty agenda, in particular expanding social service delivery in health and education, with a focus on states but spanning federal, state, local and community levels. Another issue with significant poverty focus was agriculture productivity, as a subset of the nonoil growth agenda. The regional differentiation issue, also underpinned by the poverty diagnostics, was flagged in the strategies, usually as part of the governance agenda. STRATEGY FORMULATION, IMPLEMENTATION, AND FEEDBACK LOOPS The FY05–09 CAS put poverty front and center in its motivation and context setting. It noted that Nigeria had 70 million people in poverty, the third largest number in the world after China and India, and drew on the 2004 poverty assessment and other findings for details. Given these levels of poverty, the poverty assessment was an advocacy document for more international support, noting that official development assistance was only $2 per person in Nigeria versus $28 for Africa. It made the case for consideration of debt relief, noting the government’s commitment to use the proceeds for MDG-related efforts, even if funding is fungible. The strategy articulated poverty links to individual components of the strategy, but a results chain was not explicitly developed. It identified gaps in knowledge and proposed to fill them through poverty assessment and statistical capacity building. 139 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES The FY10–13 CPS provided an update on the poverty context in Nigeria and made general observations about the importance of nonoil growth to poverty reduction. It was less explicit about how the individual strategic components related to the poverty challenges (something that was never asked for in the corporate context of that time), so there was no explicit results chain relating projected activities to poverty reduction at the national or sector levels. These, however, can be inferred. Midway through this period, the CPS shifted slightly to support priorities of the new government, including the agriculture transformation agenda, which had some pro-poor focus. The poverty diagnostics did not drive the strategies. Several other considerations drove the strategies, particularly the centrality of governance, which was a Bank institutional priority, and the political economy context, which constricted the strategy space in which the Bank operated. Because of the nature of the client, the Bank’s country strategies did not have had an effective road map or integral and consistent vision of the interventions needed to reduce poverty. Notably, governance and corruption issues associated with a resource-rich state and the complexities of federalism and regional differentiation, as well as the potential for conflict along ethnic, religious, and regional lines, in Africa’s largest country meant significant divergence between the country’s stated objectives and actual Nigerian policy and program implementation. Although the country strategies made explicit provision for scaling up, the Bank struggled to find financing modalities that actually worked to produce basic service delivery results in the Nigerian government structure. The social community-driven development (CDD) programs are not having the hoped for scaling up through effective local government linkages, despite project components aimed at doing so. Building on a decade of Bank knowledge- sharing on social protection, CDD safety net components are being supported in states with demonstrated ownership of reforms. This approach builds on existing modalities, including the government’s conditional grant scheme, including those with a focus on youth employment. Support to nonoil growth has been a second component of the Nigeria country strategies. In both diagnostics and strategy formulation, nonoil growth was deemed as having strong synergies and complementarities with the social service delivery agenda: the poor can realize benefits of services only in the face of economic growth. The poverty reduction case for support to agricultural productivity was more central, given the larger rural nature of poverty in Nigeria. This is home to the productive CDD Fadama projects, which have explicit poverty focus. They have been successful in raising incomes in supported communities and been scaled up across Nigeria. 140 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES The case for addressing the huge infrastructure deficit for growth was compelling, for example, power sector reforms where the Bank has played a transformational role, although the poverty-focused elements were more limited in projects such as rural roads and Lagos urban transport (e.g., lowering bus costs for poor). The focus on geographic growth poles made sense from an aggregate growth perspective, but with the caveat that these growth poles were not primarily in those parts of the country with the greatest poverty headcount percentages, although some were in largely populated areas with large numbers of poor. The Bank CPS adjustment in the face of the global economic crisis included a DPO supporting financial and PFM reforms without poverty linkages. The adjustment of the CPS midway, as reflected in the CPSPR, including the agricultural DPO series, was aimed at getting behind the Johnson government’s reforming Agriculture Transformation Agenda (ATA), which included at least some pro-poor elements. However, Fadama is not yet supported with the government’s own sizable resources. In addition, the trade-offs, if any, between a focus on growth (agricultural commercialization and growth poles) and poverty reduction, let alone the distinction between moderately and core poor, were less explicit in the strategy. One component of the DPO that is explicitly pro-poor is the elimination of the government’s direct fertilizer procurement and distribution system. Another potential synergy in the strategies that has not been realized is geographic. The Bank has only recently made an effort to systematically link its rural interventions, including the agricultural market–focused Rural Access and Mobility Project, with the successful scaling up of a pilot to other states. Notable from the poverty perspective is that the additional financing for Fadama 3 focuses on linking Fadama groups to the ATA commercialization agenda and coordination. The apparent increase in income equality suggests the centrality of a focus on broader based nonoil growth. The nonoil growth story also needs to be nuanced from the geographic perspective: the national average hides large differences across regions. Preliminary evidence from the (more reliable) panel survey suggests that parts of the country, such as the Southwest, have also experienced rapid declines in poverty. In parts of the North, growth has been more modest and accompanied by worsening inequality—notably in the Northeast—and hence worsening poverty. Feedback loops on poverty reduction from data to diagnostics to strategy formulation and implementation have been incomplete in Nigeria. There were no Nigerian champions demanding data and diagnostics. The Bank-financed explicit poverty-focused interventions were of small scale relative to the problems and the government’s own resources, even though several demonstrated strong technical 141 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES and interpersonal approaches to evaluation. Lending programs for Nigeria required attention, as the largest IDA program for the Africa Region, but “projectizing” drained energy and incentives away from a more concerted effort to help build a constituency for poverty reduction. It may be too much to expect that the Bank could ever have the instruments and leverage to be transformational in this country, so expectations need to be tempered. However, not all options have yet to be explored. There appear to have been few champions or strategists in Nigeria who were linking their growth-oriented agenda more strategically to poverty consideration including jobs, urbanization and agricultural commercialization agendas, compounded by the competitiveness difficulties facing any resource-rich economy. The lack of clear evidence of poverty reduction in official numbers after strong nonoil growth— combined with concern with Nigeria’s image from unrest in poorer parts of the country—might represent a window of opportunity for a more transformational approach to poverty reduction in Nigeria than has been possible in the past. The Bank’s articulation of such growth-poverty links could help move forward this bigger picture and get more traction with leadership, going beyond targeted focus on safety nets and social service access. This goes well beyond technical foundations usually covered by poverty assessments and calls for blending political economy analysis, microeconomic poverty analysis, pragmatic knowledge sharing and effective communication within the Bank’s wider program. Evidence on poverty reduction in the Nigeria program was modest. Explicit attention to poverty reduction objectives was either at a high level of generality (as in debt relief or nonoil sector growth) or detailed in only a subset of the program (as in CDD programs). The Bank program for Nigeria did monitor the MDGs, which include an indicator on income poverty. However, in general, the Bank did not produce substantial information on poverty reduction that would lead to M&E. In neither of the CASs under review was poverty included in the proposed outcomes and results to be monitored as part of assessment of the Bank’s program, including within a sectoral context. The Bank’s country programs evolved, although there were no major deviations of the implemented operations addressing poverty from the original CAS and CPS. Given the governance challenges and need to focus on champions of reform, the Bank shifted its programs accordingly. For example, it continued to stay engaged in health, enabled by pockets of reform champions at both the federal and state level, notably for maternal, child, and other basic health services that are pro-poor. The Bank responded to the global crisis with a (financial-sector/PFM) DPO combined with ramping up of social safety net dialogue to increase the visibility of the 142 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES weaknesses in Nigeria. The slight pro-poor shift was enabled by a strong Bank team, which once again is receiving a little more attention in light of unrest in the Northeast. Other obstacles in implementing the Bank’s strategy (for example, lending delays because of disagreements between the executive and legislative branches) were unrelated to the poverty focus. SUMMARY ASSESSMENT The Bank is a small player in financial terms in resource-rich Nigeria, with annual lending level representing roughly 2 percent of federal revenues. This has seriously constrained the ability of the Bank to build effective poverty reduction into its strategy. The Bank has likely had a greater impact on poverty reduction through reliance on champions—and there are committed counterparts, despite the aggregate picture. Support to champions went beyond financial resources. Some of the highest-impact interventions provided knowledge and technical support to motivated Nigerian teams within sectors and states, getting behind reform teams with solid analytics and practical support (on issues ranging from debt relief, fiscal reform, and a virtual poverty fund to impact evaluation of primary health care services at the federal level and from procurement to education learning outcomes in states). The Bank appears to have become less risk averse over time, engaging more actively on the petroleum subsidy and publicly flagging the extent of poverty in Nigeria, but its efforts have probably been excessively balanced in favor of lending over AAA. The Bank will need to play an even more prominent advocacy role, putting the poverty issue in a context that promotes domestic dialogue and debate; for example, couching it in terms of inequality or jobs or peace and security. A possible lesson is that AAA and advocacy may be even more important in a resource-rich country, especially if the timing and approach are well informed. The importance to the overall lending program of a large country such as Nigeria cannot be allowed to overshadow this imperative. Feedback loops on poverty reduction from data to diagnostics to strategy formulation and implementation have been incomplete in Nigeria. There were no Nigerian champions demanding data or diagnostics. The Bank-financed poverty- focused interventions were of small scale relative to the scale of the problems and the government’s own resources, although several demonstrated strong technical and interpersonal approaches to evaluation. The lending program for Nigeria required attention as the largest IDA program for the Africa Region, but “projectizing” drained energy and incentives away from more concerted efforts to help build a constituency for poverty reduction. It may be too much to expect that the Bank could ever have the instruments and leverage to be transformational in 143 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES Nigeria, so expectations may need to be tempered. However, not all options have yet to be explored. There appear to have been few champions or strategists in Nigeria who were linking their growth-oriented agenda more strategically to poverty considerations, including jobs, urbanization, and agricultural commercialization agendas, compounded by the competitiveness difficulties facing any resource-rich economy (from Dutch disease). The lack of clear evidence of poverty reduction in official numbers after strong nonoil growth—combined with concern with Nigeria’s image from unrest in poorer parts of the country—may represent a window of opportunity for a more transformational approach to poverty reduction in than has been possible in the past. The Bank’s articulation of such growth-poverty links could help move forward this bigger picture and get more traction with leadership, going beyond a targeted focus on safety nets and access to social services. The approach goes well beyond technical foundations usually covered by poverty assessments and calls for blending political economy analysis, microeconomic poverty analysis, pragmatic knowledge sharing, and effective communication within the Bank’s wider program. Evidence on poverty reduction in the Nigeria program was modest. Explicit attention to poverty-reduction objectives was either at a high level of generality (as in debt relief and nonoil sector growth) or detailed in only a subset of the program (as in CDD programs). The Bank program for Nigeria did monitor the MDGs, which include an indicator on income poverty. However, in general, the Bank did not produce substantial information on poverty reduction that would lead to M&E. In neither of the CASs under review was poverty included in the proposed outcomes and results to be monitored as part of assessment of the Bank’s program, including within a sectoral context. The Bank’s country programs evolved, although there were no major deviations of the implemented operations addressing poverty from the original CAS or CPS. Given the governance challenges and need to focus on champions of reform, the Bank shifted its programs accordingly. For example, it continued to stay engaged in health, enabled by pockets of reform champions at both the federal and state levels, notably for maternal/child and other basic health services which are pro-poor. The Bank responded to the global crisis with a (financial sector PFM) DPO, combined with ramping up of social safety net dialogue to increase the visibility of the weaknesses in Nigeria. The slight pro-poor shift was enabled by a strong Bank team, which is receiving a little more attention in light of unrest in the Northeast. Other obstacles in implementing the Bank’s strategy (for example, lending delays because of disagreements between the executive and legislative branches) were unrelated to the poverty focus. 144 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES Peru CONTEXT Peru grew steadily at about 6 percent a year during the last decade, and GDP per capita grew at an annual rate of about 5 percent. GNI per capita reached $9,440 in 2011. The poverty incidence fell from about 58 percent to 31 percent during 2004– 2012, with the incidence of extreme poverty falling from 17 percent to 10 percent. The stability of economic policies at the macro and regulatory level has created good incentives for domestic and foreign direct investment, which a benign external scenario has also helped. According to the latest IMF Article IV report, Peru has been one of the best macroeconomic performers in Latin America over the past decade. It continued to be a leader in high growth and low inflation in the region, thanks to prudent macroeconomic policy implementation, a far-reaching structural reform agenda, and a benign external environment. Not least because of these factors, the economy came out virtually unscathed from the 2008‒09 global financial crisis, with growth rebounding to nearly 9 percent growth in 2010 and being sustained at high levels in 2011–12. There is a strong consensus among Peruvian experts that growth has been a key driver of poverty reduction across income groups and regions. However, large regional differences remain, in both poverty rates and non-income indicators. The incidence of extreme poverty is three times the national average in the rural sierra and twice the national average in the jungle. The incidence of malnutrition in children under the age of five is 38 percent in the lowest income quintile and 2 percent in the highest quintile (the average for the entire population is 17.5 percent). POVERTY DATA Between 2000 and 2012, Peru carried out 12 National Household Surveys and seven nationally representative Demographic and Health Surveys to assess the health and nutritional status of the population. The evolution of the national statistical system faced some significant challenges; the Bank played a critical role in reforming the system. In 2004 several changes in the methodologies for the National Household Surveys took place, resulting in changes in poverty estimates. As a consequence, no official poverty estimates were made publicly available between 2004 and 2007, which added to the loss of credibility of the poverty data and of the National Institute of Statistics. The authorities requested technical assistance from the Bank to address the methodological problems and to help restore public trust. Rather than providing solely traditional technical assistance, in 2007 the Bank set up an external expert advisory committee to help reach consensus on best methodological practices to produce comparable poverty estimates. Under this new initiative, the Instituto 145 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES Nacional de Estadística e Informática (INEI) was able to issue comparable poverty figures for all years between 2001 and 2010. The communication of the results was transparent, the figures were not contested, and public trust was restored. In 2011 new methodological changes were adopted to better reflect the profound socioeconomic changes experienced by Peru during the past 15 years. As a result of the Bank’s technical assistance, INEI has become institutionally stronger, and transparency has increased substantively. DIAGNOSTICS Three sets of diagnostic work were available before the two CPSs during the period of this evaluation: the 2005 poverty assessment, the 2006 New Social Contract for Peru, and public finance studies, such as the Decentralization Process and the Public Expenditure Process and the PER of 2012. In general, these studies were of high quality in terms of the meaningful use of empirical evidence (including good-quality statistics), relevance, timeliness, depth of analysis, and identification of policy directions. A major contribution of the poverty assessment is the integrated (general equilibrium) nature of the analysis. The assessment recognizes the role of growth and sustainability issues and provides a taxonomy of the most binding constrains on urban employment expansion and the productivity growth of poor farmers. It then turns to the provision of social services to the poor and the role played by low- income households’ demand for those services. The emphasis on non-income aspects of poverty is critical in Peru, because some of these indicators are lagging with respect to improvements in incomes. For example, although children malnutrition has been reduced, it remains high, despite income growth of the poorer segments of the population, particularly in isolated rural areas. The discussion of the demand side is important and innovative. A large proportion of the poor is concentrated in indigenous groups, whose demand and use of social services depends highly on the degree to which these services are aligned with their cultural practices. There is consensus in Peru that a key issue is the fact that the remaining pockets of extremely poor are increasingly concentrated among rural indigenous groups. The New Social Contract is a good-quality diagnostic report relevant to the delivery of social services to poor families. The five-year Accountability for Social Responsibility (RECURSO) programmatic AAA program played a critical role in the diagnosis and what is needed to move the delivery system out of a low-equilibrium trap. It calls for action across a wide front, ranging from the need for standards, transparent information, and better ways to monitor quality to a new a system of 146 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES incentives and accountabilities to improve incentives on the supply side. This work supported the dialogue with the incoming authorities at the time and was disseminated widely in Peru, helping create consensus on the major areas of reform in the social sectors. This strong effort at dissemination is an especially positive aspect of the Bank’s diagnostic work in Peru. The work on public finances undertaken in 2010 and 2012 focus on the regional dimensions of poverty. It also updates the expenditure incidence analysis, calling attention to the relatively small size of the more poverty-targeted programs compared with other social transfers. The study is candid about the fact that the fiscal rules guiding the allocation of mineral revenues may not be reducing regional inequalities. The Bank’s diagnostic reports for Peru were innovative and candid, and they were disseminated widely in the country. STRATEGY FORMULATION, IMPLEMENTATION, AND FEEDBACK LOOPS The FY07–11 CPS followed the key areas identified in the diagnostic work: widening the base of growth (in particular rural infrastructure in poor areas), promoting a new social contract in the delivery of social services, and addressing the challenges of decentralization and the new responsibilities of local governments. However, the CPS had a cryptic discussion of social assistance in Peru, and the targeted programs represented only a small share of GDP compared with the rest of the Latin American and Caribbean Region. The CPS did stress some specific areas of urgency raised in the diagnostic work, such as infant malnutrition; educational quality, which has become a major priority; and the system of fiscal transfers, which may have accentuated regional inequalities. The FY12–16 CPS has a significant degree of continuity with the previous strategy, but it has a sharper focus on extreme poverty, with an emphasis on reaching the rural poor. It explicitly acknowledges the significantly higher level of poverty and child malnutrition in rural selva (jungle) areas than in the rest of the country, and it identifies a set of priorities targeted to these groups. The CPS also identifies more specific areas of targeted interventions than the earlier CPS. These targeted interventions reflect the priorities of the new administration. Overall, the proposed portfolio of lending and nonlending activities in the two CPSs is consistent with the diagnostic work and the formulated strategies. The FY07–11 CPS envisaged a series of three social sector DPLs to support key policy steps in the social sector delivery in the context of the announced decentralization of services. In spite of their potential systemic importance, there are no details on the specific policy steps to be supported by these DPLs. This area remains vague, an important shortcoming of the FY07–11CPS. There are no efforts to develop a results chain 147 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES linking these operations to objectives in this area. The strategy also contains a follow up of the RECURSO AAA program, which seems highly relevant to poverty and potentially to the design of the social DPLs. However the CPS missed an opportunity by not integrating well the discussion of the diagnostic work with the lending operations. Similarly, the program envisages five DPLs in the area of fiscal management and competitiveness, with an emphasis on examining public expenditures and result-based systems. These operations could potentially be good vehicles to address the inequalities created by fiscal transfers and to reallocate resources to the most targeted social transfer programs. Another major shortcoming of the FY07–11CPS was the lack of reference to the special situation of the indigenous populations and ways to adapt basic education and health interventions to meet the need of these groups. In contrast, the FY12–16 CPS envisages new initiatives to help consolidate the dispersed programs targeted to the poor and make Juntos, the main CCT program, the centerpiece of such assistance. The initiative is to be implemented with a tight timetable, and the CPS envisages a possible DPL or SWAp, accompanied by strong technical assistance. A SWAp in the area of nutrition (not originally planned in the first CPS) is to support expansion of Juntos in this area. An important piece of AAA, RECURSO V, is to study (through a survey) the extent to which better parental knowledge on the effect of nutrition could enhance the impact of Juntos. The envisaged activities in this area thus incorporate important complementarities between lending, technical assistance, and AAA, with a clear result chain of outcomes and milestones of progress. At the sectoral level, the FY12–16 CPS also included a new Health Reform III project to reach poor rural mothers and children, making health facilities culturally appropriate for potential beneficiaries. The explicit objective is to increase the number of institutional rural births in the nine poorest regions of the country. An education SWAp envisions introducing a result-based system involving the participation of families and communities, with the objective of improving learning outcomes of children from rural and indigenous communities. Thus in both health and education, the interventions proposed were highly targeted to the poorest families. Both CPSs also included activities to provide better infrastructure to the rural poor and help them get better connected to markets and services. The FY07–11 CPS envisaged operations to reach the rural poor in the Sierra and assist the rural transport decentralization process. The Sierra projects aimed at enhancing the connectivity of farmers to markets, which seemed to be relevant interventions. However, there is no discussion of how these isolated or pilot type interventions 148 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES would be replicated or scaled up when Bank funding ceases; a result chain in addressing this issue was missing in the CPS. The FY12–16 CPS has a richer set of proposed interventions in rural infrastructure. Projects on decentralized rural transport seek to strengthen the role of the Provincial Road Institutes and promote microenterprises in the routine maintenance of rural and provincial roads. However, the discussion does not provide information on the extent to which these interventions will focus on poorer rural areas or how the project will mobilize resources through microenterprises beyond the project period. A similar problem emerges in the discussion of projects in electricity and communications. The CPS, however, does include a targeted initiative in water and sanitation aimed at poor populations in peri-urban and rural areas, including the efficiency of small and medium-size water utilities. The results chain incorporates outcomes and milestones, as well as possible scaling up by the International Finance Corporation (IFC), providing advisory services to support public-private partnerships at the subnational level. Finally, the FY12–16 CPS discusses some activities to strengthen local government and advance the decentralization agenda. The idea is to identify key bottlenecks according to the different capacities and needs of local governments, probably through some technical assistance activities. This activity seems to be a highly relevant given the new responsibilities of local governments. However, the discussion is not sufficiently specific and does not elaborate on how priorities will be determined. The CPS documentation suggests that the operational strategy contains a more explicit and clearer set of interventions to reach the poor in the social sectors and safety nets than in the physical infrastructure sector, including the issues of scaling up and sustainability. The FY12–16 CPS also includes a richer description of envisaged operations to reach the most vulnerable groups. The support given by the Bank to the social sectors represents a significant share of its total support in Peru. Social sector DPLs accounted for about 25 percent of total lending to Peru during the review period, and the RECURSO programmatic AAA accounted for the largest bulk of the analytical work. Based on the project documentation approved and discussions in the field, the portfolio had significantly more coherence, clarity, and synergisms than what was described in the CPS documents. Several positive features characterize Bank support for the social sectors in Peru. These features, which were acknowledged explicitly in the field by Peruvian counterparts, include (i) strong complementarity (stronger than envisaged in the CPS) between the DPLs, AAA, and technical assistance, with a high level of 149 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES participation by Peruvian counterparts; (ii) good synergism between the steps supported by the DPL series and the individual operations in the sectors; and (iii) consistency between the design of projects in health, nutrition, and education and the diagnostic work on poverty, in particular in better targeting resources to the poorest households, particularly indigenous populations in rural areas that required special design features. Child malnutrition, infant mortality rates, and maternal mortality ratios were special areas of focus. The health project targeted reducing maternal and infant mortality in the nine poorest regions in the country. Peru seems to have valued most the convening role and know-how of the Bank— much more than its financing contribution. The DPLs and SWAps were very small relative to the Peruvian economy and relative to the scope of the technical assistance operation. The Bank team working with the Peruvian authorities had a high level of expertise and commitment to the Bank’s program of support for Peru, which was critical given the strong and innovative nature of the analytical work, the quality of the dialogue, and the degree of trust that was generated. Most Bank task team leaders remained engaged with Peru for at least five years. Peruvian counterparts unanimously acknowledged this commitment. Efforts to better connect the poor to services and markets focused on pilot projects that could lead to innovation, learning, and scaling up. Project documents contained result chains identifying modalities to facilitate financial sustainability, mobilizing the willingness to pay from rural communities together with some transparent subsidization to allow reaching poor groups (such as localities increasingly dispersed and far from the grid). Mobilizing private investment was important to ensure scaling up and sustainability. Peruvian experts explicitly acknowledged these efforts, particularly on roads. However, there are some concerns about scaling up and sustainability, which are explicitly discussed in project documents. Overall, the portfolio of approved operations contained a much clearer articulation of the poverty objectives and result chains than was described in the CPS documentation. There was also a positive change over time: the second CPS has a clearer articulation and poverty focus than the FY07–11 CPS, and the operations in the social services area were better aimed at reaching the poorest groups. The Bank’s self-evaluations based on CAS and CPS Progress Reports and Completion Reports, as well as status reports of individual projects, converge in the three important operational messages. First, the process of regional decentralization has made efforts to reach the poorest and more isolated rural populations more challenging. Weak institutional capacity at the local level, particularly in the poorest municipalities, limits the effectiveness of additional transfers to these municipalities. 150 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES Replacing traditional investment loans by technical assistance and capacity-building activities targeted to local governments may remedy the situation. Second, the challenge of reaching the more isolated rural populations is manifested particularly by the difficulties of achieving major improvements in maternal and child health. Neonatal mortality, early malnutrition, slow improvements in breastfeeding practices, and the slow expansion of laboratory testing for pregnant mothers are some examples of problem areas that call for solutions that need to be fine-tuned at the local level. Third, the challenges of reaching more dispersed rural groups are detected in some of the rural community development projects and electrification projects. These projects require significant coordination across levels of government, the mobilization of local finance. Subprojects play a critical role in scaling up and sustaining these investments. These messages seem to suggest the importance of experimenting with local pilot projects, complemented by technical assistance to local municipalities and incentives to scale up pilots by mobilizing resources at the local level. A distinctive contribution of the Bank program was to help Peruvian households monitor progress in the social sectors themselves. Enhancing local M&E capabilities to gauge service quality and delivery has become a major priority in Peru, with the issue raised by many stakeholders during the field visit. Creating metrics and standards to better inform users and local authorities is seen as critical. The Bank seems to have played a major role in empowering poor families by producing and disseminating information on nutrition standards and learning outcomes as a means to evaluate outcomes and the quality of social services at the local level. For example, the Bank team worked to develop easy-to-understand metrics for education and health services. The My Future, My First Centimeters initiative showed that children everywhere in the world have the same growth potential in the first five years of life and that bad nutritional practices and lack of information can contribute to stunting. The program led to a wider recognition and understanding of inequality in nutritional status and educational outcomes across localities. Since 1985 Peru has consistently collected household survey data to assess living standards and provide the evidence necessary to design social policy. The process has improved over time; today the information system is continuously up to date. The surveys include both income and non-income indicators and provide nationally and regionally representative data on different aspects of livelihood. The diagnostic reports prepared by the Bank covered the most relevant issues of poverty in Peru. Overall, the studies were of a very good quality in terms of meaningful use of empirical evidence (complemented with good-quality statistics), relevance and timeliness, the depth of analysis, and identification of policy directions. The most 151 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES recent CPS has a stronger link to poverty issues in Peru and better reflects the findings and recommendations of the Bank’s diagnostic work. A key issue in Peru has been the significant divergence between the planned portfolio, as reflected in the CPSs, and the implemented portfolios. The situation improved in more recent years, particularly with the FY12–16 CPS, and there is now strong complementarity between the Bank’s poverty-related lending and nonlending activities. Key features of the Bank’s program of support have been strong complementarity between AAA and technical assistance, with a high level of participation of Peruvian counterparts; good synergy between the steps supported by the Results and Accountability (REACT) DPL series, the Ministry of Development and Social Inclusion (MIDIS) DPL, and individual operations in the sectors in which these DPLs supported the introduction of standards and monitoring systems to start strengthening beneficiaries’ power and hold providers more accountable; strong complementarity between DPLs and technical assistance; and the specific design of projects in health, nutrition, and education, which was consistent with the diagnostics work on poverty. The Peruvian counterparts seemed to have valued the most the convening role and the know-how of the Bank over its financing contribution. The DPLs and SWAps were very small relative to the Peruvian economy, except for REACT II, which was approved during the 2009 global crisis (and consisted largely of a deferred drawdown option [DDO] with a contingent component). The lack of interest in large loans became more apparent as growth in Peru resumed and its external finances improved SUMMARY ASSESSMENT Since 1985 Peru has consistently collected household survey data to assess living standards and provide the evidence necessary to design social policy. The process has improved over time; today the information system is continuously up to date. Household surveys covers both income and non-income indicators and provide nationally and regionally representative data on different aspects of livelihood. The diagnostic reports prepared by the Bank covered the most relevant poverty issues in Peru. Overall, these studies were of a very good quality in terms of meaningful use of empirical evidence (complemented with good-quality statistics), relevance and timeliness, depth of analysis, and identification of policy directions. The most recent CPS had a stronger link to poverty issues and better reflected the findings and recommendations of the Bank’s diagnostic work. A key issue has been a divergence between the planned portfolio, as reflected in the CPS, and the implemented portfolios. The situation has improved in more recent years, particularly with the FY12–16 CPS, and there is now strong complementarity 152 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES between the Bank’s poverty-related lending and nonlending activities. Detailed review of the formulated CPSs and their implementation reveals that the operational strategy contains a more explicit and clearer set of interventions to reach the poor in the social sectors and safety nets than in the physical infrastructure sectors. In addition, concerns about scaling up and sustainability emerge from many of the interventions in infrastructure—at least in the way these interventions have been discussed in the CPS documents. Key features of the Bank’s program of support have been the strong complementarity between AAA, technical assistance, and lending with a high level of participation of Peruvian counterparts. Moreover, synergies between the steps supported by the REACT DPL series, the MIDIS DPL, and the individual operations in the sectors where these DPLs supported the introduction of standards and monitoring systems to start strengthening beneficiaries’ power and hold providers more accountable. Other important synergies included: (i)the strong complementarity between DPLs and technical assistance; (ii) the design of projects in health, nutrition, and education, which was consistent with the diagnostic work on poverty; and (iii) the fact that Peruvian counterparts seem to value the technical expertise and know-how, as well as the convening role of the Bank over its financing contribution. The DPLs and sector-wide approaches were very small relative to the Peruvian economy, except for REACT II, approved during the 2009 global crisis (and consisting largely of a development policy loan with a draw down option with a contingent component). Peru’s lack of interest in large loans grew as growth recovered and external finances improved. An important factor in the success of Bank support in Peru seems to have the fact that the Bank mobilized a team with a high level of expertise and commitment. Most of the Bank’s task team leaders remained engaged with Peru for at least five years. Its commitment and expertise seems to have been critical for the strong and innovative nature of the analytical work, the quality of the dialogue, and the high degree of trust that was generated. Philippines CONTEXT The Philippines experienced moderate annual growth of 4–5 percent during the early 2000s that recently accelerated to 7–8 percent. GDP per capita grew by an average of 3.5 percent a year during 2003–13. GNI per capita reached $4,140 by 2011. 153 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES Despite this good growth performance, the poverty headcount changed very little in the past decade, inequality has remained stubbornly high, and progress toward achieving the MDGs has been mixed. Although the country remains on track to achieve goals related to gender equality, infant and child mortality, and access to safe water, it is lagging on indicators related to basic education and maternal health. Although the under-five mortality rate declined from 40.4 per 1,000 live births in 2000 to 29.8 in 2012, the maternal mortality ratio remained stuck at 120 per 100,00 live births. Roughly one in four Filipinos (24 million people) continue to live below the national poverty line, most of them concentrated in rural areas (working in the agricultural sector or living in conflict-affected areas of Mindanao). In 2012 the incidence of food poverty (or national extreme poverty) was estimated at about 10 percent of the population, meaning that roughly 10 million people did not have sufficient income to meet their basic food requirements. About 10 percent of the population is at risk of falling into poverty and especially vulnerable to natural disasters or economic crises. Between 2003 and 2009, one in three of the poor was persistently poor, and two-thirds were transient (meaning they fluctuated in and out of poverty). As weather patterns shift the path of seasonal natural disasters, the poorest regions of the country are faced with increased vulnerability to shocks. In the second half of the decade, the economy was hit by multiple shocks, including the food and fuel price shocks, the global financial crisis, the global recession, and a series of deadly typhoons. The occurrence of the global financial crisis, the food and fuel crisis, and several highly destructive typhoons in 2008–09 was estimated to have increased poverty by nearly 4 percentage points, or an additional 3 million people. The economy has been resilient, however, rapidly recovering from these crises. The reason why growth in the Philippines has failed to translate into poverty reduction is puzzling. One of the major contributors to this mystery is that the stubbornly high level of income inequality in the Philippines (the Gini coefficient hovers at about 0.45) limits the growth elasticity of poverty reduction. Another primary constraint to translating growth into poverty reduction is the high fertility rate (particularly among the poorer segments of the population). Frequent natural disasters, including deadly typhoons that disproportionately hit poor regions, push vulnerable groups into poverty and jeopardize long-term human capital development. POVERTY DATA The Bank has provided considerable support to the Philippine Statistics Authority, contributing to the improvement of poverty data collection and management as well as the strengthening of the methodology for estimating poverty. Given the capacity constraints of the national statistical agencies, the necessary reforms on survey 154 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES design and data management have not been pushed through. The quality and timeliness of poverty data in the Philippines constrains the construction of poverty estimates as well as the monitoring of progress on poverty reduction. There are multiple sources of administrative and survey data available for constructing poverty estimates, including the Family Income and Expenditures Survey (FIES), the Labor Force Survey, and the Annual Poverty Indicators Survey. The FIES, which is the basis for computing poverty estimates, is carried out every three years. Mainly because of the limited capacity and fragmentation of the national statistical agencies, the design of the FIES survey has contributed to delays in data processing and release. The sharp decline in the number of staff in the national statistical agencies has exacerbated problems of timeliness: on average, it takes roughly two years after the completion of an FIES survey to clean the raw data, significantly undermining the timeliness of poverty estimates. There are also concerns about the quality and reliability of the FIES poverty data and the national account data. There are inconsistencies between the definitions of rural and urban in the 2003 survey round and later rounds. The lack of clarity in the methodology for establishing poverty lines, particularly related to regional price selection, renders the poverty trends not fully comparable over time. On the non-income poverty side, there is also variation in the availability and quality of data. Non-income poverty data at disaggregated levels (urban, rural, and agriculture) are particularly inadequate. Methodological problems with the construction of the national poverty lines have undermined the comparability, and thus quality, of constructed poverty statistics. Recognizing the constraints posed by quality issues, the Bank has been at the forefront of supporting the improvement of data quality in the Philippines. The Bank has full access to the raw survey data, albeit with large delays because of data- processing problems. Although there is a clear indication of appreciation from the government statistical agencies for Bank technical assistance and close collaboration at the technical level, real improvements in the timeliness and quality of poverty data are still lacking. DIAGNOSTICS Stakeholders, including representatives from the government, development partners, academia, NGOs, and civil society organizations, consider the Bank’s knowledge products on poverty to be of high quality. Overall, the quality of Bank poverty diagnostic work on the Philippines has been good, given the data quality and availability. 155 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES The Bank’s poverty diagnostic work generally made good use of the data available between 2005 and 2012, providing a reasonably complete picture of both income and non-income poverty. However, the absence of a full poverty assessment or poverty update during the entire period of evaluation limited the comprehensiveness and depth of the analytical work on poverty. Its absence, the lack of timeliness of survey data, and the delayed release of key pieces of Bank analysis limited the effectiveness of Bank poverty diagnostic work in contributing to policy dialogue. The latest poverty assessment was conducted in 2001, using 1997 FIES data. The poverty assessment planned for 2006/07 was included as a chapter in a report on inclusive growth and released only in 2010. In the absence of a full poverty assessment, the empirical underpinnings of the drivers of poverty and the unique challenges facing the extreme poor were limited in depth and breadth. Between the 2001 poverty assessment and the 2010 report on inclusive growth, poverty diagnostic work was embedded in many other pieces of analytical work produced by the Bank, including the annual Philippines Development Report, Discussion Notes, and Development Policy Updates. The Bank also conducted a series of PERs to examine the efficiency of public spending in priority areas and identify ways to improve pro-poor public actions. The most recent PER, prepared in 2011, highlighted the fact that data gaps limit the scope for analysis and the M&E of public revenues and expenditures. STRATEGY FORMULATION, IMPLEMENTATION, AND FEEDBACK LOOPS The Bank had two country strategies during the period of evaluation: the CAS for FY06–09 and the CAS for FY10–13. Both reflected the broad recommendations of the poverty diagnostics available at that time and were closely aligned with the strategies of the Philippine government. The focus in the FY10–13 CAS on the distributional aspects of growth was supported by considerable analytical underpinning, despite the absence of a formal poverty assessment or poverty update since 2001. However, the lack of timeliness of poverty data and its limited representativeness translated into similar issues with the diagnostic work, which to some extent undermined its usefulness, particularly for monitoring the effectiveness of Bank interventions. The absence of a poverty assessment or poverty update between 2001 and 2010 limited the scope and depth of poverty analysis that the FY06–09 CAS could draw from. However, as the challenges the country faced with respect to poverty reduction remained largely unchanged overall, the negative effect of the poor timeliness of poverty diagnostic work on the Bank’s country strategy formulation was limited. Seizing the window of opportunity opened by the government, the Bank focused primarily on supporting the piloting and scaling up of the conditional cash transfer 156 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES program and CDD programs, including the Kapitbisig Laban Sa Kahirapan- Comprehensive and Integrated Delivery of Social Service (KALAHI-CIDSS) CDD program. The Bank also supported the development of the National Household Targeting System for Poverty Reduction, which has become the main system of identifying the poor, providing objective information for the CCT, CDD, and national health projects. The Bank also supported several CDD programs that targeted conflict-affected areas of Mindanao including the Autonomous Region in Muslim Mindanao Social Fund, Mindanao Trust Fund for Reconstruction and Development, and the Mindanao Rural Development Program 2. In FY06–09 there were major deviations between the CAS program that was planned and that which was implemented. Only five of the 15 projects listed under the base- case lending scenario materialized. The bulk of the lending interventions occurred in response to government demands to shift from pure investment lending to sector wide national program support initiatives. One DPL, prepared in 2006, was implemented during this period; it was planned only as a possibility for the high lending scenario in the CAS. The results framework was clearer than in the previous CAS, with a stronger causal link to poverty. In FY10–13 there were major deviations between the planned and actual CAS programs, for both lending and nonlending. In the CAS, some 36 percent of lending was allocated to reducing vulnerability, 25 percent to macroeconomic stability, 21 percent to improvements in the investment climate, and 17 percent to public services. In actuality, two-thirds of the lending portfolio was allocated to macroeconomic stability, because of implementation of a series of DPLs and the Catastrophe DDO (CAT-DDO) in response to global economic crises and natural disasters. The Bank responded swiftly to support the Philippine government respond to natural disasters. The disbursement of both the emergency DPL and the CAT-DDO occurred with record speed, helping fund normal expenditures, including social expenditures. However, several stakeholders expressed concerns that the money was sitting at the national level and not used to support the poor in a timely manner. The Bank has provided considerable support to the Philippines concerning poverty data improvement, including data collection and management and the strengthening of the methodology for estimating poverty. Despite this support, the Bank has not been able to strongly influence the introduction of necessary reforms on data management and analysis. The lack of timeliness and quality of poverty statistics remains a constraint in monitoring the progress of poverty reduction. The Bank’s poverty diagnostic works were of high quality and provided strong analytical underpinnings for policy making and strategy formulation. The absence 157 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES of a full poverty assessment or poverty update since 2001 limited the depth of the work in certain aspects, including in identifying the drivers of poverty at a disaggregated level and tailoring recommendations to overcome obstacles. The CASs were closely aligned with government strategies. There were, however, significant deviations between envisaged and actual implementation of country programs. The M&E system improved both at the country program level and the project level, though it continues to focus on national level and intermediate outcomes. SUMMARY ASSESSMENT The Bank has provided considerable support to the Philippine Statistics Authority, contributing to the improvement of poverty data collection and management and strengthening the methodology for estimating poverty. Despite this support, the Bank has not been able to strongly influence the introduction of necessary reforms on data management or analysis. The timeliness and quality of poverty statistics remain a constraint in monitoring the progress of poverty reduction. The Bank’s poverty diagnostic works were of high quality and provided strong analytical underpinnings for policy making and strategy formulation. The absence of a full poverty assessment or poverty update since 2001, however, limited the depth of the work in certain aspects, including in identifying the drivers of poverty at a disaggregated level and tailoring recommendations to overcome obstacles. The CASs were closely aligned with the government’s strategies. There were, however, significant deviations between the envisaged and implemented country programs. The M&E system improved at both the program and the project level, though it continues to focus on national-level and intermediate outcomes. The Bank team in the Philippines has a strong sense of being a small player in a middle-income country with challenging political economy issues and deeply rooted vested interests. The main themes of the CASs—”islands of good governance” and seizing the window of opportunity to help “make growth work for the poor”— indicate the Bank team’s clear view of engagement in select areas. One of the key challenges the Bank faces is achieving an appropriate balance between supporting government priorities in reducing poverty and laying the groundwork to institutionalize difficult reforms needed to foster more sustained and inclusive growth. 158 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES Romania CONTEXT Romania experienced rapid growth during 2003–08, with GDP per capita growing by about 8 percent a year. However, following the severe downturn in 2008–09, the Romanian economy contracted sharply, and the pace of growth in 2009–13 has been very slow (0.2 percent a year in per capita terms). Romania’s GNI stood at $15,700 in 2011. The post-crisis recovery remains fragile and the outlook is challenging. With strong trade and financial sector linkages to the euro area, Romania has been and remains vulnerable to the regional economic slowdown. Although the economy is expected to grow at a moderately faster rate, growth momentum has been weak and lags most other emerging economies in Europe. Difficulties in absorbing structural funds from the European Union and frequent delays in advancing the structural reform agenda are weighing on the economy’s potential growth. Romania’s absolute poverty (national line) fell sharply from 30.6 percent of the population in 2001 to 4.4 percent in 2009. Despite this progress, the consumption deficit, which also had fallen, rose to more than 22 percent of the population. Inequalities remain an important issue, as the difference between poverty rates in rural and urban areas increased from 10 to more than 20 percentage points. Good progress was made in reaching the MDG targets. Between 2001 and 2009, the proportion of underweight children under the age of five fell by more than 50 percent, under-five mortality rate fell from more than 26 per 1,000 live births in 2000 to slightly more than 12 in 2012, and the maternal mortality ratio fell from 53 to 30 per 100,000 live births. In Romania, the second largest ethnic minority (and by some estimates the largest) is Roma. Most of the Roma are poor, vulnerable, and socially excluded. Among the non-Roma population of Romania, 31 percent are at risk of poverty; the figure for Roma is 84 percent. The secondary school completion rate for the Roma is only 10 percent, compared with 58 percent for non-Roma. A meaningful poverty reduction strategy must address the situation of the Roma. POVERTY DATA By 2006 Romania had a well-developed information base for poverty monitoring and analysis. It included credible household, living conditions, and labor force surveys as well as two censuses (2002 and 2011) that, combined with the household surveys, allowed for the estimation of poverty and living conditions at the local level. 159 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES The Bank has been involved in Romania’s data development since the 1990s. Its involvement began with the first poverty assessment, conducted in 1993, which provided an input into the development of data sets and methods of calculating absolute poverty when the first representative survey was implemented in 1994. Romania continued with the implementation of the survey. As new needs were identified, it added new modules, temporarily or permanently. The Bank assisted during the process. DIAGNOSTICS During the period covered by this evaluation, the Bank engaged in significant capacity-building efforts. In 2007–08 it prepared three poverty reports in close collaboration with the Ministry of Labor, Family, Social Protection, and Elderly Persons and the National Institute of Statistics. These reports included the 2007 poverty assessment and two 2008 reports, one on inclusion and social protection, the other on labor market vulnerabilities. The Bank supported the government with technical assistance throughout the preparation of these reports. Preparation of the 2007 poverty assessment included analytical workshops that focused on impact evaluation and evidence-based policy making and poverty measurement. The Bank’s credibility on technical issues, built over the years; its pioneering work on poverty as Romania transitioned from a centrally planned to a market-based economy; and the deepening of poverty as a social program were largely responsible for partnering with the government on poverty data, measurement, and policy issues. The Bank’s focus on increasing the volume and coverage of data and improving the targeting of social transfers is implicitly a recommendation to improve the equity of income after transfers. As its poverty assessments documented, the Bank paid attention to regional and urban/rural inequalities through multivariate analysis and poverty maps (showing high significance of geographic income disparities). Bank efforts on poverty monitoring covered the impact of the 2009 global crisis on poverty and focused increasingly on local conditions and exclusion. A joint Bank–UNICEF report provided a rapid assessment of the impact of the 2009 crisis. The Bank updated Romania’s poverty map with the use of 2011 census data. A recent report supported the development of poverty and inclusion indicators as the subnational level, including data on marginalized communities. These efforts responded partly to an increased focus on inclusion. The Bank’s diagnostic work was thorough and of good quality, contributing to credible and widely shared findings on the drivers of poverty. Both the 2003 and the 2007 poverty assessments conducted multivariate analyses of the predictors of adult-equivalent household consumption and reported the effects of geographic 160 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES (urban or rural area, region) and household conditions (including ethnicity, gender of household head, and other variables). The 2003 poverty assessment provided information on deprivation (nutrition, durables, housing); education, health, and employment; and social capital (from a specially designed survey). The poverty assessments also produced poverty maps. The Bank continued supporting the production of poverty and exclusion maps following these poverty assessments, using census data together with household surveys. The Bank assessed the adequacy of the country’s poverty reduction institutions, programs and funding, and poverty monitoring and evaluation arrangements. In early 2009, it prepared a set of policy notes for the government covering all sectors with a role in poverty reduction. Both in its poverty assessments and through its more direct advice to the government, the Bank proposed broad as well as specific and actionable recommendations to reduce poverty. The Bank’s broad approach was to combine measures to accelerate growth with measures to reach those vulnerable groups that would be unlikely to be reached by growth alone. The poverty assessments and sector AAA developed specific and actionable recommendations. Policies to reduce poverty through higher growth remained high on the list, with the 2004 CEM covering structural reforms across a wide range of areas. Several Bank reports covered policies on social protection, a focus of Bank support since its reengagement with Romania in the early 1990s, and developed actionable recommendations to improve the coverage, targeting, and integration of social protection arrangements. The Bank proposed actions to increase equity and quality in schooling, both regular and vocational. The Bank’s functional reviews probed deeply into delivery systems in these and other areas. The Bank earned its credibility in Romania partly by sharing international knowledge and working closely with local consultants and government line ministries and agencies. STRATEGY FORMULATION, IMPLEMENTATION, AND FEEDBACK LOOPS Through its past engagement, the Bank provided input into the government’s poverty strategies (1998, 2002, 2009). Good as the Bank diagnostic work has been, it must not obscure the contribution of other players. Before the review period of this report, the United Nations Development Programme and other United Nations agencies also supported poverty reduction efforts. The FY06–09 and FY09–13 CPSs, as well as the FY02–04 CAS and the FY14–17 CPS, consistently defined three pillars of overall Bank assistance with some variations in phrasing: growth, public sector, and poverty reduction and inclusion. Under the results chain articulated in the FY06–09 and FY09–13 CPSs, increased growth was to be achieved through privatization and improvements in the investment climate, the financial system, education and skills building, agricultural productivity, transport 161 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES (roads and rail), and energy. The FY09–13 CPS sought improved public sector performance through better public financial management and civil service administration. Both CPSs were to support poverty reduction and inclusion through improvements in social protection, programs for the most disadvantaged and vulnerable people, increased effectiveness of health services, and reductions in regional disparities. The employment dimension of the poverty reduction vision was to be achieved through increased growth (particularly nonagricultural growth) and some direct interventions to bolster the formal economy and improve employability. The CPSs’ broad results chain was reasonable, although perhaps more attention could have been afforded to addressing the jobless growth record of past performance. Through their three pillars, the CPSs were broadly consistent with the vision emanating from poverty diagnostics that identified growth, employment, and social protection as the critical targets of public policy for poverty reduction. This vision was underpinned by the poverty assessments’ diagnosis that growth was critical but not sufficient for poverty reduction, because it was unlikely to reach disadvantaged groups. Accordingly, a social protection component was critical for a comprehensive poverty reduction strategy. To achieve results under the road map to poverty reduction, interventions under the two CPSs relied on the broad results chain outlined above, although details in the two CPSs differed. Both CPSs aimed Bank inputs at similar outcomes: inclusion, improved living and social standards, human development, and reduction of urban, rural, and regional disparities. Under the FY06–09 CPS, the Bank planned three key lending inputs toward its poverty and inclusion objectives: a series of human development DPLs, a social inclusion loan, a rural and regional development loan, and a second mine closure project. The planned portfolio was complemented by AAA covering poverty, rural and regional development, and policy notes on the human development areas. Under the FY09–13 CPS, the Bank planned a DPL series that would reflect the intended content of the human development DPLs that had not materialized under the 2006 DPL. Planned AAA covered the same areas as in the FY06–09 CPS, except for rural and regional development. The country program underpinning the FY09–13 CPS appeared to have a stronger commitment to inclusion, possibly as a result of the effects of the crisis on the poor. The Bank accordingly included a social inclusion pillar. Toward the end of the CPS period, the government hired the Bank to prepare its poverty strategy. The FY06–09 CPS period was marked by a halt in Bank lending. After loan approval and effectiveness of 8 of 19 planned operations in 2006–07, little else took place on lending. As the CPSCR indicated, “Following accession and a change in the 162 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES governing coalition, interest in borrowing from the Bank waned rapidly … as Romania rapidly adjusted its financing strategy toward increased use of market finance, investment loans from the European Investment Bank and Structural and Cohesion grants from the EU [European Union].” Furthermore, continuing macroeconomic weaknesses, slippages in the reform agenda after accession, and lack of attention to strengthening institutional capacity derailed the program from a high- to a low-case lending scenario (as specified in the CPS). The quality of the portfolio declined as government interest in new loans weakened. The Bank remained engaged only through a low-case program of AAA. Bank engagement recovered under the FY09–13 CPS, as the 2009 global financial crisis raised Romania’s financial needs and reform stance. Therefore, the Bank implemented only part of its planned new lending under the FY06–09 CAS (but with some with delays), as well as preexisting projects and an unplanned operation. Preexisting projects the Bank implemented or closed during the two CPS periods accounted for the bulk of implementation. The Bank dropped 11 of the 19 planned new projects, including 3 human development DPLs, a rural and regional development project, two infrastructure projects (energy and transport), 3 programmatic policy loans, and a business environment project. Similarly, the Bank cut short the delivery of AAA under the FY06–09 CPS, after delivery of 7 of the 20 AAA it planned. Delivered activities with high poverty relevance included three poverty monitoring reports and a rural and regional technical assistance activity. Dropped activities with high relevance from the poverty perspective included a poverty monitoring note, a CEM, and three macroeconomic assessments. Unplanned delivered AAA with high poverty relevance covered (education and health), social protection, agriculture and rural development, and public expenditure. Planned and delivered lending picked up during the FY09–13 CPS period. The Bank planned and approved three DPLs with human development components and a health sector reform operation. It also approved an unplanned Social Assistance System Modernization Project. Planned AAA (28 activities) under the FY09–13 CPS was generally delivered; it covered most poverty-related sectors (poverty monitoring, social protection, health, and education). At the end of the CPS period, the Bank was implementing several reimbursable technical assistance activities covering a broad range of areas. Other projects had no obvious poverty reduction effects but could be argued to have some poverty impact. The 2006 Transport Sector Support Project, for example, while not a poverty-focused operation, is expected to provide access to schools, health, and jobs, especially in rural areas, with road and railway works generating employment. The CPS did not monitor those effects. 163 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES Evidence on sustainability is spotty. Some investment projects, such as the Social Fund and Municipal Services projects, indicate continued reliance on external funding. In the case of the Social Fund, such funding came from a different donor after the Bank project closed. DPLs that supported reforms with good project ratings may prove to have sustainable development outcomes. Two poverty reduction areas deserving more attention are skills development and employment and regional development. The Bank has explored these issues in the past, but it has not achieved the kind of engagement demonstrated in health or social protection. Efforts on employment, mainly analytical, emphasized the need to strengthen formal labor markets through tax and regulatory changes; promote economic growth through changes in the investment climate; and improve the links between education/training and employment. Efforts on regional development have yet to develop a good paradigm for action, although recent work on cities and “growth poles” shows promise. These efforts need Romania’s own efforts to develop policies in these areas, where much remains to be done. The Bank monitored overall poverty reduction outcomes closely, through its work with the Ministry of Labor, covering both the national and municipal levels (through poverty maps). Evidence from the poverty assessments and other diagnostic work informed the design of its programs. This evidence includes evidence on poverty drivers as an underpinning for support to disadvantaged groups; evidence on the relationship between growth and distribution, to underpin a strong emphasis on the growth pillar as a way of reinforcing more direct poverty reduction efforts; and more recently, evidence on the Roma, to underpin a progressively stronger focus on this population. The FY06–09 CAS articulated poverty-related targets from its projects under its Targeting Poverty Reduction and Promoting Social Inclusion pillar. These targets covered extreme poverty, living conditions at the local level and for the Roma population, delivery of early childhood education in selected communities, the quality of services to persons with disabilities, the youth employment rate, medical care of the poor, and the quality of infrastructure in rural and economically depressed areas. Under its inclusion pillar, the FY09–13 CPS monitored project targets, including Roma living conditions, the payment of benefits under the guaranteed minimum income (GMI) program and adequacy of GMI benefits, and the targeting of social assistance. The CPSs also monitored other poverty-related indicators, such as the costs of means-tested programs, consolidation of social assistance programs, and the efficiency and quality of health services. Except in follow-up projects (for example, the 2005 Mine Closure Project, which scaled up 164 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES activities under its predecessor), it is unclear how the Bank used the project monitoring data. CPSPRs and CPSCRs referred to the CPS indicators and provided updates and comments on progress with regard to the indicators. There is room for improvement of monitoring poverty outcomes at the project level. With the databases that the Bank has developed at the local level, the poverty focus of projects could be more systematically measured, even for projects that are not designed with a poverty focus. More systematic measurement of poverty impacts could strengthen the feedback loops that help inform the poverty content of new programs and operations. Overall, the Bank responded to changing conditions. Its programs changed as a result of Romania’s accession to the EU, in particular by supporting Romania on EU issues with analytic activities linked to its programs. Nevertheless, planned DPLs under the FY06–09 CPS did not go forward, as government interest in borrowing from the Bank waned. Nevertheless, the Bank responded to the effects of the 2009 global crisis, under its FY09–13 CPS, by finding an opportunity to provide DPL support that addressed both Romania’s financing needs and reform priorities. Its reform priorities included poverty-related issues, including health, education, and social protection, as the crisis briefly raised poverty numbers. The feedback loops were strong in response to findings from analytical work, perhaps less so from project M&E. Feedback loops from AAA worked in Romania primarily because the Bank has been a credible counterpart that built its reputation on poverty issues over the years with a strong record of work on data, poverty measurement, and poverty diagnostic issues. It also helped that the Bank gained traction on policy advice because of its role as an impartial observer in a very fluid political environment. The Bank could perhaps learn more about poverty by strengthening the poverty-related M&E of projects and extracting more lessons from its project experience. SUMMARY ASSESSMENT The Bank already had a long-standing engagement on poverty in Romania by the time it prepared its FY06–09 CPS. In 1991, following a nine-year hiatus in support to Romania, the Bank prepared social protection and health services operations, beginning a long-term engagement in these sectors. As poverty was not acknowledged, let alone measured, by Romania at the time, the Bank provided support on data collection and poverty indicators, partly to underpin its support for social transfer arrangements. As the Bank identified economic growth as a key driver of poverty reduction, Bank support for structural reforms also deserved attention from the poverty reduction perspective. 165 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES The Bank monitored overall poverty reduction outcomes in Romania closely through its work with the Ministry of Labor, which covered both the national and municipal levels (through poverty maps). Evidence from the poverty assessments and other diagnostic work informed the design of its programs. Evidence was gleaned from the following areas: poverty drivers as an underpinning for support to disadvantaged groups; the relationship between growth and distribution to underpin a strong emphasis on the growth pillar as a way of reinforcing more direct poverty reduction efforts; and more recently, focus on the Roma to underpin a progressively stronger focus on this sizeable minority group. The FY06–09 CAS articulated poverty-related outcome targets from its projects under its Targeting Poverty Reduction and Promoting Social Inclusion pillar. They included the project targets and covered extreme poverty, living conditions at the local level and for the Roma, delivery of early childhood education in selected communities, the quality of services to people with disabilities, the youth employment rate, medical care of the poor, and the quality of infrastructure in rural and economically depressed areas. Under its inclusion pillar, the FY09–13 CPS monitored project targets, including Roma living conditions, the payment of benefits under the guaranteed minimum income (GMI) program and the adequacy of the GMI benefits, and the targeting of social assistance. The CPSs also monitored other poverty-related indicators, such as the costs of means-tested programs, the consolidation of social assistance programs, and the efficiency and quality of health services. There is room for improvement in monitoring poverty outcomes at the project level. With the databases that the Bank has developed at the local level, the poverty focus of projects could be more systematically measured, even for projects that are not designed with a poverty focus. More systematic measurement of poverty impacts could strengthen the feedback loops that help inform the poverty content of new programs and operations. The Bank responded to changing conditions, in particular by supporting Romania on European Union issues with analytic activities linked to its programs. Nevertheless, planned DPLs under the FY06–09 CPS did not go forward, as government interest in borrowing from the Bank waned. The Bank did respond to the effects of the 2009 global crisis, under its FY09–13 CPS, by finding an opportunity to provide DPL support that addressed both Romania’s financing needs and reform priorities. The latter included in particular poverty-related issues, including health, education, and social protection, as the crisis briefly raised poverty numbers. The feedback loops were strong in response to findings from analytical work, perhaps less so from project M&E. Feedback loops from AAA worked in 166 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES Romania primarily because the Bank has been a credible counterpart that built its reputation on poverty issues over the years with a strong record of work on data, poverty measurement, and poverty diagnostic issues. It also helped that the Bank earned traction on policy advice thanks to its role as an impartial observer in a fluid political environment. Perhaps more could be learned by the Bank (and the client) about poverty by strengthening the poverty M&E in projects and by drawing lessons from the experience. There is scope for strengthening the M&E of the poverty impacts from projects. Several projects that were not explicitly poverty focused had likely effects on poverty reduction. Senegal CONTEXT Following the devaluation of the CFA franc in 1994, the Senegalese economy experienced high levels of economic growth for a decade. From 1995 to 2005, annual GDP growth averaged 4.5 percent, and inflation remained in check. Growth peaked at about 6 percent in 2003 and 2004, thanks mainly to favorable weather conditions, external environment, and domestic policies. Since 2006 the country has been buffeted by a series of domestic shocks (poor rains in 2006–07, floods and droughts in 2009 and 2012) and external shocks (world food and fuel price shocks in 2008, the global financial crisis of 2008–09). These shocks have put Senegal in a low-growth equilibrium, with average annual GDP per capita growth of only 1.3 percent between 2003 and 2013. GNI per capita stood at $1,940 in 2011. Senegal has a population of about 10 million, almost half of whom are poor. The proportion of the population living below the national poverty line decreased from 67.9 percent in 1994 to 48.3 percent in 2005–06 and 46.7 percent in 2011. The level of extreme or food poverty remained relatively unchanged throughout the entire period at 15 percent. Because of continued growth in the population (estimated at 2.5 percent a year), the number of people living below the national poverty line in Senegal has increased since 2001. Senegal has made some progress toward other MDG targets. Between 2000 and 2012, the prevalence of undernourishment declined from 24.4 percent to 21.6 percent, primary school enrollment rose from 57.4 percent to 73.3 percent, under- five mortality rate declined from 139 to 59.6 per 1,000 live births, and the maternal mortality ratio fell from 480 to 360 per 100,000 live births. Income distribution is highly unequal, with the Gini index estimated to have fallen from 39.2 in 2001 to 37.4 in 2005 and 37.8 in 2011. The lowest income quintile in 2011 167 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES accounted for 5.2 percent of total consumption while the highest income quintile accounted for 50 percent. Senegal’s Human Development Index was 0.459 in 2012— about average for Sub-Saharan Africa (0.463), ranking it 155 out of 187 countries. POVERTY DATA Poverty data in Senegal are generally of high quality and have continuously improved over time, thanks in part to the support provided by the Bank. The main limitations to income poverty data are that they are collected only once every five years; they are not sufficiently disaggregated (for example, to the village level); and raw data are not accessible to the public. The Bank (and other donors) put in place a capacity-building effort at the national statistical agency that supported the collection of income poverty data, the implementation of household consumption and poverty surveys, and the preparation of reports that present the outcome of the surveys. Non-income poverty data are abundant, but they are dispersed across the various ministries without any central depository. The National Statistics and Demographics Agency (NSDA) is responsible for household survey design and data collection in Senegal. Three principal surveys were conducted during the period under review: the 2001/02 Senegal Household Survey (ESAM II), the 2005/06 Senegal Poverty Monitoring Survey (ESPS I), and the 2011 Senegal Poverty Monitoring Survey (ESPS II). The surveys cover the entire country. The data can be broken down by area, allowing for the construction of a poverty profile. Beginning in 2008, survey data are available at the administrative unit level, which for the first time permits the creation of a poverty map. There are also limitations to the survey data: the three surveys are not fully comparable, the poverty indicators are estimated on the basis of consumption rather than income, the data are collected only once every five years, and there were long lags between the completion of the surveys and the availability of the results. Sector ministries conduct their own non-income surveys regarding sectoral data, often in collaboration with regional and international organizations. Perception surveys capture households’ perceptions of poverty. The NSDA is well staffed and well financed. Household surveys and poverty surveys are well documented and broadly in line with international standards. Survey reports are available on the NSDA website, but the underlying micro data are not readily available for use by other government officials, development partners, and civil society. The NSDA is also responsible for overseeing the quality of non-income poverty data collected by ministries, but it has yet to create a central depository for such data. 168 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES The Bank and other donors have been strengthening government statistical capacity through the Trust Fund for Statistical Capacity Building. The ANSD (National Agency of Statistics and Demography) financed throughout the period under review. Bank staff have been working closely with the ANSD regarding both the survey methodologies and the preparation of survey analyses and poverty diagnostics. DIAGNOSTICS Since the onset of the PRSP process in Senegal, in 2001, the government entity in charge of poverty monitoring has prepared good-quality poverty diagnostics and assessments, undertaken in close collaboration with the Bank. More recently, Bank support for diagnostics (including inputs for the PRSP process, poverty assessments, and poverty notes) was stepped up, with joint Bank and government poverty diagnostics underpinning preparation of the country’s poverty reduction strategies. Despite these efforts, there is no evidence that the government used the diagnostics to craft action plans for future poverty work, the preparation of subsequent PRSPs, or the design and targeting of specific interventions. There is also no evidence of linking the diagnostic to the selection of any particular intervention. Both government officials and other donors concurred that the Bank was the preeminent and lead expert regarding poverty analyses and hence used the Bank diagnostics for internal purposes. Representatives from civil society, however, complained that they were unaware of or did not have access to these diagnostics, underscoring an apparent communication gap, as these documents were in the public domain. The Bank carried out three main poverty diagnostics during the period under review: the 2004 poverty report (“La Pauvreté au Sénégal”); the 2008 poverty report (“Sénégal: Diagnostic de la Pauvreté”); and a collection of poverty notes prepared by Bank staff in 2011 (the overview diagnostic document is entitled “Poverty, Inequality, and Gender”). The 2004 report, a joint report by the Bank and the government, provides comparable estimates of poverty using survey data from the 1994/95 ESAM I and the 2001/02 ESAM II. The 2008 poverty report, also carried out jointly by Bank staff and government counterparts, provides an overview of poverty in Senegal, an analysis of employment, and a diagnostic of the education and health sectors using the 1994/95 ESAM I, the 2001/02 ESAM II, and the 2005/06 ESPS I. It develops a poverty profile and for the first time a poverty map, with data disaggregated down to the local community level. The 2013 study “Poverty, Inequality, and Gender: An Overview” uses data from the 2011 ESPS II. It derives a poverty line consisting of both a food and a nonfood 169 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES component and includes an assessment of inequality. Its main findings include the following: diagnostics have tended to be completed “just in time” before the Bank country team require it as an input into CAS and CPS preparation; the methodology for deriving Senegal’s poverty indicators is well documented and in line with international standards; development partners relied on Bank poverty diagnostics for their country strategies, but civil society appears to be unaware of the diagnostics available for its use; the poor can be characterized as people living in rural areas without access to local basic infrastructure; with a large number (10 or more) of household members; and with a head of household who is a man with a primary or lower level of education and seeking employment or economically inactive. The 2004 diagnostic was mainly descriptive. The 2006 and 2011 diagnostics conducted additional and deeper analysis of policy options available to the government. They did not rank or prioritize the policy options. None of the diagnostics discussed explicitly the impact of growth and income distribution on poverty reduction or discuss obstacles to poverty reduction or identify the constraints to address these obstacles. They did, however, discuss gender-specific components of poverty in the context of the poverty profile and the drivers of poverty. STRATEGY FORMULATION, IMPLEMENTATION, AND FEEDBACK LOOPS Although the available poverty data and poverty diagnostics underpinned the Bank’s strategy, there is no evidence that the activities proposed under the strategy were directly linked to or evolved from the diagnostics. Although the CAS/CPS used the information provided by the household surveys and diagnostics as inputs to their descriptive background chapters, they did not draw on the diagnostic’s findings in justifying its activities. Despite the weak link between diagnostics and strategy formulation and implementation, aspects of the Bank portfolio (such as rural and agricultural sector development) were in areas identified in the diagnostic work as being where the poor are located. Three CAS and CPSs were prepared during the period under review: the CAS for FY03–06, the CAS for FY07–10, and the CPS for FY13–17. The FY03–06 CAS was aligned with Senegal’s first PRSP. It focused on wealth creation, capacity building and social services, assistance to vulnerable groups, and implementation of the strategy and monitoring of its outcomes. The FY07–10 CAS was fully aligned with Senegal’s second PRSP. The FY13–17 CPS supported Senegal’s priorities as presented in its National Economic and Social Development Strategy for 2013–17 (or PRSP III), which builds on the government’s political program (the Yonnu Yokute), the joint action platform of civil society (the 170 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES Assises Nationales), and the earlier Accelerated Growth Strategy. Its three main pillars are growth, productivity, and wealth creation; human capital and sustainable development; and governance, institutions, peace, and security. The Bank’s assistance and partnership strategies for Senegal during the review period drew on the poverty data and poverty diagnostics available at the time. The poverty profile remained broadly the same, with a higher incidence of poverty in households that were large; rural; and headed by a man with low education who worked in the agriculture sector but who was currently unemployed or looking for work. The formulated strategies had a strong focus on results and articulated a reasonable results chain linking IDA interventions to desired outcomes to higher- level country goals. They presented a logical sequence of the problems facing the country and the expected outcomes of IDA involvement, including linking the proposed projects and AAA to the specific outcomes being sought. The Bank assistance strategies were aligned with Senegal’s PRSPs and selective, focusing on areas of Bank expertise and comparative advantage and given other donor interventions. However, the CAS and CPS discussion of trends in poverty incidence and the poverty profiles were largely descriptive. No attempt was made to link the proposed set of interventions to the findings of the diagnostic. They did not discuss whether or how the proposed strategy and set of interventions evolved from the poverty data and diagnostics. They did discuss targeting the poor and the most vulnerable. The CAS and CPS adopted a two-tier approach to poverty reduction. Income poverty was to be addressed through a set of interventions to stimulate private investment and private sector d under the growth and wealth creation pillars. Non- income aspects of poverty were to be addressed through activities in support of pillars on improving public service delivery, particularly in basic education, health services, and water and sanitation. None of the Bank strategy documents identified gaps in knowledge about poverty arising from the poverty data and diagnostics or proposed filling any such gaps. Implementation of Bank interventions was broadly consistent with the available poverty data and diagnostics as well as the Bank’s strategy. Projects that focused on rural areas, basic education, and the vulnerable parts of the population were particularly consistent with the poverty data. Overall, the Bank’s programs and projects were implemented as planned and hence reflected the strategic priorities for poverty reduction as outlined in the CAS and CPS, even though the feedback loops from data to diagnostics to strategy formulation and implementation in Bank country strategies for Senegal were weak. 171 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES Deviations from the interventions proposed in Bank strategies largely reflected support measures aimed at cushioning the impact of the global financial crisis and new government priorities. Although the CAS and CPS were prepared with input from the Bank’s poverty and sector specialists, the country management unit largely drove the process, and the specialists felt that they had little impact on decisions. There were also concerns over the dominant role of the country directors and the regional vice president regarding the direction and selection of Bank activities. Although other stakeholders (financial and technical partners, civil society organizations, and others) participated in discussions with Bank staff regarding the CAS and CPS, these discussions were more informational than consultative. There is a general perception by both Bank staff and development partners that once the Bank’s strategy was completed, it was placed on the shelf and it was “back to business as usual until the next progress report or completion report.” The CASCR and the IEG CASCR Review of the CAS FY03–07 concluded that its implementation was moderately satisfactory. On wealth creation, efforts to improve the climate for private sector development yielded mixed results. The Bank was successful in improving the private sector development environment, but progress was slow on measures involving entrenched interests, customs, and development of institutions. Bank support to improve public expenditure management contributed to better budget management, but it was undermined by off-budget spending and other special arrangements. Good progress was made on the capacity-building and social services pillar on extending access to primary and secondary education and improving gender parity in primary and secondary education, but little was achieved with regard to efficiency, quality, or learning outcomes. Bank support in health had mixed results, but considerable progress was made on improving access to water and sanitation. Progress in improving the living conditions of the vulnerable was positive, thanks to the Social Fund and rural infrastructure projects. Nonlending services were delivered roughly as planned, with emphasis in the early part of the period on core diagnostics to underpin planned use of the subsequent DPL instruments. IEG rated the overall outcome of the FY07–10 CAS as moderately unsatisfactory, in line with the CASCR rating. Most objectives were only partially achieved, and some were not achieved at all. Under the growth pillar, little progress was made toward removing key transport and energy bottlenecks, and the overall business environment deteriorated, despite some promising regulatory reforms. Under the human development pillar, good progress was made in broadening access to education and closing the gender gap, but there was no indication of improvement in quality. In health and nutrition, there was a large disconnect between the positive results at the project level and the disappointing outcomes. A child-focused social 172 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES cash transfer project achieved good results, but there was no information on the welfare of street children. Most of the nonlending activities were delivered, albeit with some delay. Overall, there was no significant change in the poverty focus of Bank CAS, CPS, and lending and nonlending activities during implementation, and the proposed set of interventions envisaged in the Bank’s programs was broadly consistent with the results chain in the CAS and CPS. New and unplanned activities arose largely because of new priorities of the government and in response to domestic and external shocks. Only some of these unplanned activities had an explicit poverty focus. Budget support operations were key instruments for poverty reduction, but there was no planned dialogue or conditionality in budget support operations regarding the reallocation of public spending toward programs reaching the poor. The Bank effectively leveraged its assistance, by financing pilot projects, co- financing Bank projects with other donors, and financing projects with other Bank Group entities (IFC and the Multilateral Investment Guarantee Agency). As the draft report for Senegal points out, during the period covered by this evaluation, the Bank provided strong support to Senegal for generating good- quality poverty data and diagnostics. Senegal’s government appeared to be committed to poverty reduction and made good use of the technical support provided by the Bank (and other donors) for its poverty data, poverty diagnostics, and preparation of its PRSPs. However, the feedback loops from data to diagnostics to strategy formulation and implementation in Bank country strategies were weak, and there is a widespread belief that the Bank’s strategy was not closely monitored by the government or the Bank’s own staff. SUMMARY ASSESSMENT There was no significant change in the poverty focus of Bank CASs and CPSs and lending and nonlending activities during implementation. The proposed set of interventions envisaged in the Bank’s programs was broadly consistent with the results chain in the CAS and CPS. New and unplanned activities arose largely because of new priorities of the government or in response to domestic and external shocks. Only some of these unplanned activities had an explicit poverty focus. Budget support operations were key instruments for poverty reduction, but there was no planned dialogue or conditionality in budget support operations regarding the reallocation of public spending toward programs reaching the poor. The Bank effectively leveraged its assistance, by financing pilot projects, co- financing Bank projects with other donors, and financing projects with other Bank Group entities (IFC and MIGA). Implementation of Bank interventions was broadly 173 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES consistent with the available poverty data and diagnostics as well as the Bank’s strategy, as there was no significant change in the composition of the implemented portfolio. Projects that focused on rural areas, basic education, and the vulnerable parts of the population were particularly consistent with the poverty data. During the period covered by this evaluation, the Bank provided strong support to Senegal for generating good-quality poverty data and diagnostics. The government appeared to be committed to poverty reduction and made good use of the technical support provided by the Bank (and other donors) for its poverty data, poverty diagnostics, and preparation of its PRSPs. However, the feedback loops from data to diagnostics to strategy formulation and implementation in Bank country strategies have been weak, and there is a widespread belief that the Bank’s strategy was not closely monitored by the government or the Bank’s own staff. Table A.1. Selected Development Indicators for the 10 Case Study Countries 2013 2003–13 Population Population GNI per Poverty Poverty Population Average (millions) density capita Headcount Headcount growth GDP per (people (constant ratio 2005 ratio 2005 rate capita per sq. km 2011 PPP PPP and PPP and (% per growth of land $) 1.25/day $2.00/day annum) (%) area) poverty poverty line (% of line (% of pop.) pop.) High income 1,054 147.9 37,076a [ 11.0]b 0.6 1.3 OECD (32.9 )c countries Developing 5,818.4 74.3 8,164.5 17.0d 36.3d 1.3 6.4 countries (31.0b) 10 Country 613.8 233.7d 7,080.3d 32.9d 50.0d 1.8d 3.7d Case Studies (61.4 )b Bangladesh 156.6 1,203.0 3,082 43.3e 76.5c 1.2 4.9 Egypt 82.1 82.4 10,443 1.7 f 15.4 f 1.6 2.7 Guatemala 15.5 144.3 6,901 13,7 29.8e 2.5 0.9 Lao PDR 6.8 29.3 4,402 30.3 a 62.0 a 1.8 5.7 Malawi 16.4 173.6 730 72.2c 88.1c 2.8 2.4 Nigeria 173.6 190.6 5,166 62.0 e 82.2 c 2.8 6.0 Peru 30.4 23.7 10,821 2.9 a 8.0 a 1.3 5.1 Philippines 98.4 330.0 7,598 19.0a 41.7a 1.7 3.5 Romania 19.9 86.8 18,410 g 0.0 a 1.6 a -0.6 4.4 Senegal 14.1 73.4 2,143 34.0d 60.3e 2.9 1.3 Source: World Bank, Word Development Indicators, 2014. a. information for 2012, no information for Slovak Rep. and Slovenia. b. Based on OECD definition: percent of population with household incomes of less than 50 percent of OECD median income (or less than PPP $25/day). c. Arithmetic averages. 174 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES d. Information for 2011. e. Information for 2010. f. Information for 2008 g. GNI per capita, PPP (current international $). Table A.2. Selected Millennium Development Goals for Developing Countries and the 10 Case Study Countries, 2000–2012 School Enrollment Primary (net, %) Mortality Rate, under-5 (per 1,000 live births) 2000 2005 2012 Percentage 2000 2005 2012 Percentage point point change change (2000– (2000– 2012) 2012) Developing 82 86 88 +6 84 69 52 -32 countries Middle 86 89 90 +4 71 59 45 -26 Income Countries Low Income 62 77 83 +21 135 109 79 -56 Countries Bangladesh 92 92a 88 67 43 - 45 Egypt 94 94 95b +1 45 31 22 - 23 Guatemala 86 94 93b +7 51 41 32 -19 Lao PDR 75 79 96 +21 117 97 74 -43 Malawi 174 121 71 -103 Nigeria 65 67 64a -1 188 159 122 -66 Peru 98 97 94b -4 40 28 18 -22 Philippines 89 40 36 31 -9 Romania 85 95 86 +1 27 21 13 -14 Senegal 57 70 73 +16 137 98 58 -79 Source: World Bank, Word Development Indicators, 2014. a. Information for 2010. b. Information for 2011. Table A.3. GDP Per Capita Growth, Poverty Headcount Ratio and Gini Coefficient in the 10 Case Study Countries 2003–2013 (percent per annum or percentage) 2003–08 2009–13 Countries GDP per Poverty head Gini coef. GDP Poverty head Gini coef. capita count ratio (earliest year per capita count ratio (latest year growth (earliest year available) growth (latest year available) (2003–2008) available) (2009–2013) available) Bangladesh 4.9 40.0a 33.2a 5.0 31.5b 32.1b Egypt, Arab 3.7 19.6a 32.1a 1.5 21.6c 30.8d Rep. Guatemala 1.4 51.0e 54.1f 0.3 53.7g 39.0g Lao PDR 5.5 27.6h 35.5i 6.0 27.6* 36.2j Malawi 2.6 52.4k 39.9k 2.3 50.7l 46.2l Nigeria 8.6 48.4k 40.0k 2.9 46.0l 42.9l Peru 5.7 58.7k 53.8f 4.3 23.9 45.3j 175 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES Philippines 3.5 24.9f 44.5f 3.5 25.2j 43.0j Romania 7.9 24.8 e 29.9 f 0.3 22.6 g 27.3j Senegal 2.1 48.3a 39.2a 0.3 46.7g 40.3g Source: World Bank, World Development Indicators 2014. Notes: Poverty head counts are based in each countries national poverty line; the information from Lao PDR is from the 2013 Expenditure and Consumption Survey. GDP per capita growth in Nigeria in 2004 is reported as 30.3 percent. a. Information for 2005. b. information for 2010. c. information for 2009. d. information for 2008. e. information for 2006. f. information for 2003. g. information for 2011. h. information for 2008 i. information for 2007. j. information for 2012. k. information for 2004. l. information for 2010. 176 APPENDIX A SUMMARIES OF 10 COUNTRY CASE STUDIES Table A.4. Measures of Inequality in the 10 Case Study Countries: Gini and Palma Country Survey Type Base Gini Palma Pov. Pov. Mid- Gini Palma Pov. Pov. Latest Gini Palma Pov. Pov. Year base ratio head head year mid- ratio head head year latest ratio head head year base count count year mid- count count year latest count count year ratio ratio year ratio ratio year ratio ratio base base mid- mid- latest latest year year year year year year ($1.25) ($2.50) ($1.25) ($2.50) ($1.25) ($2.50) Bangladesh Consumption 1991 27.6 0.997 70.2 96.7 2000 33.5 1.361 58.6 90.8 2010 32.1 1.272 43.3 86.2 Egypt, Arab Consumption 1990 32.0 1.261 4.5 36.9 1999 32.8 1.323 1.8 37.6 2008 30.8 1.194 1.7 32.0 Rep. Guatemala Income 1989 59.6 5.975 28.4 50.7 2000 54.8 4.189 11.8 33.9 2011 52.4 3.594 13.7 40.5 Lao PDR Consumption 1992 30.4 1.169 55.7 91.5 2002 32.5 1.289 41.2 85.2 2012 36.2 1.555 30.3 75.2 Malawi Consumption 1997 50.3 3.159 83.2 96.1 2004 39.9 1.867 74.9 94.4 2010 46.2 2.574 72.2 92.4 Nigeria Consumption 1992 44.9 2.563 61.9 86.9 2003 40.0 1.863 61.8 89.6 2010 42.9 2.189 62 88.4 Peru Consumption 1994 44.9 2.514 9.81 32.4 2000 50.9 3.481 12.5 30.8 2012a 45.3 2.435 2.9 11.6 Philippines Consumption 1991 43.8 2.284 33.2 68.7 2000 46.1 2.578 24.6 58.3 2012 43 2.189 18.9 53.4 Romania Income 1992 25.5 0.856 0.4 2.4 2001 29.4 1.069 2.5 22.3 2012b 27.3 0.945 0 3.9 Senegal Consumption 1991 54.1 4.089 65.7 86.8 2001 41.3 0.712 44.1 80.7 2011 40.3 1.901 34.1 86.8 Source: Povcalnet, as of Nov. 24, 2014. Palma ratio calculated on the basis of Povcalnet data. a. Indicates income data. b. Indicates consumption data. 1 The IBRD, IDA, and blend country classification is up to date as of October 2013. 2The government mainly focused on its own development priorities, which were tilted towards large infrastructure projects and did not prioritize the “inclusiveness” of the economic growth process. 3 See Demographic and Health Surveys, Ministry of Health and Population, 2003, 2005, 2008. 177 Appendix B. External Stakeholder Survey on the World Bank’s Support for Poverty Reduction As part of the evaluation, the Independent Evaluation Group (IEG) commissioned an independent survey firm, ICF International, to conduct a stakeholder survey in 20 client countries. To ensure the authenticity of responses, the company was instructed to conduct the survey in a confidential and anonymous manner sharing only the cumulative results with IEG. ICF International communicated to the respondents in its emails and follow-up calls that their individual responses would not be shared with IEG or the World Bank. The survey countries were selected to represent six regions of the world with consideration of the balance between the types of countries by lending (i.e., International Bank for Reconstruction and Development, IBRD, International Development Association, IDA/Blend), accessibility of data, and fragile and non- fragile countries (FCS and non-FCS). The final list of survey countries included China, Democratic Republic of Congo, Dominican Republic, Ethiopia, Ghana, Haiti, Honduras, India, Jordan, Kosovo, Kyrgyz Republic, Mexico, Morocco, Nepal, Papua New Guinea, Russia, Rwanda, Sierra Leone, South Africa, and the Republic of Yemen. For the distribution of selected countries by type and the corresponding response rates, please refer to the lists of survey countries by type and survey responses by country (See tables B.6 and B.7). The survey targeted several groups including government officials, civil society, academia, donor/international community, and private sector. The heavier weight in the respondent list was given to government officials as they are World Bank’s primary clients. The list of respondents was randomly generated from the list of respondents to the Bank’s recent country client surveys and IEG evaluation interview lists, and was complemented by the research of the independent survey firm. To improve the representativeness and size of the sampling frame and to add independently found respondent names, the survey firm supplemented the initial respondent lists. The original sampling frame contained 4,619 contact names with email addresses from which the stratified sample of 100 names in each country was drawn to be the final respondent list. In most cases, each country list included a mix of approximately 60 government and 40 other stakeholder respondents. Within the “other” stakeholder selection, the goal was to have equal representation of each group from donors, academia, and private sector. To the extent possible these criteria were followed, 178 APPENDIX B EXTERNAL STAKEHOLDER SURVEY ON BANK SUPPORT FOR POVERTY REDUCTION with the emphasis on having each stakeholder group represented. Where the numbers in the sampling frame exceeded the number required, an analyst from the survey firm randomly selected names within each target group. Additional names were available in many countries to replace the emails that bounced back. Thus, in cases where respondents’ emails bounced back and it was not possible to verify appropriate contact information, the names of individuals were substituted with additional names available in the same group of stakeholders. This was the only time when the names of respondents were substituted by additional names. The survey was fielded via email through the survey firm that followed up with multiple email reminders and individual phone calls in each country. The survey questionnaires were translated into the relevant language for each country. Every individual that had an email address and phone number listed received at least five reminders and calls. Among the five countries with most government respondents were China, the Democratic Republic of Congo, Honduras, Jordan, and Mexico. The top five countries for non-government stakeholder respondents were Ethiopia, Kosovo, Dominican Republic, Nepal, and Rwanda. A 27 percent response rate is on a par with similar stakeholder surveys administered by the World Bank and other international organizations. 1 In their research, Stoop et al. (2010)2 show that the absence of nonresponse bias and the actual size of bias depend not only on the response rate but also on the difference between respondents and nonrespondent groups, among other things. Other researchers conform with this view stating that using response rates alone as a way of judging the quality of a survey can be misleading. A high response rate can still introduce a sizable bias in an estimate 3. Some literature suggest a distinction between “active” nonrespondents and those who did not respond due to situational factors, which means that they may not be objective to the specific survey topic or surveyor or any other survey related issue. 4 Given that the number of respondents to IEG’s external survey was high enough (over 540 people), the evaluation team analyzed the respondent demographics to ensure that they were not deviating from the overall demographics of the population list. There was no particular group of stakeholders that was predominantly absent in the responses. Nonrespondent rates varied from country to country more significantly rather than in comparable groups of respondents among countries. The evaluation team also did not find “active” nonrespondent bias among those who did not respond as conversations with the independent survey firm showed that the nonrespondents were mostly those who were not reachable by phone for individual reminders and who had situational factors such as public holidays (in the Middle East and North Africa Region), political situation in the country (e.g. Russia), etc. 179 APPENDIX B EXTERNAL STAKEHOLDER SURVEY ON BANK SUPPORT FOR POVERTY REDUCTION Further, to ensure the validity of responses, the evaluation team calculated the implied error of responses at a 95 percent confidence level (z =1.96) and used it to benchmark the responses to each question. As such, the implied error of responses at the 95 percent confidence level is 3.6 percent. Overall Profile of Respondents The total number of respondents was 542 people, for a response rate of 27 percent across all targeted countries. Among all of the respondents, 305 respondents were government representatives and 237 were the representatives of other stakeholder groups. Government respondents most frequently were involved with the Bank through negotiations or discussions about loans or technical assistance (72 percent) or through reading some of the Bank’s poverty-related reports (60 percent). The primary areas of focus of government officials’ work were infrastructure, agriculture/rural development, and finance (each representing about 13 percent of respondents). More than half of all respondents had worked in their field for more than 10 years (55 percent). Reflecting on the possible areas of involvement, the greatest percentage thought that the World Bank focuses most of its attention in their respective countries on infrastructure, including urban, transport, water, and sanitation (46 percent). The distribution of the respondents in other stakeholder groups included civil society (27 percent), donors (23 percent), academia (23 percent), private sector (16 percent), and other (11 percent) 5. These respondents most frequently were involved with the Bank through reading some of the Bank’s poverty-related reports (75 percent) or through participation in Bank-initiated discussions on poverty (52 percent), which is not as “intense” an engagement as indicated by government respondents. The primary areas of focus of their work were education, public sector development, or “other.” About two-thirds of this group had worked in their field for more than 10 years. Reflecting on the possible areas of involvement, the greatest percentage also thought that the World Bank focuses most of its attention on infrastructure, including urban, transport, water, and sanitation (40percent). Constraints to Data Availability and Use Stakeholders only occasionally use data on poverty that are available from the World Bank’s website and publications. Around quarter or less of the respondents mentioned that they frequently use poverty data from the World Bank’s website and publications. As such, only 18 percent of government stakeholders and 26 percent of all other stakeholders indicated that they frequently used Bank sources for poverty 180 APPENDIX B EXTERNAL STAKEHOLDER SURVEY ON BANK SUPPORT FOR POVERTY REDUCTION data. The majority of the respondents identified that they occasionally use Bank sources for poverty data, with government officials indicating so in 49 percent of cases and other stakeholders in 48 percent. While 33 percent of government stakeholders and 26 percent of other stakeholders rarely or never use Bank sources for poverty data. The World Bank adds value to the improvement of the quality of data on poverty. Around 86 percent of government and 90 percent of non-government respondents agreed somewhat or strongly that the Bank adds value to improvement of the quality of data on poverty. Of all the government respondents only four percent disagreed somewhat with this statement and less than one percent disagreed strongly. Similarly, among non-government respondents only five percent disagreed with this statement and none disagreed strongly. Furthermore, the majority of government respondents (85 percent) agreed strongly that the Bank’s analysis of poverty data is beneficial for their agency’s work, with less than 6 percent indicating that they disagreed with this statement. Government stakeholders also agreed that the Bank has sufficient data on poverty to develop its country strategies in their countries. As such 72 percent 6 of respondents indicated that they “somewhat” or “strongly” agreed with the statement that the bank has sufficient poverty data to develop its strategies, with 31 percent agreeing strongly. Government officials identified insufficient budget to collect data (64 percent) and lack of regular household surveys (56 percent) as two primary constraints, closely followed by insufficient government capacity (53 percent) to obtaining data on poverty in the respective countries (see table B.1). Government officials most frequently identified insufficient budget to collect data (64 percent) among obstacles to obtaining poverty data. There were no large variations among IBRD, IDA/Blend, FCS and Non-FCS countries in responses to this category. However, the comparison of responses from government officials who use the data on poverty from the Bank’s website and publications “frequently” and who responded to the question about data constraints show that the frequent users of the data identify insufficient budget to collect data more often as a constraint to obtaining data than any other category (30 percent). Table B.1. What are the primary constraints in obtaining data on poverty? (responses from government officials, in percentages) Insufficient capacity to collect data 53 Insufficient budget to collect data 64 181 APPENDIX B EXTERNAL STAKEHOLDER SURVEY ON BANK SUPPORT FOR POVERTY REDUCTION Insufficient time to collect data 14 Lack of regular household surveys 56 Lack of incentives to collect data 33 Other 10 Don’t Know 4 Not applicable 5 Total Number of Respondents 279 In the category “Lack of regular household surveys,” 27 percent of respondents were from the IDA/Blend countries and 15 percent were from the IBRD countries. Responses varied between the respondents from the FCS vs non-FCS countries in the category “Lack of regular household surveys,” where 32 percent of respondents from the FCS countries chose it as an obstacle vs. 19 percent of respondents from the non-FCS countries. The two-tailed test of responses to this category shows a statistical significance of means between the respondents from the IBRD/IDA countries, FCS, and non-FCS countries, and between the respondents from the “data weak” vs. “data rich” countries. At the same time, 53 percent of government officials cited insufficient government capacity to collect data as an obstacle, which was the third most cited constraint. Government officials from the IDA/Blend countries cited this constraint in 24 percent of cases, while those from IBRD countries cited it in 18 percent of cases. The corresponding two-tailed test shows that there is a difference of means in the scale of this question between IBRD and IDA/Blend countries. There was no statistical variation and large percentage variation in this category between the responses from the FCS and Non-FCS countries. Poverty Diagnostic Work The stakeholders consider the Bank’s analytical work useful for developing their government’s poverty reduction policies and programs. All of the listed products (i.e., Poverty Assessments; Poverty Assessments and Poverty and Social Impact Analyses, PSIAs; and Public Expenditure Reviews, PERs) received more than half of the positive responses ranging between somewhat and to a great extent, while the most positive feedback was received for PSIAs (73 percent). Government officials believe “to a great extent” in 32 percent of cases that PSIAs are useful in developing their ministry’s or agency’s programs and policies for reducing poverty, while they express the same sentiment about the Public Expenditure Reviews in 25 percent of cases. There was some variation between the government respondents to this 182 APPENDIX B EXTERNAL STAKEHOLDER SURVEY ON BANK SUPPORT FOR POVERTY REDUCTION question from the IDA/Blend and IBRD countries. As such, 78 percent of respondents from the IDA/Blend countries agreed that PSIAs are “somewhat” or “to a great extent” useful in developing poverty reduction-related policies and programs in their agencies, while only 65 percent of the respondents from the IBRD countries mentioned so. Similarly, around 71 percent of respondents from the IDA/Blend countries mentioned Poverty Assessments being useful to “somewhat” or “a great extent” vs. respondents from the IBRD countries. In regards to PERs, the respondents from the FCS countries were more likely to say that the PERs were useful than respondents from the non-FCS countries with 65 percent and 56 percent, respectively (agreeing strongly in 32 percent and 22 percent of responses, respectively.) In further analysis around 49 percent of respondents who indicated they are “fully aware” and 48 percent who indicated that they are “somewhat aware” of the poverty assessments also found the Bank’s diagnostic work on poverty useful to their ministry’s or agency’s work to “a great extent.” The percentages are calculated with the exclusion of those who indicated that they are not aware of poverty assessments or they are not applicable. See table B.2 for more information. 7 Table B.2. To what extent have the World Bank Poverty Assessment been useful in developing your ministry's or agency's programs and policies for reducing poverty? (percent of respondents) Indicate your level of awareness of the World Bank’s Poverty To a great Not Assessments extent Somewhat Very little Not at all applicable Don’t know Fully Aware 45 27 20 25 14 5 Somewhat 44 61 57 50 52 52 Aware Not Aware at 6 8 17 25 19 24 All Not Applicable 1 4 3 0 14 5 No Answer 4 4 3 0 0 14 The vast majority of both government and non-government respondents were positive about the quality of the Bank’s diagnostic work on poverty and its recommendations. The responses were similar between the two groups and across various categories. A detailed analysis of this question shows that around 75 percent of government respondents and 73 percent of non-government stakeholders mentioned that Poverty Assessments and PSIAs provided an identification of key constraints to poverty reduction; 75 percent of government and 69 percent of non- 183 APPENDIX B EXTERNAL STAKEHOLDER SURVEY ON BANK SUPPORT FOR POVERTY REDUCTION government stakeholders believe that these products provide actionable recommendations; and 73 percent of government and 74 percent of non-government stakeholders believe that the products provide a well-grounded analysis. Slight variations included that government officials most frequently agreed “to a great extent” that the Bank’s Poverty Assessments and PSIAs helped identify constraints to poverty reduction (30 percent government vs. 27 percent of non-government), while other stakeholders most frequently agreed “to a great extent” the that Bank’s products provided a well-grounded analysis of poverty (22 percent of government vs. 30 percent of non-government). Government officials believe that, overall, Bank instruments address the poverty focus in their country. All of the three listed instruments (Development Policy Operations, DPOs; Investment Lending, IL; and Analytical Works) were deemed to address the Bank’s poverty focus “somewhat” or “to a great extent” in the 70 to 78 percent range, although 11 to 14 percent of respondents don’t know or did not answer. As such, DPOs were believed to address the poverty focus in some or great extent by 74 percent of respondents, ILs by 71 percent of respondents, and Analytical Works (AW) by 78 percent of respondents. Overall, ILs had higher percentages of officials believing they address the poverty focus “to a great extent,” with 38 percent of respondents indicating so vs. the other two instruments, where the frequency was 32 percent. Further analysis shows that, in general, IDA/Blend countries more frequently mention that DPOs somewhat or to a great extent address the Bank’s focus on poverty (77 percent) vs. respondents from the IBRD countries (68 percent). Additionally, the IDA/Blend countries more frequently believed that the DPOs address poverty focus than IBRD countries with 81 percent of the responses vs. 70 percent, respectively. For more details see table B.3. Table B.3. To what extent do the following instruments address the World Bank’s focus on poverty in your country? (Percent of respondents) IDA/Blend IBRD IDA/Blend IBRD DPO IL AW DPO IL AW Very Little 9 14 8 8 8 8 Somewhat 46 34 45 35 29 46 To a Great Extent 31 34 35 33 43 27 Not At All 0 3 0 2 2 2 Don't Know 5 7 5 17 11 13 Not Applicable 2 2 2 4 3 3 184 APPENDIX B EXTERNAL STAKEHOLDER SURVEY ON BANK SUPPORT FOR POVERTY REDUCTION Country Strategies and Projects Looking at Bank products, the government respondents were more frequently “fully aware” of the Country Partnership Strategy (37 percent) than poverty analytical work or data. Overall, government respondents also were more “fully aware” about country strategies than other stakeholders (27 percent). The Bank needs to increase its focus on the poorest segments of population. Overall, a total of 66 percent of government officials and 63 percent of non-government stakeholders agreed with the statement “strongly” or “somewhat” that the World Bank’s country strategy in their respective countries focuses on the poorest segments of the population, with less than a third (27 percent) of government officials and less than a quarter (16 percent) of other stakeholders agreeing strongly with the statement. There were no significant variations between government respondents from the IDA/Blend vs. IBRD countries and FCS vs. non-FCS countries. However, both government and other stakeholders disagreed somewhat or strongly between 25 and 28 percent. Government and other stakeholders from the IDA/Blend countries were disagreeing more frequently than those from the IBRD countries. As such, government respondents disagreeing with this statement in 63% percent of cases were from the IDA countries, while 37 percent were from the IBRD countries. Among the other stakeholders who disagreed with the statement, 73 percent were from the IDA/Blend and 27 percent were from the IBRD countries. More government respondents from non-FCS countries disagreed than those from the FCS countries with a 74 percent vs. a 26 percent frequency. Similarly, more non- government stakeholders from non-FCS countries disagreed than from the FCS countries with a 70 percent vs. a 30 percent frequency. When government officials were asked whether they agree that the benefits of Bank- funded projects continue after the projects are completed, around 66 percent agreed with the statement, with only a quarter of the respondents agreeing strongly. Furthermore, another quarter of the respondents disagreed with the statement and nine percent of respondents indicated they did not know an answer to the statement. There were no large variations in the responses agreeing with the statement between the government respondents from IDA/Blend and IBRD countries, but those from the IDA/Blend countries more frequently disagreed with the statement (29 percent) than those from the IBRD countries (17 percent). The biggest gap in disagreement range was observed in the “disagree strongly” category, where 11 percent of IDA/Blend countries chose it vs. three percent from the IBRD countries. At the same time, respondents from the FCS countries were more likely to disagree with the statement (39 percent) than those from the IBRD 185 APPENDIX B EXTERNAL STAKEHOLDER SURVEY ON BANK SUPPORT FOR POVERTY REDUCTION countries (20 percent), with the larger gap in the “disagree strongly” category that had 18 percent vs. four percent, respectively. FEEDBACK LOOPS The Bank’s coordination between donors and the government needs further strengthening. A little over half (57 percent) of government respondents agreed strongly or somewhat that the World Bank coordinates priorities between donors and the government, while 27 percent disagreed strongly or somewhat with this statement. When donors were asked whether the World Bank coordinates priorities between the donors and government, only 53 percent agreed with the statement, of which only 11 percent agreed strongly. At the same time, around 44 percent disagreed with the statement, of which 13 percent disagreed strongly. This question had the highest total percentages of people disagreeing with the statement both among the government officials and the donors. The analysis of the negative tale of the responses shows that government officials and donors from IDA countries were more likely to disagree with this question than those from other types of countries. As such, 72 percent of government officials disagreeing with this statement were from IDA/Blend countries, while 28 percent were from IBRD. Similarly, donors responding negatively to this question were 88 percent from IDA/Blend and 13 percent from IBRD. More respondents from non-FCS countries disagreed with this statement than those from FCS countries. Among government respondents disagreeing with the statement, 65 percent were from the non-FCS countries and among the donors 67 percent. The Bank seeks feedback more frequently from donors compared to other stakeholder groups. Both government officials and other stakeholders responding to the survey felt that the World Bank seeks feedback while developing its country strategies more frequently from donors (including bilateral and multilateral organizations) than other listed stakeholder groups. Overall, 71 percent of government respondents and 67 percent of other stakeholders believed that the Bank seeks feedback from donors to “some” or a “great extent” from donors. The next most frequently mentioned group was civil society, with 65 percent of government respondents and 64 percent of other stakeholders choosing this category. Both government officials and other stakeholders indicated less frequently that the Bank seeks feedback from academia (with 56 percent of government officials agreeing and 52 percent of other stakeholders agreeing) and private sector actors (with 54 percent of government officials agreeing and 53 percent of other stakeholders agreeing). For more details, please see table B.4 and B5). 186 APPENDIX B EXTERNAL STAKEHOLDER SURVEY ON BANK SUPPORT FOR POVERTY REDUCTION Table B.4. To what extent does the World Bank seek feedback from the following group when developing its partnership strategy in your country? (Percent of government respondents) To a great Some Little Not at Don’t Not extent extent Extent all know applicable Donors 34 37 8 1 17 1 Academia 18 38 20 4 18 0 Civil Society 22 43 16 1 15 0 Private Sector 13 39 18 4 22 0 Other 5 12 8 1 25 7 Table B.5. To what extent does the World Bank seek feedback from the following group when developing its partnership strategy in your country? (Percent of non-government respondents) To a great Some Little Not at Don’t Not extent extent Extent all know applicable Donors 37 30 10 0 18 0 Academia 17 37 18 3 22 1 Civil Society 22 42 18 2 13 1 Private Sector 16 37 21 2 21 0 Other 6 7 3 0 19 6 The issue of better coordination was mentioned in responses to the open-ended question on the lessons the World Bank can use to help government to strengthen the poverty focus of its policies or strategies. Around 7 percent of comments from government stakeholders answering the question mentioned better coordination among donors and government as an important lesson. One of the comments was: “Our government has a strong focus on poverty reduction. The support that the [World Bank] could give is more in ensuring that reliable data are available to support the analysis as well as the monitoring of the efforts. [The World Bank] could also play a better role in helping coordinate/facilitate the discussions between donors and recipient country government[s] on better policies to reduce poverty. It seems that often the [World Bank] acts like just another donor and also that it does not like to get involved in strategic policy change discussions.” In 24 percent of comments, government respondents mentioned the importance of ensuring more local ownership and beneficiary engagement in developing projects and strategies. The comments highlighted the necessity for the Bank to have a “bottom-up” approach to developing poverty-related strategies and policies to ensure greater ownership and sustainability. Some comments included: 187 APPENDIX B EXTERNAL STAKEHOLDER SURVEY ON BANK SUPPORT FOR POVERTY REDUCTION “It is essential to include local governments and communities in the development of policies to ensure the sustainability of the actions arising from the implementation of these policies. It is important to establish a system to monitor the impact of actions taken on behalf of the population in poverty with sufficient visibility to allow these people to demand accountability.” “The discussion of strategies should be included with civil society activists, parliamentarians, people from the villages, [and] the beneficiaries, in order to receive the full and objective information.” “Support is needed during the entire project process, develop population-based project impact assessments, and develop projects from the bottom-up, as part of a policy of ‘closeness’ and an approach that is participatory.” In 81 percent of cases, government respondents mentioned that they agree somewhat or strongly that the lessons from projects implemented by the Bank and other development agencies inform the development of the Bank’s country strategies. Among those who agreed, 34 percent agreed with the statement strongly. There were no significant variations between the respondents from the IDA/Blend vs. IBRD and FCS vs. non-FCS countries. Table B.6. List of Survey Countries (by Country Types) Borrower status FCS status Country IDA/Blend IBRD FCS Non-FCS Ghana X X Ethiopia X X Nepal X X India X X Papua New Guinea X X Kosovo X X South Africa X X Dominican Republic X X Mexico X X Honduras X X Morocco X X Sierra Leone X X Rwanda X X Haiti X X Congo, Dem. Rep. X X Russia X X Kyrgyz Republic X X Jordan X X Yemen, Rep. X X China X X 188 APPENDIX B EXTERNAL STAKEHOLDER SURVEY ON BANK SUPPORT FOR POVERTY REDUCTION Table B.7. List of Survey Completions by Country Government Other Country Completes Completes Total China 39 11 50 Congo, Dem. Rep. 22 16 38 Dominican Republic 10 18 28 Ethiopia 17 22 39 Ghana 15 6 21 Haiti 17 7 24 Honduras 23 8 31 India 5 11 16 Jordan 23 8 31 Kosovo 15 21 36 Kyrgyz Republic 13 13 26 Mexico 25 12 37 Morocco 5 2 7 Nepal 18 21 39 Papua New Guinea 12 7 39 Russia 8 6 14 Rwanda 7 18 25 Sierra Leone 8 6 14 South Africa 10 8 18 Yemen 13 17 30 Total Responses 305 237 542 QUESTIONNAIRE FROM THE SURVEY WITH GOVERNMENT OFFICIALS 1. Please characterize your primary involvement with what the World Bank does in your country Options Percent I have participated in negotiations and/or discussions about World Bank 72 loans and/or technical assistance I have analyzed or provided feedback on the World Bank's knowledge 40 products I have participated in World Bank-initiated discussions on poverty 41 I have read some of the World Bank's poverty-related reports 60 I have not engaged in any of the activities listed above 0 Total Responses 305 189 APPENDIX B EXTERNAL STAKEHOLDER SURVEY ON BANK SUPPORT FOR POVERTY REDUCTION 2. What field is the primary focus of your work? Options Percent Agriculture and rural development 13 Education 7 Energy 3 Environment, natural resources, and climate change 7 Finance 13 Infrastructure (including urban, transport, water, and sanitation, etc.) 13 Health, nutrition, and population (including HIV/AIDS, pandemics, etc.) 5 Public sector development (including governance and anticorruption) 9 Private sector development 2 Macroeconomic and fiscal policy 8 Labor and social protection (including pensions, social safety nets, etc.) 7 Other (including gender or other cross-cutting issues. Please specify.) 14 Total Responses 305 3. How many years have you worked in this field? Options Percent Less than 3 years 8 Between 3 - 10 years 36 More than 10 years 55 Total Responses 305 4. In which of the following areas do you believe the World Bank focuses most of its attention in your country? Options Percent Agriculture and rural development 32 Education 30 Energy 25 Environment, natural resources, and climate change 26 Finance 16 Infrastructure (including urban, transport, water, and sanitation, etc.) 46 Health, nutrition, and population (including HIV/AIDS, pandemics, etc.) 22 Public sector development (including governance and anticorruption) 23 Private sector 6 Macroeconomic and fiscal policy 18 Labor and social protection (including pensions, social safety nets, etc.) 14 Other (including gender or other cross-cutting issues) 3 Don’t know 1 Total Responses 305 190 APPENDIX B EXTERNAL STAKEHOLDER SURVEY ON BANK SUPPORT FOR POVERTY REDUCTION 5. Please indicate the extent of your awareness of the following World Bank products for your country. (Percent of respondents) Somewhat Not aware at Not Fully aware aware all applicable Country Partnership Strategy 37 49 10 2 Analytical work (such as country poverty 30 53 10 2 assessment, etc.) Poverty data 32 51 8 3 Total Responses 305 6. In your work, how often do you use data on poverty that are available from the World Bank's website and publications? Would you say...? 8 Options Percent Frequently 18 Occasionally 49 Rarely 24 Not At All 9 Total Responses 305 7. Please indicate the extent to which you agree or disagree with the following statements as they apply to your country. (Percent of respondents) Agree Agree Disagree Disagree Do not Not Options strongly somewhat somewhat strongly know applicable I can easily access data on 28 47 12 2 4 5 poverty that are related to my work The World Bank adds value 40 46 4 0 6 2 to the improvement of the quality of data on poverty The World Bank’s analysis of 39 46 4 1 2 6 data on poverty is beneficial for my agency’s work The World Bank has 31 41 12 2 10 3 sufficient data on poverty to develop its country strategy Total Responses 279 191 APPENDIX B EXTERNAL STAKEHOLDER SURVEY ON BANK SUPPORT FOR POVERTY REDUCTION 8. What are the primary constraints in obtaining data on poverty in your country? Options Percent Insufficient capacity to collect data 53 Insufficient budget to collect data 64 Insufficient time to collect data 14 Lack of regular household surveys 56 Lack of incentives to collect data 33 Other 10 Don’t know 4 Not applicable 5 Total Responses 279 9. To what extent have the following World Bank products been useful in developing your ministry's or agency's programs and policies for reducing poverty? (Percent of respondents) To a great Don’t Not Options extent Somewhat Very little Not at all know applicable Poverty Assessments 32 34 12 2 7 9 Poverty and Social 32 41 10 1 7 7 Impact Analyses Public Expenditure 25 33 15 3 10 9 Reviews Total Responses 304 10. To what extent do the following instruments address the World Bank’s focus on poverty in your country? (Percent of respondents) To a great Do not Not Options extent Somewhat Very little Not at all know applicable Development Policy 32 42 9 1 10 3 Operations Investment Lending 38 33 12 2 9 2 Analytical Works 32 46 8 1 8 2 Total Responses 304 192 APPENDIX B EXTERNAL STAKEHOLDER SURVEY ON BANK SUPPORT FOR POVERTY REDUCTION 11. Please indicate the extent to which you agree or disagree with the following statements as they apply in your country. (Percent of respondents) Agree Agree Disagree Disagree Don’t Not strongly somewhat somewhat strongly know applicable The World Bank’s country strategy focuses on the poorest 27 39 19 6 6 2 segments of the population. Based on my experience, the benefits to the poor of World 25 41 17 8 9 0 Bank-funded projects continue after the projects are completed Lessons from projects implemented by the World Bank and other development 34 47 4 3 10 0 agencies inform the development of the World Bank’s country strategies Total Responses 301 12. To what extent has the World Bank’s diagnostic work on poverty, such as Poverty Assessments and Poverty and Social Impact Analyses, provided the following in your country? (Percent of respondents) To a great Some Little Not at Don’t Not extent Extent Extent all know applicable Well-grounded analysis of 22 52 11 1 10 2 poverty Identification of key constraints 30 46 9 2 9 1 to poverty reduction Actionable recommendations 25 51 10 2 10 1 to reduce poverty Total Responses 301 13. When other donors provide direct support to reduce poverty in your country, the World Bank coordinates priorities between these donors and your government. Would you say that you...? Options Percent Agree Strongly 20 Agree Somewhat 38 Disagree Somewhat 20 Disagree Strongly 7 Not Applicable 1 Don't Know 14 Total Responses 300 193 APPENDIX B EXTERNAL STAKEHOLDER SURVEY ON BANK SUPPORT FOR POVERTY REDUCTION 14. To what extent does the World Bank seek feedback from the following groups when developing its partnership strategy in your country? (Percent of respondents) To a Some Little Not at Don’t Not great Extent Extent all know applicable Donors (including bilateral and 34 37 8 1 17 1 multilateral organizations) Academia (including research 18 38 20 4 18 0 institutes and think tanks) Civil society (including 22 43 16 1 15 0 international and locally-based nongovernmental organizations) Private sector 13 39 18 4 22 0 Other (Please specify ) 5 12 8 1 25 7 Total Responses 300 15. Please share one or two lessons that the World Bank can use to help your government to strengthen the poverty focus of its policies or strategies. Count Total number of open-ended responses 244 QUESTIONNAIRE FROM THE SURVEY WITH DONORS, ACADEMIA, CIVIL SOCIETY AND PRIVATE SECTOR 1. Please characterize your primary involvement with what the World Bank does in your country Options Percent I have participated in negotiations and/or discussions 34 about World Bank loans and/or technical assistance. I have analyzed or provided feedback on the World 30 Bank's knowledge products. I have participated in World Bank-initiated discussions 52 on poverty. I have read some of the World Bank's poverty-related 75 reports. I have not engaged in any of the activities listed 0 above. Total Responses 237 194 APPENDIX B EXTERNAL STAKEHOLDER SURVEY ON BANK SUPPORT FOR POVERTY REDUCTION 2. What field is the primary focus of your work? Options Percent Agriculture and rural development 8 Education 10 Energy 6 Environment, natural resources, and climate change 6 Finance 7 Infrastructure (including urban, transport, water, and 8 Health, nutrition, and population (including HIV/AIDS, 6 Public sector development (including governance 10 Private sector development 8 Macroeconomic and fiscal policy 8 Labor and social protection (including pensions, 6 Other (including gender or other cross-cutting issues. 17 Total Responses 237 3. How many years have you worked in this field? Options Percent Less than 3 years 8 Between 3 - 10 years 30 More than 10 years 64 Total Responses 237 4. Please indicate your professional affiliation: Options Percent Donor (including bilateral and multilateral 23 Academia (including research institutes and think 23 Civil society (including international and locally based 27 Private sector 16 Other (Please Specify) 11 Total Responses 237 195 APPENDIX B EXTERNAL STAKEHOLDER SURVEY ON BANK SUPPORT FOR POVERTY REDUCTION 5. Please consider the following statement and indicate your level of agreement. When other donors provide direct support to reduce poverty in the country in which you work, the World Bank coordinates priorities between these donors and the government Options Percent Agree somewhat 42 Agree strongly 11 Disagree strongly 13 Disagree somewhat 31 Don't know 4 Not applicable 0 Total Responses 55 6. Considering the country in which you work, in which of the following areas do you believe the World Bank focuses most of its attention? Options Percent Agriculture 35 Education 25 Energy 26 Environment, natural resources, and climate change 16 Finance 19 Infrastructure (including urban, transport, water, and sanitation, etc.) 40 Health, nutrition, and population (including HIV/AIDS, pandemics, etc.) 19 Public sector development (including governance and anticorruption) 29 Private sector development 16 Macroeconomic and fiscal policy 24 Labor and social protection (including pensions, social safety nets, etc.) 12 Other (including gender or other cross-cutting issues). 4 Don’t know 1 Total Responses 235 7. Please indicate the extent of your awareness of the following World Bank products for the country in which you work: (Percent of respondents) Fully Somewhat Not aware Not aware aware at all applicable Country Partnership Strategy 27 56 14 1 Analytical Work (such as country poverty 33 57 8 0 assessment, etc.) Poverty data 39 48 11 1 Total 237 196 APPENDIX B EXTERNAL STAKEHOLDER SURVEY ON BANK SUPPORT FOR POVERTY REDUCTION 8. In your work, how often do you use data on poverty that are available from the World Bank's website and publications? Would you say...? Options Percent Frequently 26 Occasionally 48 Rarely 19 Not At All 7 Total Responses 237 9. Please indicate the extent to which you agree or disagree with the following statements as they apply to your country. (Percent of respondents) Agree strongly Agree Disagree Disagree Do not Not Options somewhat somewhat strongly know applicable I can easily access data on 38 45 10 1 3 1 poverty that are related to my work. The World Bank adds value to 43 47 5 0 4 0 the improvement of the quality of data on poverty. Total Responses 219 10. To what extent has the World Bank's diagnostic work on poverty, such as Poverty Assessments and Poverty and Social Impact Analyses, provided the following in the country in which you work? (Percent of respondents) To a Not great Some Little Not at Don’t applicabl Options extent Extent Extent all know e Well-grounded analysis of poverty 30 44 12 2 8 1 Identification of key constraints to poverty 27 47 14 1 8 0 reduction. Actionable recommendations to reduce 18 52 18 1 7 1 poverty Total Responses 234 234 197 APPENDIX B EXTERNAL STAKEHOLDER SURVEY ON BANK SUPPORT FOR POVERTY REDUCTION 11. Please indicate the extent to which you agree or disagree with the following statements as they apply to the country in which you work. (Percent of respondents) Do Agree Agree Disagree Disagree not Not Options strongly somewhat somewhat strongly know applicable The World Bank has contributed to 31 49 13 3 4 0 an understanding of the causes of poverty. The World Bank's country strategy 16 47 23 6 7 0 focuses on the poorest segments of the population. Total Responses 232 12. To what extent does the World Bank seek feedback from the following groups when developing its partnership strategy in your country? (Percent of respondents) To a great Some Little Not at Don’t Not extent Extent Extent all know applicable Donors (including bilateral and 37 30 10 0 18 0 multilateral organizations) Academia (including research institutes 17 37 18 3 22 1 and think tanks) Civil society ( including international 22 42 18 2 13 1 and locally based nongovernmental organizations) Private sector 16 37 21 2 21 0 Other (Please specify ) 6 7 3 0 19 6 Total Responses 231 1For example, the World Health Organization’s Internal Survey conducted as part of the Stakeholder Perception Survey received a 25 percent response rate. The World Bank’s 2013 Access to Information Stakeholder Survey had a 27 percent response rate and the Client Feedback Survey of FY13 Analytical & Advisory Activities had a 31 percent response rate. 2Ineke Stoop, Jaak Billiet, Achim Koch, Rory Fitzgerald . Improving Survey Response: Lessons Learned from the European Social Survey. 2010. P.5. 3 Bautista, R. Handbook of Survey Methodology for the Social Sciences. 2012. P.43. 4 Stoop, I. Handbook of Survey Methodology for the Social Sciences. 2012. P.142 198 APPENDIX B EXTERNAL STAKEHOLDER SURVEY ON BANK SUPPORT FOR POVERTY REDUCTION 5Some of the mentions in the “Other” category still included references to working in government or international/development agencies. Some respondents in this category also indicated that they worked for more than one category i.e., public sector and private sector or government and international institution. 6The percentages of responses are calculated based on the total number of respondents in each group (i.e. government and other stakeholders) which included the missing values. 7Please note that cross-tabulated data in this table is calculated proportional to responses in each category and analyzed by column 8 Pleasenote that those respondents who chose the option “Not at all” in this question, were redirected to Question 9, skipping the next two questions. 199 Appendix C. Staff Survey on the World Bank’s Support for Poverty Reduction The Independent Evaluation Group (IEG) conducted a survey among World Bank staff to capture their perceptions of best practices and constraints to obtaining poverty data, creating poverty diagnostic work, and translating analytic work into country strategies and policies. The survey was conducted between April 15 and May 13, 2014, and was sent to 4,150 Bank staff, of whom 866 responded (a rate of 21 percent). The universe of 4,150 respondents was narrowed to exclude those working on services that are not directly related to country strategies or operations. Among the excluded units were procurement, human resources, information and technology, business solutions, the World Bank Institute, and IEG. 1 The full list of excluded units and staff titles is provided at the end of the appendix. Since the evaluation team narrowed the list down to include staff working on operational and poverty related issues, it was decided not to sample the list further. A similar technique is used in a number of Bank-administered staff surveys and the response rate (21 percent) is within the range of similar surveys. 2 The survey probed three main issues related to the Bank’s poverty work: (i) constraints to obtaining poverty data; (ii) best practices and challenges to creating poverty diagnostic work; and (iii) challenges in translating poverty diagnostic work into country strategies. Survey limitations are described in box C-1. Box C.1. Limitations of the Survey The results of the internal survey are subjective and based on staff perception rather than factual evidence. The following aspects need to be considered when interpreting the survey’s results.  There is self-selection bias among respondents, which exists in all surveys where respondents are given an option to choose whether or not to respond to the survey.  There was an overrepresentation of staff self-identified as knowledgeable about certain countries. As such, roughly six percent of survey respondents self-identified as most knowledgeable about Indonesia, while just over five percent self-identified as most knowledgeable about India. Other countries identified by a significant number of respondents include: Brazil (3.1 percent), Vietnam (2.9 percent), Bangladesh (2.8 percent), Ethiopia (2.7 percent), Nigeria (2.4 percent), Pakistan (2.4 percent), Tanzania (2.4 percent), Kenya (2.3 percent), and China (2.2 percent). Over 80 percent of respondents self-identified as mapped to one of the six Regions and only 16 percent as mapped to the networks. The highest percentage of respondents self-identified as mapped to the Africa Region. 200 APPENDIX C STAFF SURVEY OF WORLD BANK’S SUPPORT FOR POVERTY REDUCTION  Since the survey was not randomly sampled and was sent to the universe of respondents, it does not have a sampling error and did require confidence level testing. Despite the survey limitations mentioned in box C.1., the comparison of the distribution of population and that of the respondents by mapping, location, and grade levels shows that there is no significant variation between the groups. The differences were in 1 and 2 percent between each of the compared groups. For more details, please refer to figures C.1. and C.2. This shows, overall the respondent group is not significantly different from the population chosen for this evaluation. Figure C.1. Comparison of Survey Population and Respondents by grade Levels( Percent of respondents) 47% 48% 50% 45% 40% 35% 27% 28% 30% 25% 23% 25% 20% 15% 10% 5% 0% GF GG GH + Population Survey 201 APPENDIX C STAFF SURVEY OF WORLD BANK’S SUPPORT FOR POVERTY REDUCTION Figure C.2. Comparison of Survey Population and Respondents by Mapping ( Percent of respondents) 25% 24% 20% 20% 15% 14% 12% 12% 12% 12% 11% 10% 10% 10% 8% 7% 6% 6% 5% 5% 3%3% 2% 3% 3% 2% 1% 0% AFR EAP ECA LCR SAR MNA OPCS HDN SDN PREM FPD Population Survey For the purposes of the analysis, some responses were analyzed by Bank country type classification to find differences between countries that are either International Bank for Reconstruction and Development (IBRD) or International Development Association (IDA)/blend and fragile and conflict-affected states (FCS) or non-FCS. Some of the survey results were also analyzed across countries that have no or only one Household Income and Expenditure Survey and Living Standard Measurement Survey versus those with more than one survey conducted between 2000 and 2012 and available in the central data platforms of Development Economics (such as the Micro data Library and PovcalNet). Overall, 79 percent of respondents identified themselves as most knowledgeable with countries with more than one survey, while a little over 18 percent of respondents identified themselves as knowledgeable about countries with one or no survey. 202 APPENDIX C STAFF SURVEY OF WORLD BANK’S SUPPORT FOR POVERTY REDUCTION Overall Respondent Profiles Five fields that respondents most commonly selected as the area in which they were most knowledgeable were: infrastructure (14 percent); macroeconomic and fiscal policy (10 percent); agriculture and rural development (9 percent); public sector development (9 percent); and poverty (9 percent). About 58 percent of respondents indicated that they had worked on country strategies and 52 percent indicated that they had worked on projects and instruments targeting poverty in the past five years. Over one-third of the respondents (around 35 percent) indicated that they contributed to or led diagnostic work on poverty in the past five years. The majority of respondents were senior-level staff members, as roughly 48 percent of respondents self-identified as GG level 3 and 24 percent identified as GH level or above, while 28 percent of respondents identified themselves as GF-level staff. This distribution is similar to the distribution of grade levels of the respondent universe. 4 Additionally, over 68 percent of respondents have been employed by the World Bank for more than five years, of whom over 39 percent of respondents have been at the Bank for more than 10 years. This means that the respondent pool is over-represented by staff who have been at the Bank for five or more years. The geographic distribution of the respondents was almost evenly spread between the World Bank headquarters and country offices. However, the representation of Regions and networks was uneven. Staff from the Regions took a more active part in the survey than staff from the networks. The highest participation rates were observed among staff from the Africa (24 percent) and East Asia and Pacific (14 percent) Regions. Among the networks, the highest participation was from the respondents in the Sustainable Development Network (SDN, six percent) followed by Finance and Private Sector Development (FPD, three percent). 5 Key Messages The majority of respondents agreed that the Bank has sufficient access to data to provide policy advice to client governments. Roughly 77 percent of respondents agreed with the statement that the World Bank has sufficient data to make policy advice to client countries on poverty-related issues; 20 percent agreed strongly with this statement. This feedback was shared almost uniformly across GF, GG, and GH-level staff and across Regions. There were no significant variations between the positive responses from staff working on IBRD, IDA, and blend 6 countries, with 75 percent, 75 percent, and 83 percent of responses, respectively. Similarly there were no significant variations between FCS and non-FCS countries in the proportional percentage of 203 APPENDIX C STAFF SURVEY OF WORLD BANK’S SUPPORT FOR POVERTY REDUCTION positive responses, with 70 percent and 78 percent, respectively. However, 65 percent of staff working on countries with one or no survey between 2000 and 2012 agreed that the Bank has sufficient data to make policy advice while 79 percent of respondents from countries with more than one survey responded similarly to the question. Although the difference in the percentages of responses between these two groups of respondents is not very large, the two-tailed test comparing the means of responses show statistical differences in responses, with staff from one or no survey countries tending to respond more negatively. Staff who indicated that they have been at the Bank longer than 10 years were more likely to agree strongly that the Bank has sufficient data to make policy advice (24 percent) to governments than those who said that they have been at the Bank for less than two years (15 percent). Staff believe that the Bank has the needed quality of data to understand the causes of poverty in client countries. When asked whether the quality of the data that the Bank has is high enough to understand the causes of poverty in client countries, the overall feedback was positive, with 13 percent agreeing strongly and over 51 percent agreeing somewhat. The majority of respondents believed that the primary challenges to obtaining data on poverty are the client government’s insufficient capacity, budget, and political will. As such, staff identified insufficient capacity (57 percent), insufficient budget (42 percent), and insufficient political will of the government agencies (39 percent) as the top three constraints to obtaining data (see table C.2). Table C.1. Perceptions on the Primary Constraints to Obtaining Poverty Data, by Region (Percent of respondents) Perception AFR EAP ECA LCR MNA SAR Insufficient capacity of the government 64 64 53 48 51 67 agency or agencies responsible for collecting data Insufficient budget of the government agency 50 50 41 41 47 32 or agencies responsible for collecting data Insufficient political will within the 45 33 43 34 33 43 government Insufficient capacity within the World Bank to 8 9 2 6 10 10 help the government collect data on poverty 204 APPENDIX C STAFF SURVEY OF WORLD BANK’S SUPPORT FOR POVERTY REDUCTION Insufficient budget within the World Bank to 27 15 19 32 20 13 help the government collect data on poverty Insufficient interest from senior management 7 5 1 7 10 11 in your VPU to collect poverty-related data Not applicable 5 7 8 14 12 8 Do not know Other (please specify) Total (number) 173 101 88 90 49 84 Note: AFR = Africa; EAP = East Asia and Pacific; ECA = Europe and Central Asia; LCR = Latin America and the Caribbean; MNA = Middle East and North Africa; SAR = South Asia In general, a higher percentage of staff from the Regions identified insufficient government capacity and budget as primary constraints to obtaining poverty, compared to the networks. Among the Regions, staff from the Africa, East Asia and Pacific, and South Asia Regions were more likely than others to choose the government’s weak capacity as a constraint to obtaining data on poverty. As such, from 64 percent to 67 percent of respondents from each of these Regions chose the government’s insufficient capacity as one of the primary constraints. Furthermore, 69 percent of staff working on IDA countries believed that insufficient capacity of government agencies is one of the two major constraints as opposed to 56 percent of staff working on IBRD countries. Similarly, 76 percent of staff working on FCS countries believed the government’s weak capacity is one of the major constraints versus 54 percent of staff working on non-FCS countries. Additionally, staff working on countries with one or no survey were more likely to cite weak government capacity as a constraint than those working on countries with more than one survey with 95 percent and 60 percent, respectively. Bank staff mentioned the lack of sufficient government budget (42 percent) as the second most frequently cited constraint. This feedback was more frequently shared by staff most knowledgeable of IDA, blend, and FCS countries. It was defined as an issue by 53 percent of staff working on FCS countries, 40 percent of staff working on non-FCS countries, 52 percent of staff working on IDA countries, 32 percent of staff working on IBRD countries, and 35 percent of staff working on blend countries. Similarly, staff working on data poor countries indicated more frequently that the budget is an issue than those working on data rich countries, with 63 percent versus 45 percent, respectively. Another top obstacle identified by staff was insufficient political will (39 percent). Staff working on blend countries were more likely (50 percent) to identify insufficient 205 APPENDIX C STAFF SURVEY OF WORLD BANK’S SUPPORT FOR POVERTY REDUCTION political will within government as a main constraint compared to staff working on IDA countries (40 percent) and IBRD countries (34 percent). Staff have alternative sources of data to rely on when they don’t have poverty data. Respondents indicated that when data are scarce, the census or household surveys (74 percent), the Bank’s poverty assessments (37 percent), and administrative data on government programs (36 percent) are among the top three sources of data used to create diagnostic work. The smallest percentage of respondents (12 percent) indicated that they use experiences from other countries with similar poverty-related issues. The flow of data among the Bank networks and Regions is inconsistent. Fewer than half of the respondents (49 percent) believed that Bank staff almost always or frequently shared data on poverty within or across the Bank’s networks and Regions, while over 43 percent of responses ranged between sometimes, occasionally, and hardly ever. Challenges from data sharing in the Bank were flagged in an open-ended question when respondents were asked to provide feedback on why the Bank’s diagnostic work may not have an impact on government policies. As one respondent said: “Data collection is something Poverty Reduction and Economic Management [PREM] needs to take on as a core part of their program—we need spatially disaggregated, robust data. Currently we rely on the government's data, which PREM colleagues say is ‘confidentially’ shared by [the] Office of Statistics, and do not share this with others in the Bank although they freely use it for their own work. Sector teams do not get access to it, although one part of the Bank has it. It is not shared on the grounds of ‘building client relationships and trust.’” Bank staff appear to have positive outlook on the influence of poverty diagnostic work on government policies. As such, when asked about the influence of the Bank’s diagnostic work on poverty, 72 percent of respondents indicated that they agreed with the statement, of whom 54 percent agreed somewhat and 18 percent agreed strongly (see figure C.3). There were no significant variations among staff mapped to Regions (see table C.2). However, staff from the Europe and Central Asia and Middle East and North Africa Regions chose “agree strongly” less often than staff from other Regions (9 percent and 10 percent, respectively). 206 APPENDIX C STAFF SURVEY OF WORLD BANK’S SUPPORT FOR POVERTY REDUCTION Figure C.3. Perceptions on the Influence of Diagnostic Work (percent of respondents) Table C.2. Perceptions on the Influence of Diagnostic Work, by Region (Percent of respondents) AFR EAP ECA LCR MNA SAR Agree strongly 20 30 9 17 10 21 Agree somewhat 56 54 59 53 61 54 Disagree somewhat 15 8 19 15 14 18 Disagree strongly 5 2 7 4 8 2 Do not know 5 6 6 11 6 5 Note: The percentages are calculated proportional to the total number of respondents per Region. AFR = Africa; EAP = East Asia and Pacific; ECA = Europe and Central Asia; LCR = Latin America and the Caribbean; MNA = Middle East and North Africa; SAR = South Asia. Those who did not think that the Bank’s diagnostic work influences government policies were invited to explain why they thought so. Over 100 people responded to the open-ended question by sharing experiences and feedback that ranged from poor data quality and data sharing practices in the Bank, to the lack of practical knowledge among Bank staff, to lack of political will in the government, all the way to a perception that the Bank’s work provided little value added. As such, the lack of political will and different government priorities was mentioned by 39 percent of those who responded to the open-ended question, lack of coordination with government by 15 percent, corruption 207 APPENDIX C STAFF SURVEY OF WORLD BANK’S SUPPORT FOR POVERTY REDUCTION and fragility in countries by 14 percent, and lack of political knowledge by 11 percent. Some of the quotes from the responses are: • “The government has its own political and economic objectives and is very reluctant to be influenced by external diagnostic works, however evidence-based they might be.” • “We focus too much on data and think if the Bank provides relevant data- informed analysis this will somehow automatically translate into the right policy choices, while we neglect the adverse political economy that holds up policy choices for other reasons and [are] related to strong entrenched patronage.” • “Lack of political will in the government to make and implement the difficult decisions required. The World Bank has used its influence for getting various policies made, but has been quite ineffective in improving implementation.” Respondents indicated that the main challenges that Bank staff face in preparing certain analytical products are related to lack of sufficient data, budget, or time. Around 38 percent of respondents indicated a lack of sufficient poverty data as one of the major constraints to creating Poverty and Social Impact Analyses (PSIAs) and around 37 percent felt the same constraint while creating poverty assessments (PAs). In both PSIAs and PAs, staff also felt that time and budget constraints were key challenges. As such, over a quarter of the respondents said that they lacked sufficient time to do the PSIAs, while only 19 percent identified this constraint for PAs. Insufficient budget was also mentioned by 32 percent of respondents for PSIAs and 26 percent for PAs. Overall, staff have a positive perception that the Bank instruments address the poverty focus of country strategies. Over half of the respondents (55 percent) felt that development policy operations (DPOs) focus on the poverty issues outlined in the Bank’s country strategies either to a great extent or somewhat, while over one-third felt that it did so to a very limited extent or not at all (30 percent). Respondents mapped to Operations Policy and Country Services (OPCS, 33 percent) were among those who most frequently identified DPOs as being focused on poverty (see table C.3).to a great extent. In contrast, over 79 percent of respondents indicated that investment lending (IL) addresses the poverty focus as described in the Bank’s country strategies to a great extent or somewhat. Respondents agreeing that IL focuses on poverty to a great extent were more frequently mapped to the Africa Region (47 percent), South Asia Region (40 percent), Latin America and the Caribbean Region (44 percent), and SDN (41 percent). The more negative responses came from staff mapped to the Middle East and North Africa Region (25 percent) and PREM (25 percent) compared to others (see table C.4). 208 APPENDIX C STAFF SURVEY OF WORLD BANK’S SUPPORT FOR POVERTY REDUCTION Table C.3. Staff Perceptions about the Poverty Focus of Development Policy Operations, by Region and Network (Percent of respondents) Respons AF EA EC LC MN SA HD FP OPC PRE SD Othe Tota e R P A R A R N D S M N r l To a great 20 11 15 33 20 23 13 108 extent 17 14 11 13 6 Somewha 39 53 35 33 25 30 35 280 t 42 46 41 40 35 Very little 18 17 21 33 25 21 16 0 33 35 18 6 144 Not at all 11 9 15 5 19 9 0 20 0 5 7 6 70 Do not 11 21 30 0 15 23 39 100 know 11 13 12 9 15 Note: AFR = Africa; EAP = East Asia and Pacific; ECA = Europe and Central Asia; LCR = Latin America and the Caribbean; MNA = Middle East and North Africa; SAR = South Asia; HDN = Human Development Network; FPD = Finance and Private Sector Development; OPCS = Operations Policy and Country Services; PREM = Poverty Reduction and Economic Management; SDN = Sustainable Development Network. Other refers to staff who did not choose any of the provided mapping categories. Table C. 4. Staff Perceptions about the Poverty Focus of Investment Lending, by Region and Network (Percent of respondents) Respons AF EA EC LC MN SA HD FP OPC PRE SD Othe Tota e R P A R A R N D S M N r l To a great 46 31 24 41 21 40 26 26 33 20 40 13 247 extent Somewha 38 52 55 44 48 46 47 32 56 45 29 33 314 t Very little 9 10 17 9 15 7 11 11 11 25 16 37 89 Not at all 3 2 0 0 10 0 0 5 0 0 0 0 13 Do not 5 5 5 6 6 7 16 26 0 10 16 17 53 know Note: AFR = Africa; EAP = East Asia and Pacific; ECA = Europe and Central Asia; LCR = Latin America and the Caribbean; MNA = Middle East and North Africa; SAR = South Asia; HDN = Human Development Network; FPD = Finance and Private Sector Development; OPCS = Operations Policy and Country Services; PREM = Poverty Reduction and Economic Management; SDN = Sustainable Development Network. Other refers to staff who did not choose any of the provided mapping categories. 209 APPENDIX C STAFF SURVEY OF WORLD BANK’S SUPPORT FOR POVERTY REDUCTION GH staff were also more positive about IL’s focus on poverty than DPOs, as 36 percent of respondents at the GH and above level believed to a great extent that DPOs are poverty focused versus 17 percent who indicated the same for IL (see figure C.4). Figure C.4. Perceptions on the Extent to which IL and DPOs Address the Poverty Focus of the Country Assistance Strategy. GH and above 1% 36% 49% 13% 2% 2% GG 35% 44% 12% 7% IL To a great 2% extent GF 33% 41% 12% 12% Somewhat 13% GH and above 17% 43% 23% 4% Very little 10% Not at all DPO GG 16% 38% 20% 16% 7% Do not GF 13% 41% 18% 20% know 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Over 75 percent of respondents were positive that the Bank’s analytical work (i.e., economic and sector work, SW) and technical assistance (TA) addressed the poverty focus of the Bank’s country strategies. Only a quarter of respondents in each of the options felt that ESW and TA focus on poverty to a great extent. Staff mapped to the Africa Region were more likely to respond that ESW addresses country strategies’ focus on poverty to a great extent than staff from other Regions and networks (see table C.5), while staff mapped to the Human Development Network (HDN) felt similarly about TA (see table C.6). Respondents at the GH and above levels were more positive (33 percent) that ESW addresses the poverty focus to a great extent than staff at GF and GG levels (see figure C.5). 210 APPENDIX C STAFF SURVEY OF WORLD BANK’S SUPPORT FOR POVERTY REDUCTION Figure C.5. Perceptions on the Extent to which ESW and TA Address the Poverty Focus of the Country Assistance Strategy (in percent of respondents). 1% GH and above 24% 54% 15% 5% 1% GG 26% 51% 14% 8% TA 3% To a great extent GF 28% 51% 11% 7% Somewhat 1% Very little GH and above 33% 54% 11% 1% Not at all 1% ESW GG 24% 52% 14% 9% Do not know 3% GF 22% 52% 12% 10% 0% 20% 40% 60% 80% 100% Source: IEG Staff Survey Table C.5. Staff Perceptions about the Poverty Focus of ESW, by Region and Network (Percent of respondents) Response AFR EAP ECA LCR MNA SAR HDN FPD OPCS PREM SDN Other To a great 35 27 17 18 30 15 25 26 22 25 20 28 extent Somewhat 49 51 62 69 45 58 40 53 67 60 51 31 Very little 11 15 15 6 13 16 10 16 11 15 9 19 Not at all 1 1 2 2 2 2 0 0 0 0 2 0 Do not know 4 6 3 4 10 8 25 5 0 0 18 22 Note: AFR = Africa; EAP = East Asia and Pacific; ECA = Europe and Central Asia; LCR = Latin America and the Caribbean; MNA = Middle East and North Africa; SAR = South Asia; HDN = Human Development Network; FPD = Finance and Private Sector Development; OPCS = Operations Policy and Country Services; PREM = Poverty Reduction and Economic Management; SDN = Sustainable Development Network. Other refers to staff who did not choose any of the provided mapping categories. Table C.6. Staff Perceptions about the Poverty Focus of TA, by Region and Network (Percent of respondents) Response AFR EAP ECA LCR MNA SAR HDN FPD OPCS PREM SDN Other To a great 29 29 13 18 25 33 43 26 22 15 20 38 extent Somewhat 52 55 56 55 49 49 38 63 44 45 51 41 Very little 11 11 22 16 14 13 14 5 33 20 9 12 211 APPENDIX C STAFF SURVEY OF WORLD BANK’S SUPPORT FOR POVERTY REDUCTION Not at all 1 0 2 4 2 0 0 0 0 5 2 3 Do not 7 5 7 6 10 6 5 5 0 15 18 6 know Note: AFR = Africa; EAP = East Asia and Pacific; ECA = Europe and Central Asia; LCR = Latin America and the Caribbean; MNA = Middle East and North Africa; SAR = South Asia; HDN = Human Development Network; FPD = Finance and Private Sector Development; OPCS = Operations Policy and Country Services; PREM = Poverty Reduction and Economic Management; SDN = Sustainable Development Network. Other refers to staff who did not choose any of the provided mapping categories. Overall, staff feel that the Bank has sufficient data for country strategies. Of respondents, 71 percent agreed that the Bank has sufficient data to develop its strategies, with only 14 percent agreeing strongly. Staff working on IDA 7 countries were less likely to agree with the statement (68 percent agreeing) compared to staff working on blend countries (75 percent agreeing) and IBRD countries (73 percent agreeing). Also, 65 percent of staff working on FCS countries agreed while 72 percent of staff working on non-FCS countries agreed. (See table C.7). Table C.7. Sufficiency of Data to Create Strategies, by Borrower Status of Country (Percent of respondents) Borrower Status RESPONSE BLEND IBRD IDA TOTAL Agree somewhat 61 59 54 57 Agree strongly 14 14 14 14 Disagree somewhat 19 17 22 20 Disagree strongly 3 3 6 4 Do not know 3 8 4 5 IEG Staff Survey Staff who have been at the Bank for more than 10 years were more likely to agree strongly that the Bank has sufficient data to create country strategies (20 percent) than staff who have been at the Bank for less than two years (10 percent) and between two and five years (11 percent). Staff believe that the Bank addresses the causes of poverty in its strategies. Over 82 percent of respondents believed that the Bank addresses the causes of poverty in its strategies. More than half of the respondents (52 percent) believed that the Bank does so somewhat while only one-third (30 percent) believed that it does to a great extent. The extent to which respondents believed the Bank somewhat or to a great extent addressed the causes of poverty in its strategies increased according to a respondent’s length of service at the Bank. Of the respondents with less than two years of experience, 69 percent believed that the Bank integrates the causes of poverty into its strategies somewhat or to a great extent compared with 86 percent of similar responses from staff with more than five, but less than 10 years of experience, and 88 percent of responses from staff with 10 or more years of experience. 212 APPENDIX C STAFF SURVEY OF WORLD BANK’S SUPPORT FOR POVERTY REDUCTION Government reluctance to accept the findings of analytical work and time lag were identified as the two most important constraints to translating the Bank’s analytical work into country strategies. The top two constraints chosen by staff were government reluctance (38 percent) and time lags between the release of analytical work and the drafting of country strategies (37 percent). As seen in table C.8, a higher percentage of respondents from the Europe and Central Asia Region identified government reluctance as one of the two major constraints (48 percent) while staff from the Middle East and North Africa Region were least likely to cite it as a reason (22 percent), compared to other Regions. Table C.8. Staff Perceptions on the Translation of Poverty Analytical Work into Country Strategies, by Staff Mapping (Percent of respondents) Options AFR EAP ECA LCR MNA SAR Inadequate quality of poverty 23 17 14 14 20 18 diagnostic Disparities between the time when a 38 41 31 39 37 44 poverty diagnostic is released and the time when a country partnership strategy or country assistance strategy is produced Reluctance of the country team to 10 11 9 12 8 17 incorporate recommendations and findings of diagnostic work on poverty into strategies. Reluctance of government officials to 39 38 48 22 31 38 incorporate diagnostic work on poverty into the country’s poverty-related strategies Not applicable 19 24 19 31 18 20 Note: AFR = Africa; EAP = East Asia and Pacific; ECA = Europe and Central Asia; LCR = Latin America and the Caribbean; MNA = Middle East and North Africa; SAR = South Asia Although there were no significant variations across respondents mapped to Regions, a higher percentage of staff (46 percent) working on FCS countries believed that time disparities between the release of a country strategy and the poverty diagnostic work is a constraint versus staff working on non-FCS countries (36 percent), as shown in table C.9. 213 APPENDIX C STAFF SURVEY OF WORLD BANK’S SUPPORT FOR POVERTY REDUCTION Table C.9. Staff Perceptions on the Translation of Poverty Analytical Work into Country Strategies, by FCS and non-FCS Countries (Percent of respondents) Options Non-FCS FCS Inadequate quality of poverty diagnostic 17 25 Disparities between the time when a poverty diagnostic is released and the time when a 36 46 country partnership strategy or country assistance strategy is produced Reluctance of the country team to incorporate recommendations and findings of 13 13 diagnostic work on poverty into strategies. Reluctance of government officials to incorporate diagnostic work on poverty into the 38 36 country’s poverty-related strategies Not applicable 23 17 Other (please specify) 10 15 Note: FCS=fragile and conflict-affected states. In the open-ended responses provided by those who selected “Other,” staff further shared their insights. In almost 21 percent of open-ended responses (total n = 78), staff mentioned that, in some cases, analytical products are too technical, not grounded in country realities, or too general to be used in country strategies. A few respondents commented: • “Diagnostic's recommendations are often very general and are of little value to technical sector specialists who would like more details and narrower sector perspective, which is often missing.” • “General poverty analysis [is] not helpful, [it] requires focus on key sectors such as health in which data is scarce, and there is [a] lack of interest on the part of government.” • “The diagnostic work tends to be nonspecific and very aggregate. For example, determinants of poverty are usually about personal characteristics (large family, low education), but does not capture the impact of infrastructure on productivity, job opportunities, or wages; it doesn't even capture the impact of health on poverty status. 214 APPENDIX C STAFF SURVEY OF WORLD BANK’S SUPPORT FOR POVERTY REDUCTION Survey Questionnaire VPUs and Titles Excluded from the Survey Units Excluded Staff Titles Excludeda IFC Accounting officer MIGA Knowledge management (including officers, senior officers, coordinators, and lead officers) BPS Communications (including officers, senior officers, and lead officers) CTR Procurement specialists (including officers and senior officers) ECR Resource management specialists (including officers and senior officers) EXCb Conferences officer GSD Business Solutions (including officers and senior officers) HRD Information (including officers and senior officers) ICS Financial officer (including officers and senior officers)c IEG Ethics officer INT Executive secretary ITS Counsel (including senior and lead counsel) LEG External Affairs (including officers and senior officers) TRE Investigator SEC Special assistant WBId Young professional WBT Liaison Trust fund coordinator Engineer Risk officer Note: BPS Budget, Performance Review, and Strategic Planning; CTR = Controller's; ECR = External and Corporate Relations; EXC = Office of the President; GSD = General Services; HRD = Human resource Development; ICS =International Centre for Settlement of Investment Disputes ;IEG = Independent Evaluation Group ; IFC =International Finance Corporation ; INT = Integrity Vice Presidency; ITS = Information Technology Solutions ; LEG = Legal; MIGA = Multilateral Investment Guarantees Agency; SEC = Corporate Secretariat; TRE = Treasury ; WBI = World Bank Institute; WBT = World Bank Tribunal . a. Those that are mapped to networks and Regions but do not carry out the main line of work. Ideally they should be mapped to one of the excluded units, but they may have chosen to go with their VPU mapping. b. EXC includes the Office of Ethics and Business Conduct (EBC) c. Not to be confused with financial sector specialists who are included in the shortened list. d. WBI is currently renamed to the Leadership, Learning, and Innovation (LLI) unit 215 APPENDIX C STAFF SURVEY OF WORLD BANK’S SUPPORT FOR POVERTY REDUCTION SURVEY QUESTIONS AND RESPONSES 1. Among the countries with which the World Bank works, please identify the one country with which you are most knowledgeable as a result of your employment at the World Bank. (Please note that only the 10 countries with highest response rates have been listed below.) ANSWER OPTIONS RESPONSE RESPONSE COUNT PERCENT Bangladesh 2.8 24 Brazil 3.1 27 Ethiopia 2.7 23 India 5.3 46 Indonesia 6.4 55 Kenya 2.3 20 Nigeria 2.4 21 Pakistan 2.4 21 Tanzania 2.4 21 Vietnam 2.9 25 Answered question 866 Skipped question 2 2. Considering the country you have just selected, please identify the one field within that country with which you are most knowledgeable. ANSWER OPTIONS RESPONSE RESPONSE PERCENT COUNT Agriculture and rural development 8.9 77 Education 6.4 55 Energy 5.9 51 Environment, natural resources, and climate change 7.0 61 Finance 5.3 46 Infrastructure (including urban, transport, water, and sanitation, etc.) 14.0 121 Health, nutrition and population (including HIV/AIDS, pandemics, etc.) 6.1 53 Public sector development (including governance and anticorruption) 9.0 78 Private sector development 4.4 38 Poverty 8.5 74 Macroeconomic and fiscal policy 10.3 89 Social protection and labor 6.5 56 Other (including gender, fragile and conflict states, etc.) (Please specify.) 7.7 67 Answered question 866 Skipped question 2 216 APPENDIX C STAFF SURVEY OF WORLD BANK’S SUPPORT FOR POVERTY REDUCTION 3. At any time in the past five years have you been involved in one or more of the following activities in the country you selected above as part of your work at the Bank? Please check all that apply. ANSWER OPTIONS RESPONSE RESPONSE PERCENT COUNT Contributing to or leading diagnostic work on poverty 34.5 280 Advising how to collect or collecting poverty data 20.6 167 Working on country strategies 57.7 468 Working on projects and instruments specifically targeting poverty 51.5 418 Not applicable 13.3 108 Other poverty-related work (Please specify.) 9.5 77 Answered question 811 Skipped question 57 4. Do you agree that the World Bank has sufficient data to provide policy advice on poverty-related issues to government in the country and field you selected above? ANSWER OPTIONS RESPONSE RESPONSE PERCENT COUNT Agree strongly 19.6 160 Agree somewhat 57.0 466 Disagree somewhat 13.8 113 Disagree strongly 5.9 48 Do not know 3.8 31 Answered question 818 Skipped question 50 5. The quality of data available to understand the causes of poverty in the country you selected above is sufficient. ANSWER OPTIONS RESPONSE RESPONSE PERCENT COUNT Agree strongly 13.4 110 Agree somewhat 51.3 420 Disagree somewhat 21.6 177 Disagree strongly 9.0 74 Do not know 4.5 37 Answered question 818 Skipped question 50 217 APPENDIX C STAFF SURVEY OF WORLD BANK’S SUPPORT FOR POVERTY REDUCTION 6. If applicable, what are the primary constraints to obtaining data on poverty in the country you selected above? Please select up to three. ANSWER OPTIONS RESPONSE RESPONSE PERCENT COUNT Insufficient capacity of the government agency or agencies responsible for collecting 57.3 445 data Insufficient budget of the government agency or agencies responsible for collecting 42.2 328 data Insufficient political will within the government 38.9 302 Insufficient capacity within the World Bank to help the government collect data on 7.2 56 poverty Insufficient budget within the World Bank to help the government collect data on 21.5 167 poverty Insufficient interest from senior management in your VPU to collect poverty-related 6.9 54 data Not applicable 8.8 68 Do not know 7.9 61 Other (Please specify.) 8.9 69 Answered question 777 Skipped question 91 7. World Bank staff share data on poverty within and across the Bank’s Networks or Regions. ANSWER OPTIONS RESPONSE RESPONSE PERCENT COUNT Almost always 14.0 109 Frequently 35.1 272 Sometimes 28.9 224 Occasionally 9.1 71 Hardly ever 5.4 42 Do not know 7.5 58 Answered question 776 Skipped question 92 218 APPENDIX C STAFF SURVEY OF WORLD BANK’S SUPPORT FOR POVERTY REDUCTION 8. When there are constraints to obtaining data on poverty in the country you selected, which of the following are most often used to create diagnostic work on poverty? Please check all that apply. ANSWER OPTIONS RESPONSE RESPONSE PERCENT COUNT Administrative data on government programs 35.9 277 Data from a census or from household surveys (such as Household Budget Surveys, 73.6 568 Living Standards Measurement Studies, Labor Surveys, etc.) Surveys of firms or enterprises 18.7 144 Experiences from other countries with similar poverty-related issues 11.9 92 Evaluations of World Bank projects 20.7 160 Information from a non-representative sample or subsample of the population of 15.5 120 interest World Bank Poverty Assessments 36.9 285 Do not know or not applicable 13.3 103 Other (Please specify.) 3.8 29 Answered question 772 Skipped question 96 9. The World Bank has sufficient data on poverty to develop strategies in the field and country I selected above. ANSWER OPTIONS RESPONSE PERCENT RESPONSE COUNT Agree strongly 13.9 108 Agree somewhat 56.6 439 Disagree somewhat 19.7 153 Disagree strongly 4.4 34 Do not know 5.3 41 Answered question 775 Skipped question 93 10. To what extent are the causes of poverty addressed in the World Bank’s strategies (such as in a Country Partnership Strategy) in the country you selected above? ANSWER OPTIONS RESPONSE PERCENT RESPONSE COUNT To a great extent 30.0 232 Somewhat 52.1 403 Very little 10.2 79 Not at all 1.6 12 Do not know 6.1 47 Answered question 773 Skipped question 95 219 APPENDIX C STAFF SURVEY OF WORLD BANK’S SUPPORT FOR POVERTY REDUCTION 11. The World Bank’s diagnostic work on poverty has influenced the government’s policies or strategies in the country and field I selected above. ANSWER OPTIONS RESPONSE PERCENT RESPONSE COUNT Agree strongly 18.4 142 Agree somewhat 53.9 416 Disagree somewhat 14.5 112 Disagree strongly 4.4 34 Do not know 8.8 68 Answered question 772 Skipped question 96 12. If you think the Bank’s diagnostic work on poverty has had little or no influence on government policies, please tell us why you think so. RESPONSE COUNT Answered question 106 Skipped question 762 220 APPENDIX C STAFF SURVEY OF WORLD BANK’S SUPPORT FOR POVERTY REDUCTION 13. Based on your experience in the country you selected above, what are the main challenges, if any, that World Bank staff faces when developing the following products? Please check all that apply. (Response percent) ANSWER OPTIONS INSUFFICIENT INSUFFICIENT INSUFFICIENT INSUFFICIENT INSUFFICIENT OTHER DO NOT RESPONSE TIME BUDGET DATA ON ATTENTION COORDINATION (PLEASE KNOW OR COUNT POVERTY FROM THE BETWEEN AND WITHIN SPECIFY NOT BANK’S BANK REGIONS OR BELOW) APPLICABL MANAGEMENT NETWORKS E Poverty and Social 24.2 30.7 36.2 13.2 15.9 6.4 22.9 Impact Analysis Poverty 17.5 24.7 34.3 9.8 11.0 5.6 28.0 Assessment and poverty updates/notes Public Expenditure 14.1 21.0 20.9 9.0 13.2 7.5 34.6 Review Country Economic 10.3 11.9 14.7 5.6 9.6 5.2 46.4 Memorandum Other (Please specify.) Answered question 737 Skipped question 131 14. How well do the following instruments address the poverty focus of the Bank’s country strategies in the country you selected above?(Response percent) ANSWER OPTIONS TO A GREAT EXTENT SOMEWHAT VERY LITTLE NOT AT ALL DO NOT KNOW RESPONSE COUNT Development Policy Operations 14.8 38.2 19.4 9.9 13.8 Investment Lending 33.2 43.7 12.0 1.7 7.3 Economic and Sector Work (ESW) 24.8 51.5 12.5 1.5 7.6 Technical Assistance 25.6 50.5 13.3 1.5 7.3 Answered question 751 Skipped question 117 221 APPENDIX C STAFF SURVEY OF WORLD BANK’S SUPPORT FOR POVERTY REDUCTION 15. The World Bank’s poverty analytical and advisory services (such as technical assistance and ESW) complement the Bank’s lending instruments in the country I selected above. ANSWER OPTIONS RESPONSE RESPONSE PERCENT COUNT Agree strongly 29.8 223 Agree somewhat 51.7 387 Disagree somewhat 10.8 81 Disagree strongly 1.5 11 Do not know 6.3 47 Answered question 749 Skipped question 119 16. If applicable, what are the primary constraints to translating diagnostic work on poverty into the Bank’s country strategies in the country you selected above? Please select no more than two. ANSWER OPTIONS RESPONSE RESPONSE PERCENT COUNT Inadequate quality of poverty diagnostic 18.2 132 Disparities between the time when a poverty diagnostic is released and the time 37.2 270 when a Country Partnership Strategy or Country Assistance Strategy is produced Reluctance of the country team to incorporate recommendations and findings of 12.9 94 diagnostic work on poverty into strategies Reluctance of government officials to incorporate diagnostic work on poverty into the 37.7 274 country’s poverty-related strategies Not applicable 22.0 160 Other (Please specify.) 10.7 78 Answered question 726 Skipped question 142 222 APPENDIX C STAFF SURVEY OF WORLD BANK’S SUPPORT FOR POVERTY REDUCTION 17. To what extent does the World Bank seek feedback from the following groups when developing its country partnership strategies in the country you selected above? (Response percent) GROUP TO A SOMEWHAT VERY NOT DO NOT RESPONSE GREAT LITTLE AT ALL KNOW COUNT EXTENT Donor (including bilateral and 39.7 38.9 8.2 1.2 11.6 730 multilateral organizations, and UN agencies) Academia (including research 19.6 43.1 22.4 2.6 11.9 730 institutes and think tanks) Civil society (including 23.6 43.7 19.2 1.5 11.5 729 international and locally based nongovernmental organizations) Private sector 14.2 38.1 26.1 6.1 14.3 724 Other 4.8 3.0 3.3 1.0 20.2 236 If you chose Other above, please specify. 48 Answered question 733 Skipped question 135 18. Lessons and findings from diagnostic work inform the development of the World Bank’s strategies in the country I selected above. ANSWER OPTIONS RESPONSE PERCENT RESPONSE COUNT Agree strongly 26.9 199 Agree somewhat 56.2 416 Disagree somewhat 8.6 64 Disagree strongly 2.0 15 Do not know 6.2 46 Answered question 740 Skipped question 128 19. My current grade level is: ANSWER OPTIONS RESPONSE RESPONSE COUNT PERCENT GF 28.2 208 GG 48.3 356 GH and above 23.5 173 Answered question 737 Skipped question 131 223 APPENDIX C STAFF SURVEY OF WORLD BANK’S SUPPORT FOR POVERTY REDUCTION 20. Where have you worked most of the time during the past two years? If you have been employed at the World Bank for less than two years, in which part of the Bank have you worked the most since joining the Bank? ANSWER OPTIONS RESPONSE PERCENT RESPONSE COUNT AFR 23.5 173 EAP 13.7 101 ECA 12.0 88 LCR 12.2 90 MNA 6.7 49 SAR 11.4 84 FPD Anchor 2.9 21 HDN Anchor 2.6 19 PREM Anchor 2.7 20 SDN Anchor 6.3 46 OPCS 1.2 9 Other (Please specify.) 4.8 35 Answered question 735 Skipped question 133 21. Where have you been physically located most of the time during the past two years? If you have been employed at the World Bank for less than two years, where have you been physically located most of the time since you joined the Bank? ANSWER OPTIONS RESPONSE PERCENT RESPONSE COUNT Headquarters 50.1 368 Country Office 49.6 364 Other (Please specify.) 0.3 2 Answered question 734 Skipped question 134 1 Please note that effective July 1, 2014, many vice-presidencies and sector units were replaced by the new Global Practices. 2 For instance, the World Development Report 2015 on Mind, Society, and Behavior, which is the largest flagship report of the World Bank, reported staff survey results at a 39 percent response rate. The survey was not sampled and was sent to 4,797 World Bank staff from all sectors of the World Bank to participate in a survey designed to measure perceptions. The WDR further reports that this response rate is well above the needed rate for representativeness (See WDR 2015, page 190). Similarly, DEC’s research paper on influence of World Bank research fielded a survey among all senior operations staff at grades GG and 224 APPENDIX C STAFF SURVEY OF WORLD BANK’S SUPPORT FOR POVERTY REDUCTION above and received a 19 percent response rate (555 respondents out of 2,900). (See Research at Work: Assessing the Influence of World Bank Research. Development Economic Unit. World Bank 2012). 3The Bank uses 10 grade levels as a way to structure its workforce. Letters are used to signify the various grades which reflect increasing levels of responsibility, skills, and requirements 4The distribution of grade levels in the respondent universe is: 27 percent, GF staff; 47 percent, GG staff; and 25 percent, GH and above staff. 5 Refer to the disclaimer about country and regional distribution of responses in Box D-.1. 6The analysis of the internal survey results by the types of countries (IBRD, IDA, blend, FCS and non-FCS) excludes the responses from staff indicating that they are most knowledgeable about the countries that are not classified by the World Bank. These countries include Australia, Cuba, Finland, Hungary, Kuwait, Latvia, Lithuania, Saudi Arabia, South Sudan, Spain, Sweden, the United Arab Emirates, and West Bank and Gaza. One exception is South Sudan, which did not become a World Bank member until 2012. The responses from all of these countries were included in the cumulative numbers of responses in each question. 7 Economies are divided into IDA, IBRD, and Blend countries based on the operational policies of the World Bank. International Development Association (IDA) countries are those with low per capita incomes that lack the financial ability to borrow from the International Bank for Reconstruction and Development (IBRD). Blend countries are eligible for IDA loans but are also eligible for IBRD loans because they are financially creditworthy. Source: World Bank Webiste, https://datahelpdesk.worldbank.org/knowledgebase/articles/378834–how-does-the- world-bank-classify-countries 225 Appendix D. Key Findings from Focus Group Discussions with Staff As part of the evaluation, the Independent Evaluation Group (IEG) conducted six focus group discussions with Bank staff to gather insights on how the availability of poverty data affects the Bank’s poverty diagnostic and whether and how they translate into country strategies. The focus groups collected qualitative feedback based on the participants’ personal experiences and views. As such, the responses were treated as exploratory and were triangulated with other sources of data in the final report. Focus groups brought together 22 staff members working on countries across all six Regions of the world. Participants were selected with the consideration of the balance between the types of countries by lending (i.e., International Bank for Reconstruction and Development, IBRD; International Development Association, IDA/blend), fragile and non-fragile countries, and availability of data. Given resource and time constraints, focus group discussions gathered staff working on 18 countries: Afghanistan, Botswana, Cambodia, Democratic Republic of Congo, India, Indonesia, Jordan, Kosovo, Kyrgyz Republic, Moldova, Pakistan, the Russia Federation, South Sudan, Sri Lanka, Papua New Guinea, West Bank and Gaza, and the Republic of Yemen as well as the Organization of Eastern Caribbean States. 1 The homogeneity of each focus group was in the type of work participants performed as their main duties. As such, each focus group distinguished between task team leads (TTLs) or co-leads of poverty assessments and poverty updates, country economists, and TTLs and co-TTLs of country partnership strategies. Invitations were extended to all identified types of staff in each of the above mentioned countries. In the cases where poverty assessments for the identified countries were produced prior to 2009, the country economists were invited to participate instead of the assessment TTLs. Not everyone who was invited to participate in the discussions was able to join. Every effort was made to accommodate the busy and varying schedules of the invitees to ensure their participation. The main questions raised in the discussions were the same across all the focus groups. The same facilitator led discussions in each focus group. Each focus group discussion was conducted for an hour and a half. When necessary, the facilitator asked for additional information and probed for emerging trends. The list of the main questions appears below. The discussions about the use and translation of poverty analytics and data in country strategies varied among groups. As such, poverty assessment leaders and country economists did not respond to all of the 227 APPENDIX D KEY FINDINGS FROM THE FOCUS GROUP DISCUSSIONS WITH STAFF questions regarding country strategies since they felt their experience in this field was relatively limited. Similarly, a few of the country partnership strategy (CPS) TTLs and co-TTLs were not comfortable responding to some of the questions related to data and poverty diagnostic. Main Questions DATA AVAILABILITY • Do you think the Bank has sufficient poverty data to make policy advice related to poverty in client countries? • What are the main constraints to obtaining poverty data in your line of work? QUALITY AND CONSTRAINTS TO POVERTY DIAGNOSTIC WORK • What are the main constraints that the World Bank faces when developing analytical work on poverty? TRANSLATION OF POVERTY DATA AND DIAGNOSTIC INTO COUNTRY STRATEGIES • What are the primary constraints to translating analytical work on poverty into the Bank’s country strategies? • In your experience, what plays an important role in the design and focus of country assistance strategies (CASs) and CPSs? FEEDBACK LOOPS • How much does the Bank consult with external stakeholders while creating country strategies? • What are the obstacles for an effective integration of stakeholder feedback, if any? Perceptions on Availability of Poverty Data Despite many data constraints, the participants do not consider the adequacy of data a main constraint to making policy advice, particularly when the economists and experts know the country context well. In general, all of the participants were in agreement that data availability and data quality vary by country and region. However, participants in the focus groups did not express strong feedback that, in the countries where there has been only one or no Household Income and Expenditure Survey (HIED) or Living Standard Measurement Survey (LSMS) between 2000 and 2012, 2 staff face significantly more constraints in producing their analytical work, providing policy advice, or creating country strategies when 228 APPENDIX D KEY FINDINGS FROM THE FOCUS GROUP DISCUSSIONS WITH STAFF compared to countries with the greater number of surveys. The general explanation was that the Bank has sufficient data to make policy advice and uses its strong knowledge of country context to do its work even in the countries with limited data. Discussions further highlighted that in countries where there is no sufficient HIES or LSMS data, Bank staff become more opportunistic in using other sources of data and try to do quicker assessments on the ground. The Bank needs to have long-term investments in data collection and analytical work, including building client country capacity. Most of the participants agreed that the Bank has a comparative advantage and client-country demand to do capacity building work. Capacity building work benefits the Bank in the long run because it ensures the systematic collection of higher quality data and greater supervision over data collection. The discussions also highlighted that in certain cases, the Bank engages in one-off deals to collect HIES data in low-income countries or uses limited engagement through trust funds for capacity building. Both strategies may not be sustainable if there is no long-term planning. Some participants also highlighted different types of challenges the Bank faces in countries while helping build capacity. According to the participants, in low-income countries and in the Sub-Saharan Africa Region, particularly, an influx of donors compete for governments’ attention and in some cases have competing agendas. In those situations the Bank could be a smaller player and may either duplicate efforts in building data capacity within the client governments or not be able to convince the governments that higher data collection capacity may have long-term positive impact. The discussants also highlighted that working on middle-income countries is also challenging because donors are not as willing to invest in these countries as they are in low-income countries and often the Bank can’t secure its own trust funds to do this work. Discussions with country program coordinators pointed out that building data collection capacity of the client governments not only helps governments collect better data, but also ensures greater country ownership and consideration of country context. In addition to capacity building work, poverty assessment leads and country economists pointed out that there is a need for additional budget and long-term investments in the data and diagnostic work. Participants highlighted the current tension between the Bank’s “belt-tightening” and staffing policies, and its need to produce more data and poverty analytics to better understand the root causes of poverty and to achieve the Bank’s new twin goals. Some of the participants highlighted that the Bank’s budget and organizational structure does not support having a cadre of well-trained staff who work continuously on data collection and 229 APPENDIX D KEY FINDINGS FROM THE FOCUS GROUP DISCUSSIONS WITH STAFF data harmonization as well as retaining highly specialized poverty economists on open-ended contracts. In all of the discussions with country economists and poverty assessment leads, and in one of the focus groups with the CPS leads, staff highlighted that data-related work is mainly done by short-term consultants who work on an ad-hoc basis. Because of this, available data are not harmonized, and, often times, is not comparable for the purposes of proper analysis. It is important to consider political economy around the collection and sharing of poverty data. In general, participants indicated that the Bank’s access to data in its client countries is improving. At the same time, the discussions highlighted a highly sensitive political nature of poverty data in the countries that participants have worked in. Based on this experience, the discussions also stressed the importance for the Bank to establish trustworthy and long-term relationships with client governments to collect the needed data. The CPS and CAS leads stressed the importance of building relationships with client governments as a building block for obtaining poverty-related data after they are collected. Several discussion groups pointed to the Bank’s important role in making data available to all interested parties in the client countries. Some of the staff working on the fragile and conflict-affected countries noted that client governments share data with the Bank that have not been released publicly or shared with others. Thus, the Bank has a critical role to play in making such data transparent and available. The Bank needs to make better use of other sources of data and consider changing its data sources to have a more comprehensive picture of its impact and engagements. Most of the country economists and poverty assessment leads, along with some CPS leads, mentioned that relying only on LSMS and HIES data is not sufficient to understand the impact of the Bank’s work on poverty, and the Bank needs to rethink its data needs and investments. Participants mentioned that there is a need for collection of data on the impact of Bank’s and other donors’ projects on poverty. Discussions highlighted that HIES and LSMS data are very useful in creating poverty profiles but are not dense and multidimensional as they focus only on income-based dimensions and do not cover the large number of issues important to poverty work. Collecting detailed and multidimensional data will help poverty specialists respond to specific questions from country directors and country management units (CMUs) on policy choices. For instance, some poverty assessment leads mentioned that they are being asked to add language on shared prosperity to all documents and strategies because it is the new trend in the Bank, but there is not enough data to do so. 230 APPENDIX D KEY FINDINGS FROM THE FOCUS GROUP DISCUSSIONS WITH STAFF The issue of relating and harmonizing various datasets reverberated through all of the discussions and was pointed out as an imperative going forward. As a few of the participants pointed out, including some of the CPS and CAS leads, the Bank needs to ensure that the “various datasets talk to each other.” Some participants highlighted that harmonization and centralization of data will require better collection and use of the project level data, and could accordingly shed light on the impact of the Bank’s interventions on the ground. Additionally, one of the CPS leads groups and one of the poverty assessment leads groups highlighted the urgent need for the Bank to strengthen its data availability at the subnational level, where government capacity is usually weaker and data are much scarcer. The discussions highlighted that since the Bank aims to work on eliminating extreme poverty, it needs to better understand the underlying causes of poverty and the impact of its interventions at the subnational level. Perceptions on the Quality of and Constraints to Analytical Work on Poverty In general, the participants agreed that the Bank’s work on data quality and poverty analytics has significantly improved in the past 10 years, particularly since it created a centrally located team of poverty experts. Some of the discussions with CAS and CPS leads pointed out that multidimensional analysis of issues, including the use of both quantitative and qualitative tools, helped create better pictures of poverty in their countries. Furthermore, the Bank has created useful analytical tools, such as poverty maps, incidence analysis, and STATCAP that create more opportunities for working with governments on capacity building and data collection because these tools provide mutually beneficial services. The overall feedback was that the quality of the analysis in poverty assessments is usually strong but often does not provide specific policy recommendations. The focus groups with the poverty assessment leads and country economists highlighted that poverty assessments are not designed to provide specific policy recommendations, but rather they serve as background information and analytical tools. Some of them pointed that poverty data currently collected by the Bank do not lend themselves for being used for policy recommendations in diagnostic work. ), . Poverty assessment leads felt that they lack time and cross-sectoral collaboration to deliver comprehensive products. Poverty diagnostic work needs better continuity through long term planning and budgeting. Poverty assessment leads and country economists mentioned that poverty diagnostic work deserves greater continuity. Currently incentives to be a 231 APPENDIX D KEY FINDINGS FROM THE FOCUS GROUP DISCUSSIONS WITH STAFF poverty economist are not high because staff contracts have limited terms (two to three years), and internal recognition of the work is insufficient. Thus, the Bank does not have enough well-trained staff with extensive backgrounds in poverty work. Furthermore, some of the poverty assessment, CAS, and CPS leads mentioned that poverty assessments and knowledge work heavily depend on the availability of trust funds, particularly in middle-income countries where governments are more selective about Bank engagement. This sentiment was broadly shared by the staff working on IBRD countries and countries with rich data. As a positive example, poverty assessment leads and country economists noted the benefits of adopting a programmatic approach to poverty work more widely. In four groups with poverty assessment leads and country economists, participants underscored that more systematic poverty assessments and updates leading to the production of shorter and more regular notes have more potential of being used both by the clients and the country teams. For example, a large poverty assessment right before the CPS process starts may not have an impact on it as much as regular poverty updates submitted to the country team. The programmatic approach also allows TTLs to secure funds and to plan better for the dissemination of poverty work. The participants further highlighted that the dissemination of poverty work is usually neglected. The Bank needs to create more opportunities and incentives for staff to disseminate the findings of poverty diagnostics. Participants pointed out that the Bank does not do enough to disseminate the findings of its poverty work internally and even more so in the client countries. Poverty assessment leads also pointed out that the Bank’s communications team (External and Corporate Relations) does not provide adequate support on how to best package messages from the poverty work based on the needs and interests of various stakeholders. Some of the recommendations from the focus group participants included building in time and sufficient budget into poverty analytical work for proper dissemination and collaboration with the government agencies, which may mean working closely on capacity development and communication with the Bank’s teams. Additionally, the participants pointed out that there is more institutional recognition of and rewards for large poverty assessments than smaller updates and poverty notes, because the larger products tend to have a “big splash” in the media. However, it is not evident that larger studies have more impact. Thus, the Bank needs to reward staff for good work on smaller products that are tailored to country needs or for informal consultations with clients, where most of the learning may happen. 232 APPENDIX D KEY FINDINGS FROM THE FOCUS GROUP DISCUSSIONS WITH STAFF There is a need for better collaboration and coordination among World Bank networks and sectors. Poverty assessment leads highlighted that the engagement of sector specialists across networks and countries depends on personal connections rather than institutional policy and official staff terms of reference. As such, poverty assessment leads often use their personal connections with experts they know to provide inputs to chapters in their assessments. Some participants pointed out that the lack of collaboration between poverty teams and sector staff negatively impacts the results of the analysis and specificity of recommendations in poverty work. A positive example of cross-sectoral collaboration cited by some poverty assessment leads was Poverty and Social Impact Analysis, which according to the discussants, required formal coordination and involvement across sectors and thus had strong multisectoral analysis. Two of the focus group discussions with poverty assessment leads highlighted that it was easier to collaborate with the social protection teams than with the education and health teams. Perceptions on How Poverty Data and Analytics Translate into Country Strategies CMUs and country directors play a critical role in CAS and CPS design processes. There was a general agreement in all of the focus group discussions that CMUs and country directors, in particular, play an important role in CAS and CPS processes. 3 Both CPS and poverty assessment leads mentioned that country directors’ preferences and past experiences have a great impact on the CPS, which could be both a positive learning experience and a drawback in cases when they don’t align with a specific country context. A division of opinions was observed between the two focus groups including CAS and CPS leads where one group believed that a country director’s experience and agenda play a critical role and can change the trajectory of country engagements whereas the other group believed that a country director can change the pace and leadership style rather than the strategy direction. One of the explanations could be the size of the client countries and Bank’s engagement as well as IBRD/blend countries versus IDA countries. Poverty diagnostic is not consistently and comprehensively used in CAS and CPS. The discussions with the poverty assessment leads and country economists revealed a unanimous sentiment that typically only a small percentage of poverty work is used in country strategies. They added that usually the time and efforts spent to prepare the poverty works are not used efficiently as only a small percentage of their work is used in strategies. They mentioned that poverty experts are usually called on to provide either descriptive information or support a point with specific data that the CAS and CPS want to promote. A number of the discussants in the groups including poverty assessment leads and country economists highlighted that some 233 APPENDIX D KEY FINDINGS FROM THE FOCUS GROUP DISCUSSIONS WITH STAFF issues are included in CAS and CPS because they may be current trends in the Bank and are merely used to “tick the box,” citing gender and shared prosperity as two examples. At the same time, some of the poverty assessment leads in various groups acknowledged that country teams may not be ignoring poverty diagnostic on purpose, but they omit poverty analysis because it is not the “same language” they are used to, which makes it harder for them to understand a more nuanced approach to poverty. The discussions with the CAS and CPS leads confirmed that they view poverty diagnostics as key to creating poverty profiles but not to providing concrete policy advice or even being directly used in the CAS and CPS documents. However, these underpinnings provide the Bank with the ability to have an informed dialogue with governments and have a strategic vision around its interventions. Perceptions on How the Bank Creates and Uses Feedback Loops for Designing Its Country Strategies The Bank has a limited engagement with stakeholders and needs to adopt a better approach for creating stronger feedback loops. In all of the focus group discussions there was a sentiment that although the Bank consults with stakeholders on country strategies, it is usually more of a formal process. The CPS leads pointed out that it is better for the Bank to engage in ongoing consultations with the stakeholders rather than expect to have informed and well-prepared feedback from a smaller group of stakeholders in the short timeframe during which a CAS or a CPS is prepared. They added that often one-off and quick consultations raise unnecessary expectations among stakeholders as they hope to see the reflection of their ideas and advice in the final document. Additionally, some of the poverty assessment and CAS and CPS leads pointed out that often the stakeholders don’t know enough about the Bank’s work on the ground to be able to provide feedback on the strategy or any other document. It was recommended that the Bank work on raising awareness and engaging with stakeholders. 1Although the Organization of Eastern Caribbean States includes six independent countries and three British Overseas Territories, the focus group discussions covered only the six independent countries: Antigua and Barbuda, Dominica, Grenada, St. Kitts and Nevis, St.Lucia, and St. Vincent and the Grenadines since they are World Bank client countries. 2 Forthe purpose of this evaluation, data accessibility is measured by the data available in Development Economic’s central data platforms (such as PovCalNet) and looks at the frequency of the Household Income and Expenditure Survey and the Living Standard Measurement Survey. 234 APPENDIX D KEY FINDINGS FROM THE FOCUS GROUP DISCUSSIONS WITH STAFF 3 Only one person disagreed with this general statement, saying that usually it is delegated to other people in a country management unit, and thus the country partnership strategy usually looks more like a collection of sector-based pieces than a strategy. 235 Appendix E. Selection Criteria and Methodology Used to Assess the Quality of Poverty Assessments from 20 Countries The objective of this Poverty Assessment (PA) Quality Review is to determine the extent to which the PAs (a) provided sufficient background information on available poverty surveys and data, (b) made good use of available survey and other data to provide a clear understanding of the extent and drivers of poverty, (c) assessed the adequacy of the countries’ poverty reduction institutions, programs and funding, (d) evaluated poverty monitoring and evaluation arrangements, (e) proposed specific and actionable recommendations for reducing poverty, and (f) influenced the countries’ poverty reduction strategies and programs, helped build in-country capacity, and supported joint work and partnerships. The methodology for evaluating the quality of the twenty PAs is consistent with the World Bank’s 2004 Guidance Note on Poverty Assessments. 1 The review period is 2002–13. The term “poverty assessment (PA)” includes full PAs as well as poverty notes, updates, reports, and TAs. The term “poverty profile” refers to estimates of the levels and trends in poverty (a) at the national level and disaggregated by regions and social groups, and (b) across income, consumption, and non-income indicators. The term “poverty diagnostics” refers to the examination of the (a) key drivers of income and non-income poverty at the national and regional levels and across social groups, (b) determinants of the changes in poverty incidence over time, including growth and distributional changes, and (c) obstacles to poverty reduction nationwide, most-affected regions, and across social groups. Poverty Assessment Selection The Review examined the quality of 20 country poverty assessments (PAs), notes, and poverty reduction technical assistance projects. There is some overlap with the study’s country case studies (Bangladesh, Egypt, Lao PDR, and Nigeria). Most of the 20 countries have at least one publically available poverty assessment and many have poverty notes, poverty updates, and technical assistance reports. The 20 countries were selected to (i) provide equal coverage of each of the Bank’s six Regions (four countries each for the Africa and East Asia and Pacific Regions, and three countries each for the remaining four Regions); (ii) countries with greater rates 237 APPENDIX E SELECTION CRITERIA AND METHODOLOGY USED TO ASSESS THE QUALITY OF POVERTY ASSESSMENTS FROM 20 COUNTRIES of poverty either as a proportion of the developing world’s poor or as a share of country population; and (iii) at least one weak data country in each Region. 2These selection criteria should identify any gaps in the quality of Bank poverty diagnostics across the Regions—across those countries with the greatest poverty reduction challenges in absolute or relative terms—and countries with weak data. The 20 countries, with weak data countries in italics, are: • Africa (4): Democratic Republic of Congo, Ethiopia, Mozambique, and Nigeria • East Asia and Pacific (4): China, Indonesia, Lao People’s Democratic Republic, and Papua New Guinea • Europe and Central Asia (3): Armenia, the Kyrgyz Republic, and Moldova • Latin America and the Caribbean (3): Brazil, Colombia, and Guyana • Middle East and North Africa (3): the Arab Republic of Egypt, Iraq, and Republic of Yemen • South Asia (3): Afghanistan, Bangladesh, and India POVERTY HEADCOUNT Using the most recent year that data were collected, nine countries were selected based on having the greatest poverty headcounts under an international poverty line of $ 1.25 per day per person at 2005 purchasing power parity (PPP) prices. These nine countries are (year of PA completion and the number of poor in millions are in parentheses): Nigeria (2010, 109), the Democratic Republic of Congo (2006, 49), Ethiopia (2005, 37), Mozambique (2008, 14), China (2009, 157), Indonesia (2010, 44), Brazil (2009, 12), India (2010, 394), and Bangladesh (2010, 65). The number of poor in these nine countries totaled to about 881 million, or nearly three-quarters of the 1.2 billion poor in all developing countries in 2010. One additional country was selected on the basis of national poverty: Egypt (2009, 17). Not surprisingly, there is a heavy overlap between the rankings of countries with the greatest poverty headcounts based on the international and national poverty lines. POVERTY RATES An additional six countries (namely, Armenia, Colombia, Iraq, Kyrgyz Republic, Lao PDR, and the Republic of Yemen) were selected to bring the regional distribution into balance (that is, two additional countries each for Europe and Central Asia and the Middle East and North Africa, and one country each for East Asia and Pacific and Latin America and the Caribbean. These six countries were selected from a sample of 46 countries with high rates of headcount poverty as reported in the 2008 World Development Report. The 46 countries had headcount poverty rates greater than 40 percent on the basis of the national poverty line or greater than 20 percent on 238 APPENDIX E SELECTION CRITERIA AND METHODOLOGY USED TO ASSESS THE QUALITY OF POVERTY ASSESSMENTS FROM 20 COUNTRIES the basis of the international poverty line. Of these 46 countries, 15 were eliminated due to a lack of poverty diagnostics during 2000–2013, and the 11 countries in the Africa Region were eliminated since 4 countries had been selected on the basis of the poverty headcount criteria. Of the remaining 20 countries, 6 were selected for the Review in accordance with those having the greatest number of publically available poverty diagnostic works. These six countries are Armenia, Colombia, Iraq, Kyrgyz Republic, and Republic of Yemen. WEAK DATA The list of 16 countries selected for the study on the basis of poverty headcount and poverty rates includes three weak data countries (namely, the Democratic Republic of Congo, Iraq, and Republic of Yemen) in Africa and the Middle East and North Africa. In order to cover at least one weak data country in each region, three additional weak data countries were added: Afghanistan, Guyana, and Papua New Guinea. They were selected from a list of the weak data countries, and two of these countries (i.e., Guyana, and Papua New Guinea) had no surveys available in PovCalNet during 2000–2012. Moldova is included in the list for the Europe and Central Asia Region. POVERTY DIAGNOSTIC WORK Many of the 20 countries selected for the study include a mixture of poverty assessments, notes, technical assistance, and other poverty diagnostic work. This amalgam allows for a preliminary assessment of the 2008 trend in Bank poverty diagnostic work away from full poverty assessments toward poverty notes, updates, and technical assistance. QUALITY ASSESSMENT METHODOLOGY AND TEMPLATE The methodology used for the assessment of the quality of the twenty PAs closely follows the 2004 Guidance Note. 3 The 2004 Guidance Note did specify that the PAs would be prepared “in close coordination with national institutions, partners and civil society groups” and would cover three topics: (a) an assessment of the poverty situation, (b) an analysis of the impact of growth and public actions on poverty, and (c) the appraisal of poverty monitoring and evaluation systems. Further, a PA was to include: (a) an analytical synthesis of existing knowledge on these three topics, (b) an identification of key knowledge gaps with respect to these three topics, and (c) new analysis that addresses selected gaps or complements existing work. The 2004 Guidance Note also calls for the consideration of qualitative data and sociological/anthropological studies (see para. 4), the inclusion of specialists in the relevant sectors on the task team (see para. 26), and the wide dissemination “of poverty work within the Bank and outside” (see para. 27). Importantly, the 2004 239 APPENDIX E SELECTION CRITERIA AND METHODOLOGY USED TO ASSESS THE QUALITY OF POVERTY ASSESSMENTS FROM 20 COUNTRIES Guidance Note states that good-practice PAs aim to inform good pro-poor policy and that “good technical analysis is a means to this end” (see para. 7), and the need for strong linkages between the PAs and CASs, lending operations, and nonlending activities is noted. Consistent with the World Bank’s 2004 guidance note on poverty assessments, the systematic review of the quality of the Poverty Assessment follows six criteria and 26 sub-criteria as below: • Surveys and Data. Did the PA provide sufficient background on the: ◦ data used to undertake the poverty diagnostics, including the survey type, year, location, and content (income, consumption, education, health, and/or other non-income indicators), ◦ survey methodology and whether the data were publically available, and, ◦ Institutional arrangements and capacity for the survey design and implementation and, where appropriate or needed, was capacity building planned and/or undertaken? • Poverty Profiles and Diagnostics. Did the PA: ◦ identify and provide an analytical synthesis of the poverty statistics and knowledge available from alternative sources, ◦ explain the methodology used to determine the PA’s poverty statistics including types of data used, poverty line(s) used, and possible alternative poverty lines, ◦ make use of analytical tools developed in recent years such as ADePT, poverty mapping, and micro-simulation, ◦ provide estimates of the levels and trends in poverty (i) at the national level and disaggregated by regions and social groups, and (ii) across income, consumption, and non-income indicators, ◦ examine extreme poverty, ◦ report any participatory assessments of poverty, ◦ examine the key drivers of income and non-income poverty at the national and regional levels and across social groups, ◦ consider the determinants of the changes in poverty incidence over time, including growth and distributional changes, and, ◦ explore the obstacles to poverty reduction nationwide, most-affected regions, and across social groups? • Institutions and Public Actions. Did the PA evaluate the: 240 APPENDIX E SELECTION CRITERIA AND METHODOLOGY USED TO ASSESS THE QUALITY OF POVERTY ASSESSMENTS FROM 20 COUNTRIES ◦ government’s response to poverty including its key institutions, strategies, funding, and programs for poverty reduction, ◦ impact of past policies (macroeconomic, structural and sectoral) and programs (including targeted poverty reduction, social protection, specific public expenditure, and other programs) on the well-being of the poor and different segments of the poor, and, ◦ government, donors, and other partners’ efforts to support empowerment of poor communities and participatory poverty reduction work? • Monitoring and Analysis System. Did the PA: ◦ consider whether the country’s systems and capacity to monitor and analyze trends in poverty are adequate and, where problems are identified, summarize possible remedial actions, ◦ assess whether sufficient systems and capacity exist to evaluate the poverty impact of policy and program interventions, and, ◦ check for participatory, qualitative, or other alternative assessments of poverty? • Recommendations. Did the PA: ◦ provide a concise and clearly prioritized set of recommendations for poverty reduction work, and, ◦ specify costs, possible sources of funding, administrative responsibilities, and timing for these poverty reduction measures? • Influence and Impact. Did the PA report on: ◦ the extent to which core government agencies, key donors and other partners were engaged in the design, data acquisition and analytical work, compilation, and review of the PA, ◦ the support, if any, provided to improving poverty monitoring, analysis, and evaluation, ◦ whether any longer-term capacity building process was considered and/or initiated, ◦ whether any support was provided to in-country participatory processes—for reaching broad consensus on methodologies, findings, strategies and priority actions—during and after conducting analytical work, and, ◦ whether the PA was (i) made publically available and widely disseminated through printed publications and electronic media, (ii) 241 APPENDIX E SELECTION CRITERIA AND METHODOLOGY USED TO ASSESS THE QUALITY OF POVERTY ASSESSMENTS FROM 20 COUNTRIES adequately discussed with government, donors, other partners, and the poor themselves, and (iii) linked with country-based and owned processes which aim to develop a poverty reduction strategy and/or inform public actions. List of Poverty Assessments The 20 PAs reviewed by the study were (with year of publication and year of survey in parentheses): • Poverty Status in Afghanistan (2010; 2008) • Armenia Poverty Update, Report No. 24339–AM (2002; 1999) • Bangladesh Poverty Assessment: Assessing a Decade of Progress in Reducing Poverty, Bangladesh Development Series Paper No. 31 (2013; 2010) • Brazil: Measuring Poverty Using Household Consumption, Report No. 36358–BR (2007; 2003) • From Poor Areas to Poor People: China’s Evolving Poverty Reduction Agenda (2009; 2003) • Columbia Poverty Report, Report No. 24524–CO (2002; 1999) • Democratic Republic of Congo Poverty Diagnostic: Report No. 36489–DRC (2007; 2005) • Arab Republic of Egypt Poverty Assessment Update, Report No. 39885–EG (2007; 2005) • Ethiopia: Well-Being and Poverty in Ethiopia, Report No. 29468–ET (2005; 1999) • Guyana Poverty Assessment: Accelerating Poverty Reduction, Report No. 43702–GY (2008; 2006) • Perspectives on Poverty In India: Stylized Facts from Survey Data (2011; 2005) • Indonesia: Making the New Indonesia Work for the Poor, Report No. 37349– ID (2006; 2004) • Confronting Poverty in Iraq (2011; 2007) • Kyrgyz Republic Enhancing Pro-poor Growth, Report No. 24638–KG, (2003; 2001) • Lao PDR Poverty Assessment Report: From Valleys to Hilltops – 15 Years of Poverty Reduction, Report No. 38083–LA, (2006; 2003) • Recession, Recovery and Poverty in Moldova, Report No. 28024–MD (2004; 2002) 242 APPENDIX E SELECTION CRITERIA AND METHODOLOGY USED TO ASSESS THE QUALITY OF POVERTY ASSESSMENTS FROM 20 COUNTRIES • Mozambique Beating the Odds: Sustaining Inclusion in a Growing Economy, Report No. 40048–MZ (2008; 2003) • Nigeria Poverty Assessment, Report No. 40903–NG, (2007; 2004) • Papua New Guinea: Poverty Assessment (2004; 1996) • Republic of Yemen Poverty Assessment (2007; 2006). 1 World Bank. 2004a. “Guidance Note on Poverty Assessments.” Washington, DC: World Bank. 2The Europe and Central Asia Region is an exception. During the period of 2000–12, most countries in the Region had at least two rounds of household survey data available in PovCalNet. Turkmenistan did not have surveys during the period of evaluation but it did not have a poverty assessment after 2001. Armenia, the Kyrgyz Republic, and Moldova were selected for the review of PAs. 3 The 2004 Guidance Note was intended to provide good practice guidance for the preparation of PAs. Its requirements were not binding and allowed for considerable flexibility in the content and focus of the PAs in accordance with country needs and specific circumstances. 243 Appendix F. Technical Note on the Methodology Used in the CASCRR/CPSCRR Review In order to assess general trends in the poverty focus of the Bank’s country strategies, this evaluation reviewed a sample of IEG’s Country Partnership Strategy Completion Report Reviews (CPSCRRs) and Country Assistance Evaluations (CAEs). Country Partnership Strategy Completion Report Reviews A typical CPSCRR is a desk study that discusses the relevance and implementation of the strategy (including projects and AAA) and assesses the extent to which the Bank achieved the strategic pillars and objectives within a given Country Partnership Strategy (CPS) period. Based on this assessment, it rates the overall outcome of each pillar or set of objectives. There are 105 countries that had CPSCRRs between FY2004–2013. 1 This evaluation examined the 66 countries 2 that had at least two CPSCRRs during this period. 3 To assess the information from the CPSCRRs, a data base was assembled covering basic country data, presence of poverty-related pillars and objectives, ratings of poverty related pillars, and monitoring of poverty related indicators. A list of CPS pillars and objectives was obtained from the Bank’s Business Warehouse systems. Poverty-related pillars and objectives were defined as those that directly focused on “poverty,” “education,” “health,” “social protection,” “social development,” “agriculture and rural development” and “basic infrastructure.” To determine which pillars and objectives fell into these broad categories, the team first conducted a comprehensive word search for key words and phrases. The list of words used in this search are shown in table F.1. 4 Following the word search, the team individually reviewed all pillars and objectives that contained the relevant wording identified by the search to ensure correct categorization and poverty focus. CPS indicators are typically designed to measure progress on CPS objectives. To determine which indicators were poverty-focused, the results indicators of CPSs (listed both in the CPSs and the Country Partnership Strategy Completion Reports) were individually reviewed. They were then categorized into the following groupings: Poverty, school enrollment, school quality, health access, health outcomes, social transfers, basic infrastructure, and “other.” 244 APPENDIX F TECHNICAL NOTE ON THE METHODOLOGY USED IN THE CASCRR/CPSCRR REVIEW Table F.1. Search Terms and Phrases for Poverty-Related Pillars and Objectives Poverty Educationb Health Social Social Agriculture Basic Reductiona Protectionc Developmentd and Rural Infrastructuree Development Poverty, Education, Health, Social Social Agriculture, Basic service, Poor School Mortality, protection, development, Rural, Social service, Immunization, Safety net, Pro-poor Farmers, Community Births, Social growth, Sharing Farm, Crop, service, Basic Maternal, assistance, growth, Yields, infrastructure, Mother, Social Participation in Irrigation, Water, HIV/AIDS, transfers, growth, Growth Small holder Sanitation, Nutrition CCT, more pro-poor, Electricity, Conditional Broad based Energy cash growth, Sharing transfer, the fruits of Pension, growth, Benefits, Inclusive Insurance, growth, Equity, Social Access, service, Inclusion, Community Exclusion, service, Equitable, Vulnerability, Services, Vulnerable Social a. The use of the term “poor” will lead to a pillar being categorized under the poverty column if the pillar or objective references increasing the income or the poor, or employment, and so on. This category excludes pillars that address “pro- poor growth” as this falls under Social Development. b. Excludes tertiary education. c. Many pillars and objectives with the term “social service” were categorized as being under the Social Development category. References to vulnerability to environmental damage, health epidemics, or violence have been excluded from this category. Vulnerability to epidemics is captured under the Health Category, and vulnerability to violence will be captured in the Social Development Category. d. Delivery of “unspecified services” and access to services were included under Social Development, this includes “urban services” and “public services.” This does not include specified services such as health or education services, which are captured under other categories. The search term “social” was used to identify objectives and pillars about cohesion and areas of social development not covered by other search terms. Objectives and pillars identified from a search of “poor” that were relevant to equity or access were included in this category. e. Energy efficiency, water management and sustainability were not included in the Infrastructure category. Access to water was included in this category. Country Assistance Evaluations Country Assistance Evaluations (CAEs) provide detailed discussions of the Bank’s engagement in a country over a broad time period, often covering multiple Country Assistance Strategies (CASs). The CAEs are based on field visits and include input from government and other stakeholders. Fourteen CAEs were reviewed for this evaluation. 5 The selected CAEs covered periods ranging from 7 to 15 years since 2003, all of which included at least 2 years of engagement post 2005. Accordingly, the CAEs aggregated Bank objectives over two or more CPS periods and rated 245 APPENDIX F TECHNICAL NOTE ON THE METHODOLOGY USED IN THE CASCRR/CPSCRR REVIEW outcomes during the total period, although in some cases they provided separate ratings for the different periods they covered. The fourteen available CAEs were reviewed primarily to identify examples of successful and unsuccessful results, which partly reflect implementation. 1This includes the six OECS countries (Antigua and Barbuda, St. Kitts and Nevis, Dominica, Grenada, St. Lucia, and St. Vincent and the Grenadines), who have a joint CAS and CASCRR. CPSCRRs also may cover final Country Assistance Strategies (CAS), CAS drafts, as well as Interim Strategy Notes (ISNs). 2 CASCRRs/CPSCRRs from the following countries were reviewed: Albania, Argentina, Armenia, Azerbaijan, Bangladesh, Belarus, Benin, Bhutan, Bosnia and Herzegovina, Brazil, Bulgaria, Burkina Faso, Chile, China, Colombia, Costa Rica, Croatia, Djibouti, Ethiopia, Gabon, The Gambia, Georgia, Ghana, Guatemala, Honduras, India, Indonesia, Jamaica, Jordan, Kenya, Kyrgyz Republic, Lesotho, Macedonia (former Yugoslav Republic of Macedonia), Malawi, Mauritania, Mexico, Moldova, Montenegro, Morocco, Mozambique, Nicaragua, Niger, Nigeria, OECS Countries, Pakistan, Peru Philippines, Poland, Romania, Russian Federation, Rwanda, São Tomé and Príncipe, Senegal, Serbia, South Africa, Sri Lanka, Tajikistan, Tanzania, Turkey, Uganda, Ukraine, Uruguay, Uzbekistan, Vietnam, Republic of Yemen, Zambia. 3 Because of variation in the timing of CPSs and CPSCRRs for different countries, period 1 and period 2 CPSCRRs are not divided by year, only the sequence within a given country. As a result, the years covered by a CPSCRR in period 1 in a given country may overlap with the years covered by a CPSCRR in period 2 in another country. 4 Various spellings and permutations of these words were searched for using STATA. All terms in the data set were put into lower case letters to facilitate searches, and all hyphens were removed. Some objectives and pillars fit under multiple sections are counted as such. 5 The CAEs covered include: Afghanistan Country Program Evaluation, 2002–11: An Evaluation of the World Bank Group Program (2012); Bangladesh Country Assistance Evaluation (2009); Cambodia: An IEG Country Assistance Evaluation 1999–2006 (2010); Egypt Country Assistance Evaluation, Fiscal 1999–2007 (2009); The World Bank in Georgia, 1993–2007 (CAE) (2008); Mozambique Country Program Evaluation (2011); The World Bank in Nepal, 2003–2008 (Country Program Evaluation) (2011); The World Bank in Nigeria 1998– 2007: Nigeria Country Assistance Evaluation (2010); Peru: Country Program Evaluation for the World Bank Group, 2003–09 (2011); Timor-Leste Country Program Evaluation, 2000– 2010 (2011); Uganda Country Assistance Evaluation 2001–07 (2009); The World Bank Group in West Bank and Gaza, 2001–2009 (2010). Liberia Country Program Evaluation (2012); Brazil Country Program Evaluation (2014). 246 Appendix G. A Technical Note on Calculating the Proxy for Poverty Focus in Bank Lending In order to broadly gauge the poverty focus of the Bank’s interventions, the information from the theme and sector code system was used to calculate a weighted ratio of the Bank’s lending, Development Policy Operations and Investment Lending, respectively, that goes to areas that are more directly related to poverty reduction as a share of total lending. The theme and sector code system, which was initiated in July 2002, provides the basis for analyzing and reporting on the content of Bank activities (e.g., Bank budget allocations to strategic goals and priority sectors). Theme and sector codes are assigned to policy lending, investment lending, economic and sector work, technical assistance, research services, client training, and other activities that directly serve the Bank’s external clients. 1 Activities that serve the Bank's internal needs—e.g., quality assurance, country assistance strategies, sector strategy papers, knowledge products, training of Bank staff—are not coded for sectors and themes. Sector codes relate more to the part of a country’s economy that is targeted or impacted by a Bank operation (e.g. Education, Energy & Mining, or Transportation, etc.) whereas theme codes have more to do with the strategic end-goal and development objectives of the Bank (e.g. HIV/AIDS, Rural Markets, Pollution Management and Environmental Health, etc.). Themes refer to the goals/objectives of Bank activities; they are also consistent with the Bank's corporate advocacy and global public goods priorities and are used to capture Bank support to the Millennium Development Goals. Themes are not methods/instruments of delivery or ways of doing business, nor are they reflections of the structure of networks in the Bank. They may also be somewhat overlapping (i.e., not mutually exclusive). 2 Out of the 82 total theme codes available, the 31 themes in italics in table G.1 were chosen because they relate to the poorest or most vulnerable populations and, thus, serve as a proxy for “poverty-focus” in Bank lending. Table G.1. World Bank Theme Codes Economic Management 20 Analysis of Economic Growth 21 Debt Management and Fiscal Sustainability 247 APPENDIX G A TECHNICAL NOTE ON CALCULATING THE PROXY FOR POVERTY FOCUS IN BANK LENDING 22 Economic Statistics, Modeling, and Forecasting 23 Macroeconomic Management 24 Other Economic Management Public Sector Governance 25 Administrative and Civil Service Reform 26 Decentralization 27 Public Expenditure, Financial Management, and Procurement 28 Tax Policy and Administration 29 Other Accountability/Anti-Corruption 30 Other Public Sector Governance 90 Managing for Development Results 94 E-Government Rule of Law 31 Access to Law and Justice 32 Judicial and Other Dispute Resolution Mechanisms 33 Law Reform 34 Legal Institutions for a Market Economy 35 Legal Services 36 Personal and Property Rights 37 Other Rule of Law Financial and Private Sector Development 38 Corporate Governance 39 Infrastructure Services for Private Sector Development 40 Regulation and Competition Policy 41 Micro, Small, and Medium Enterprise Support 42 International Financial Standards and Systems 43 State-Owned Enterprise Restructuring and Privatization 95 E-Services 96 Financial Consumer Protection and Financial Literacy 248 APPENDIX G A TECHNICAL NOTE ON CALCULATING THE PROXY FOR POVERTY FOCUS IN BANK LENDING 97 Anti-Money Laundering and Combating the Financing of Terrorism 98 Other Financial Sector Development 99 Other Private Sector Development Trade and Integration 45 Export Development and Competitiveness 47 Regional Integration 48 Technology Diffusion 49 Trade Facilitation and Market Access 50 Other Trade and Integration Social Protection, Labor, and Risk Management 51 Improving Labor Markets 52 Natural Disaster Management 53 Poverty Strategy, Analysis, and Monitoring 54 Social Safety Nets (Social Assistance and Social Care Services) 55 Social Protection and Labor Policy and Systems (not available in BW) 56 Other Social Protection and Risk Management Income Support for Old Age, Disability and Survivorship Social Development, Gender, and Inclusion 57 Participation and Civic Engagement 58 Conflict Prevention and Post-Conflict Reconstruction 59 Gender 60 Indigenous Peoples 62 Other Social Development 100 Social Inclusion Human Development 63 Child Health 64 Other Communicable Diseases 65 Education for All 66 Education for the Knowledge Economy 249 APPENDIX G A TECHNICAL NOTE ON CALCULATING THE PROXY FOR POVERTY FOCUS IN BANK LENDING 67 Health System Performance 68 Nutrition and Food Security 69 Population and Reproductive Health 70 Other Human Development 88 HIV/AIDS 89 Non-Communicable Diseases and Injury 92 Malaria 93 Tuberculosis Urban Development 71 Urban Services and Housing for the Poor 72 Municipal Finance 73 Municipal Governance and Institution Building 74 Other Urban Development 101 Urban Planning and Housing Policy 102 City-Wide Infrastructure and Service Delivery 103 Urban Economic Development 104 Cultural Heritage Rural Development 75 Rural Markets 76 Rural Non-Farm Income Generation 77 Rural Policies and Institutions 78 Rural Services and Infrastructure 79 Other Rural Development 91 Global Food Crisis Response Environment and Natural Resources Management 80 Biodiversity 81 Climate Change 82 Environmental Policies and Institutions 83 Land Administration and Management 250 APPENDIX G A TECHNICAL NOTE ON CALCULATING THE PROXY FOR POVERTY FOCUS IN BANK LENDING 84 Pollution Management and Environmental Health 85 Water Resources Management 86 Other Environmental and Natural Resources Management Calculating the thematically weighted commitment amount (table G.2) for each project entailed the following steps: first, the list of all Bank projects was downloaded into Excel from the Bank’s Business Warehouse (BW) Operations database using Project Theme Detail Report 2.c.2.1, which provides thematic coding for all Bank operations of all instrument types and all product lines. Second, filters (in BW) were used to retrieve only those projects that were approved between FY2000–12 and that have any of the 31 “poverty-focused” theme codes listed for themes 1, 2, 3, 4, or 5. Next, the “# Projects” variable in BW was modified under the macros function in BEx Analyzer/Excel to automatically calculate the “poverty themed” weight for each project; that is, for every project that is comprised of any of the 31 “poverty-focused” themes, the theme percentages for those poverty themes is added up to get one total “poverty theme” percentage. This number represents the total share of that project that was designated by the Bank’s project team to go towards those “poverty focused” themes. This percentage was then multiplied by the Commitment Amount variable to yield the “Poverty-Weighted Commitment Amount (PWCA).” Finally, all of the PWCAs were aggregated across all projects and analyzed according to their various product lines and instrument types (e.g. International Bank for Reconstruction and Development [IBRD], International Development Association [IDA], Adjustment, Investment) to yield the total amount of money committed to any of the 31 poverty-focused themes in any given year. The example below shows the calculations for a generic project. Here, the PWCA for Project A is derived by: PWCA= [(Theme 1 + Theme 4 + Theme 5) x Commitment Amount] = [(.2+.25+.05) x 25.2] = 12.6. For Project B, the calculation would be: PWCA= [(Theme 2 + Theme 4) x Commitment Amount] = [(.2+.15) x 73.2] = 25.62. 1http://intresources.worldbank.org/INTOPCS/Resources/theme- sector_quickref_guide.pdf. 2http://intranet.worldbank.org/WBSITE/INTRANET/UNITS/INTOPCS/0,,contentMDK: 21806751~pagePK:51455324~piPK:51455326~theSitePK:380832,00.html. 251 Table G.2. Example of Calculation of Poverty Weighted Commitment Amount Project Theme 1 Theme Theme 2 Theme Theme 3 Theme Theme 4 Theme Theme 5 Theme Com PWCA 1% 2% 3% 4% 5% Amt. Project Rural 20 Decentralization 30 Administrative 10 Rural 25 Participation 5 25.20 12.6 A services and and civil policies and civic infrastructure service and engagement reform institutions Project State-owned 10 Rural services 20 Infrastructure 30 Child 15 Climate 25 73.20 25.62 B enterprise and services for Health change restructuring infrastructure private sector and development privatization 252 Appendix H. Poverty Data Availability in Micro Data Catalog and Cost Estimation of Statistical System Improvement Micro Data Catalog The International Household Survey Network (IHSN) provides a public catalog with listings for 4,224 surveys and censuses conducted in low- and middle-income countries, along with metadata, survey questionnaires, manuals, and reports, if available. Roughly 1,200 are identified as containing income, consumption, or other poverty-related data. Data cannot be obtained directly from the IHSN catalog, but a third of the micro data sets are available from external data repositories. Through the Central Micro data catalog 1 the World Bank provides a public listing of 1,762 surveys of various types along with their metadata and related documentation including 60 LSMS surveys. Micro data can be downloaded directly from the source, while others require approval before data are made available. A larger micro data catalog is maintained for internal use by World Bank staff. It contains more than 5,000 surveys, of which somewhat less than a thousand contain household income or expenditure data. This catalog includes the harmonized datasets used for much of the World Bank’s poverty work. However restrictive covenants placed on many of underlying surveys by the owners or producers as a condition of their transfer to the World Bank limit their use even within the World Bank. Because the World Bank lacks a uniform policy on data acquisition, the terms of use for data sets acquired from countries or other producers are often ad hoc and poorly documented. There are also issues related to sharing data inside the Bank. Internal sector silos hinder data harmonization and dissemination. Within the Bank, data appear to have not been adequately shared among staff working on poverty. The lack of a consistent flow of poverty data among the Bank networks and regions was cited as a key internal constraint to obtaining data beyond the external constraints posed by countries. In the Staff Survey, Focus Groups, and the case studies, Bank staff frequently noted that data needed to be shared better across the regions and the networks. Although about half (49 percent) of the internal survey respondents believed that Bank staff “almost always” or “frequently” shared poverty data, 253 APPENDIX H POVERTY DATA AVAILABILITY IN MICRO DATA CATALOG AND COST ESTIMATION OF STATISTICAL SYSTEM IMPROVEMENT slightly less than half (43 percent) maintained that this occurred only “sometimes,” occasionally,” or “hardly ever.” The creation of the Global Poverty Working Group, regional data platforms, and the internal Micro data catalog containing harmonized datasets from different sources are all efforts to address some of these problems. However work on the harmonized databases, important for conducting cross-country studies and monitoring progress over time, remains fragmented, undertaken by different units at different times and without sufficient resources committed to ensure continuity. 2 Bank-sponsored surveys are typically planned by local Bank teams with national counterparts. Although the teams may receive advice from the LSMS unit or other experienced staff, decisions about sampling, content of the questionnaire, and processing of the data may be made without adequate consideration of comparability or adherence to recommended standards. Likewise the terms under which data can be used and disseminated are often idiosyncratic, failing to take into account the Bank’s and general public interest in open access. Surveys obtained by World Bank staff for use in their work, may or may not be made available to other units or included in the Micro data Catalog. And because of the lack of common standards, considerable effort must be expended on ex post harmonization of data sets, which may still fail to resolve inconsistencies. The Cost of Improving Statistical System The World Bank and other partners have attempted to estimate the shortfall in funding needed to upgrade statistical systems in developing countries. In 2004 the consensus estimate included in the Marrakech Action Plan for Statistics (MAPS) was $140–$160 million per year in additional resources for low-income and lower- middle-income countries. The MAPS cost estimates and the work program helped donors coordinate their support and encouraged partnerships for statistical capacity building. Although MAPS achieved its objectives and stimulated additional investments in statistics, it did not address all of the shortcomings in national statistical systems. The World Health Organization (WHO) has recently estimated the cost of scaling up investment in global civil registration and vital statistics (CRVS) for 73 countries to be on the order of $3.82 billion over 10 years (WHO 2012). Taking into account domestic contributions and recurrent expenditures, they conclude that an additional $1.99 billion is required over a 10–year period, or an average of $199 million per year; $40 million more than the Marrakech estimate per annum, for just one important statistical tool. 254 APPENDIX H POVERTY DATA AVAILABILITY IN MICRO DATA CATALOG AND COST ESTIMATION OF STATISTICAL SYSTEM IMPROVEMENT Morten Jerven at Simon Fraser University (2014) has estimated that using survey techniques for collecting data for the eight MDGs along with a population census would cost $1.08 billion per annum, assuming that the majority of MDG development data are survey-based and that poverty analysis requires annual collection of survey data (Jerven 2014). Demombynes and Sandefur (2014) have refined Jerven’s estimate by identifying the funding gap taking into account preexisting spending on household surveys. Focusing on countries below $2,000 per capita GDP in PPP dollars yields a total cost to international donors of closing all remaining survey gaps of less than $300 million per annum, which they point out is a fairly small share of global aid budgets. The differences between these estimates demonstrate, as much as anything, the lack of consensus on the statistical tools needed to monitor a comprehensive development agenda. There are efforts underway with involvement from the World Bank to estimate the full cost and incremental investment needed to monitor the post-2015 agenda using broadly agreed assumptions and estimation methods. These estimates will be presented to the Financing for Development Conference scheduled for July 2015. References Demombynes, G., and J. Sandefur. 2014. Costing a Data Revolution, Data for Development Viewpoint. Copenhagen: Copenhagen Consensus Center. Jerven, M. 2014. “Benefits and Costs of Data for Development: Targets for the Post-2015 Development Agenda.” Data for Development Assessment Paper, Copenhagen Consensus Center, Copenhagen, Denmark. WHO (World Health Organization). 2012. The Case for Investment in Civil Registration and Vital Statistics Systems. Geneva: Health Metrics Network. 1See World Bank Central Microdata Catalogue http://microdata.worldbank.org/index.php/catalog/central. 2 Someof the efforts to construct harmonized data sets and analytical tools include CLSP and I2D2 (DECRG), SEDLAC and LABLAC (LAC Region), ECAPOV (ECS Region), and SHIP (Africa Region). 255 Bibliography Benjamin, Dwayne, Loren Brandt, and John Giles. 2011. “Did Higher Inequality Impede Growth in Rural China?” Economic Journal 121 (557): 1281–1309. Berg, Andrew, Jonathan D. Ostry, and Jeromin Zettelmeyer. 2012. “What Makes Growth Sustained?” Journal of Development Economics 98 (2): 149–166. Bourguignon, François. 2004. “The Poverty-Growth-Inequality Triangle.” Working Paper 125, Indian Council for Research on International Economic Relations, New Delhi. Christiaensen, Luc, and Lionel Demery. 2007. Down to Earth: Agriculture and Poverty Reduction in Africa. Washington, DC: World Bank. Dang, Hai-Anh, Peter F. Lanjouw, and Umar Serajuddin. 2014a. “Updating Poverty Estimates at Frequent Intervals in the Absence of Consumption Data: Methods and Illustration with Reference to a Middle-Income Country.” Policy Research Working Paper 7043, World Bank, Washington, DC. Dang, Hai-Anh, Peter Lanjouw, Jill Luoto, and David McKenzie. 2014b. “Using Repeated Cross- Sections to Explore Movements in and out of Poverty.” Journal of Development Economics 107: 112–128. Davies, Tim. 2014. “Open Data in Developing Countries-Emerging Insights from Phase I.” Open Data in Developing Countries Working Papers 2. Berlin: World Wide Web Foundation. Demombynes, G., and J. Sandefur. 2014. “Costing a Data Revolution,” Data for Development Viewpoint. Copenhagen: Copenhagen Consensus Center. Dollar, David, Tatjana Kleinberg, and Aart Kraay. 2014. “Growth, Inequality, and Social Welfare: Cross-Country Evidence.” Policy Research Working Paper 6842, World Bank, Washington, DC. Ferreira, Francisco H. G. and Nora Lustig, eds. Forthcoming. “Appraising Cross-National Income Inequality Databases.” Journal of Economic Inequality, Special Issue. Ferreira, Francisco H. G., Phillippe G. Leite, and Martin Ravallion. 2010. “Poverty Reduction Without Economic Growth? Explaining Brazil’s Poverty Dynamics, 1985–2004.” Journal of Development Economics 93: 20–36. Filmer, Deon, and Norbert Schady. 2006. Getting Girls into School: Evidence from a Scholarship Program in Cambodia. Washington, DC: World Bank. Fosu, Augustin Kwasi. 2010. “Income Distribution and Growth’s Ability to Reduce Poverty: Evidence from Rural and Urban African Economies,” UNU–WIDER Working Paper 2010/92, United Nations University, Helinski, Finland. 256 BIBLIOGRAPHY Fox, Louise. 2008. Beating the Odds: Sustaining Inclusion in Mozambique's Growing Economy. Washington, DC: World Bank. Gasparini, Leonardo, Federico Gutiérrez, and Leopoldo Tornarolli. 2005. “Growth and Income Poverty in Latin America and the Caribbean: Evidence from Household Surveys,” CEDLAS, Working Paper 30, CEDLAS, Universidad Nacional de La Plata, Argentina. Herzer, Dierk, and Sebastian Vollmer. 2012. “Inequality and Growth: Evidence from Panel Cointegration,” Journal of Economic Inequality 10: 489–503. IEAG (Independent Expert Advisory Group on a Data Revolution for Sustainable Development). 2014. A World That Counts: Mobilising The Data Revolution for Sustainable Development. New York, NY: Independent Advisory Group Secretariat. IEG (Independent Evaluation Group). 2010. Poverty Reduction Support Credits: An Evaluation of World Bank Support. Washington, DC: World Bank. _____. 2011a. Assessing IFC’s Poverty Focus and Results. Washington, DC: World Bank. _____. 2011b. The World Bank’s Involvement in Global and Regional Partnership Programs: An Independent Assessment. Washington, DC: World Bank. _____. 2011c. The World Bank Group's Response to the Global Economic Crisis: Phase I. Washington, DC: World Bank. _____. 2011d. Social Safety Nets: An Evaluation of World Bank Support, 2000–2010. Washington, DC: World Bank. _____.2012a. The World Bank Group’s Response to the Global Economic Crisis—Phase II. Washington, DC: World Bank. _____. 2012b. Liberia Country Program Evaluation: 2004–2011. Washington, D.C.: World Bank. _____.2012c. World Bank Group Impact Evaluations: Relevance and Effectiveness. Washington, DC: World Bank. _____.2013. “Senegal-Country Assistance Strategy for the period FY07–FY10.” Country Assistance Strategy Completion Report Review. Washington, DC: World Bank. _____.2014. Learning and Results in World Bank Operations: How the Bank Learns. Washington, DC: World Bank. _____. 2015. Results and Performance of the World Bank Group 2014. Washington, DC: World Bank. IMF (International Monetary Fund) 2012. “Data Quality Assessment Framework—Generic Framework.”Washington, DC: IMF. Inchauste, Gabriela, Sergio Olivieri, Jaime Saavedra, and Hernan Winkler. 2012. “What is Behind the Decline in Poverty Since 2000,” Policy Research Working Paper 6199, World Bank, Washington, DC. 257 BIBLIOGRAPHY IYCN (Infant and Young Child Nutrition) Project. 2011. Consulting with Caregivers: Formative Research to Determine the Barriers and Facilitators to Optimal Infant and Young Child Feeding in Three Regions of Malawi. Washington, DC: IYCN. Jolliffe, Dean. 2010. Poverty Status in Afghanistan—A Profile Based on the National Risk and Vulnerability Assessment (NRVA) 2007–08. Washington, DC: World Bank. Knowles, Stephan. 2005. “Inequality and Economic Growth: The Empirical Relationship Reconsidered in the Light of Comparable Data.” Journal of Development Studies 41 (1): 135–159. Kraay, Aart. 2006. “When Is Growth Pro-Poor? Evidence from a Panel of Countries.” Journal of Development Economics 80: 198–227. Libresco, Brett, and Thomas O'Brien. 2007. Philippines—Client Perspectives on Elements of World Bank Support. Washington, DC: World Bank Loayza, Norman V., and Claudio Raddatz. 2006. “The Composition of Growth Matters for Poverty Alleviation.” Policy Research Working Paper 4077, World Bank. Lustig, Nora, Omar Arias, and Jamele Rigolini. 2002. “Poverty Reduction and Economic Growth: A Two-Way Causality,” IDB Publications 53919, Inter-American Development Bank, Washington, DC. McKinsey Global Institute. 2011. Big Data: The Next Frontier for Innovation, Competition, and Productivity. Washington, DC: McKinsey Global Institute. OECD (Organisation for Economic Co-Operation and Development). 2007. “Promoting Pro-Poor Growth: Private Sector Development.” In Promoting Pro-Poor Growth: Policy Guidance for Donors, OECD, DAC Guidelines and Reference Series, OECD Publishing. OED (Operations Evaluation Department). 2005. 2004 Annual Review of Development Effectiveness: The World Bank's Contributions to Poverty Reduction. Washington, DC: World Bank QAG (Quality Assurance Group). 2003. QAG Assessment Quality of ESW in FY02. Washington, DC: World Bank. Ravallion, Martin. 1997. “Can High Inequality Development Countries Escape Absolute Poverty?” Economics Letters 56: 51–57. _____. 2001. “Growth, Inequality and Poverty: Looking Beyond Averages,” World Development 29(11): 1803–1815. _____. 2007. “Inequality is Bad for the Poor.” In Inequality and Poverty Re-Examined, J. Micklewright and S. Jenkins (eds.). Oxford: Oxford University Press Ravallion, Martin, and Gaurav Datt. 1996. “How Important to India’s Poor is the Sectoral Composition of Economic Growth?” World Bank Economic Review 10 (1): 1–25. _____. 2002. “Is India's Economic Growth Leaving the Poor Behind?” Journal of Economic Perspectives 16 (3): 89–108. 258 BIBLIOGRAPHY Ravallion, Martin, and Shaohua Chen. 1997. “What Can New Survey Data Tell Us About Recent Changes in Distribution and Poverty?” World Bank Economic Review 11(2): 357–82. ______. 2007. “China’s (Uneven) Progress Against Poverty.” Journal of Development Economics 82 (1): 1–42. ______.2009. ‘The Impact of the Global Financial Crisis on the World’s Poorest.” Vox CEPR’s Policy Portal (blog), April 30. http://www.voxeu.org/article/impact-global-financial-crisis-world- s-poorest Thomson, Anne, Graham Eele, Felix Schmieding. 2013. Independent Evaluation of the International Household Survey Network (IHSN) and Accelerated Data Program (ADP). Final Report. Oxford, UK: Oxford Policy Management. UNDP (United Nations Development Programme). 2004. Unleashing Entrepreneurship: Making Business Work for the Poor. New York, NY: UNDP. Voitchovsky, Sarah. 2005. “Does the Profile of Income Inequality Matter for Economic Growth?: Distinguishing Between the Effects of Inequality in Different Parts of the Income Distribution,” Journal of Economic Growth 10: 273–296. World Bank. 1990. World Development Report 1990: Poverty. New York: Oxford University Press. _____. 2001a. World Development Report (WDR) 2000/2001: Attacking Poverty. Washington, DC: World Bank. _____. 2001b. Romania—Country Assistance Strategy. Washington, DC: World Bank. _____. 2001c. Country Assistance Strategy for the People’s Republic of Bangladesh. Washington, DC: World Bank. _____. 2001d. Egypt Country Assistance Strategy. Washington, DC: World Bank. _____. 2002a. Poverty in Bangladesh: Building on Progress. Washington, DC: World Bank. _____. 2002b. Peru—Country Assistance Strategy. Washington, DC: World Bank. _____. 2003a. Poverty in Guatemala. Washington, DC: World Bank. _____. 2003b. Senegal—Country Assistance Strategy FY03–06. Washington, DC: World Bank. _____. 2003c. Kyrgyz Republic: Enhancing Pro-poor Growth. Washington, DC: World Bank. _____. 2004a. “Guidance Note on Poverty Assessments.” Washington, DC: World Bank. _____. 2004b. Papua New Guinea: Poverty Assessment. Washington, DC: World Bank. _____. 2004c. Recession, Recovery and Poverty in Moldova. Washington, DC: World Bank. 259 BIBLIOGRAPHY _____. 2005a. 2004 Annual Review of Development Effectiveness (ARDE): The World Bank's Contributions to Poverty Reduction. Washington, DC: World Bank. _____. 2005b. World Development Report 2006: Equity and Development. Washington, DC: World Bank. _____. 2005c. Ethiopia—Well-Being and Poverty in Ethiopia: The Role of Agriculture and Agency. Washington, DC: World Bank. _____. 2005d. Peru Opportunities for All: Peru Poverty Assessment. Washington, DC: World Bank. _____. 2005e. Philippines—Country Assistance Strategy for the Period FY2006–2008. Washington, DC: World Bank. _____. 2005f. IDA Country Assistance Strategy for the Lao PDR. Washington, DC: World Bank. _____. 2005g. IBRD Country Assistance for the Republic of Guatemala. Washington, DC: World Bank. _____. 2005h. Country Partnership Strategy for the Federal Republic of Nigeria (2005–2009). Washington, DC: World Bank. _____. 2005i. Country Assistance Strategy for the Arab Republic of Egypt for the Period FY06–FY09. Washington, DC: World Bank. _____. 2006a. Making the New Indonesia Work for the Poor. Washington, DC: World Bank. _____. 2006b. Lao PDR—Poverty Assessment Report: From Valley's to Hilltops—15 Years of Poverty Reduction. Washington, DC: World Bank. _____. 2006c. Country Assistance Strategy for the People’s Republic of Bangladesh for the Period FY06–09. Washington, DC: World Bank. _____. 2006d. Romania—Country Partnership Strategy. Washington, DC: World Bank. _____. 2006e. Peru—Country Partnership Strategy for the Period FY12–FY16. Washington, DC: World Bank. _____. 2007a. Country Assistance Strategy of the World Bank for the Republic of Malawi. Washington, DC: World Bank. _____. 2007b. Senegal—Country Assistance Strategy FY07–10. Washington, DC: World Bank. _____. 2007a. Brazil—Measuring Poverty Using Household Consumption. Washington, DC: World Bank. _____. 2007b. Arab Republic of Egypt: Poverty Assessment Update. Washington, DC: World Bank. _____. 2007c. Republic of Yemen Poverty Assessment. Washington, DC: World Bank. _____. 2007d. Malawi Poverty and Vulnerability Assessment: Investing in Our Future. Washington, DC: World Bank. 260 BIBLIOGRAPHY _____. 2008a. Guyana Poverty Assessment: Accelerating Poverty Reduction. Washington, DC: World Bank. _____. 2008b. Poverty Assessment for Bangladesh: Creating Opportunities and Bridging the East-West Divide. Washington, DC: World Bank. _____. 2008c. Sénégal—Diagnostic de la Pauvreté. Washington, DC: World Bank. _____. 2008d. IBRD and IFC Country Partnership Strategy for the Republic of Guatemala. Washington, DC: World Bank. _____. 2009a. Nigeria—Employment and Growth Study. Washington, DC: World Bank. _____. 2009b. Guatemala—Poverty Assessment: Good Performance at Low Levels. Washington, DC: World Bank. _____. 2009c. Philippines—Country Assistance Strategy for the Period FY2010–2012. Washington, DC: World Bank. _____. 2009d. Romania—Country Partnership Strategy for the period July 2009–June 2013. Washington, DC: World Bank. _____. 2009e. Country Partnership Strategy for the Federal Republic of Nigeria (2010–2013). Washington, DC: World Bank. _____. 2010a. World Development Report 2010: Development and Climate Change. Washington, DC: World Bank. _____. 2010b. Philippines: Fostering More Inclusive Growth. Washington, DC: World Bank. _____. 2010c. Poverty in Lao PDR 2008: Lao Expenditure and Consumption Survey 1992/03–2007/08. Washington, DC: World Bank. _____. 2010d. Country Assistance Strategy for the People’s Republic of Bangladesh for the Period FY11–14. Washington, DC: World Bank. _____. 2011a. World Bank Corporate Scorecard 2011: Integrated Results and Performance Framework. Washington, DC: World Bank. _____. 2011b. World Bank for Results 2011. Washington, DC: World Bank. _____. 2011c. 2011 Philippines Development Report: Generating Inclusive Growth to Uplift the Poor. Washington, DC: World Bank. _____. 2011d. Perspectives on Poverty in India: Stylized Facts from Survey Data. Washington, DC: World Bank. _____. 2012a. World Development Report 2013: Jobs. Washington, DC: World Bank. _____. 2012b. IDA Country Partnership Strategy for the Lao PDR. Washington, DC: World Bank. 261 BIBLIOGRAPHY _____. 2012c. Peru—Country Partnership Strategy for the Period FY2007–FY2011. Washington, DC: World Bank. _____. 2012d. IBRD and IFC Country Partnership Strategy for the Republic of Guatemala for the Period 2013–2016. Washington, DC: World Bank. _____. 2012e. Country Assistance Strategy for the Republic of Malawi for the Period FY13–FY16. Washington, DC: World Bank. _____. 2012f. Interim Strategy Note for the Arab Republic of Egypt. Washington, DC: World Bank. _____. 2013a. Philippine Development Report: Creating More and Better Jobs. Washington, DC: World Bank Group. _____. 2013b. Bangladesh Poverty Assessment: Assessing a Decade of Progress in Reducing Poverty 2000– 2010. Washington, DC: World Bank. _____. 2013c. Country Partnership Strategy for the Federal Republic of Nigeria for the Period FY 2014–FY 2017. Washington, DC: World Bank. _____. 2013d. Country Partnership Strategy FY13–17 for the Republic of Senegal. Washington, DC: World Bank. _____. 2014a. A Measured Approach to Ending Poverty and Boosting Shared Prosperity: Concepts, Data, and the Twin Goals. Washington, DC: World Bank. _____. 2014b. The World Bank Group Strategy. Washington, DC: World Bank. _____. 2014c. Trust Fund for Statistical Capacity Building, Annual Progress Report: April 1, 2013 – March 31. TFSCB Internal Management Committee and TFSCB Administration Unit paper series. _____. 2014d. Prosperity for All / Ending Extreme Poverty: A Note for the World Bank Group Spring Meetings 2014. Washington, DC: World Bank. _____. 2014e. Romania—Country Partnership Strategy. Washington, DC: World Bank. _____. 2015a. Purchasing Power Parities and the Real Size of World Economies: A Comprehensive Report of the 2011 International Comparison Program. Washington, DC: World Bank. _____. 2015b. A Measured Approach to Ending Poverty and Boosting Shared Prosperity: Concepts, Data, and the Twin Goals. Policy Research Report. Washington, DC: World Bank. World Bank and IMF (International Monetary Fund). 2004. Global Monitoring Report 2004: Policies and Actions for Achieving the Millennium Development Goals and Related Outcomes. Washington, DC: World Bank. _____. 2015. Global Monitoring Report 2014/2015: Ending Poverty and Sharing Prosperity. Washington, DC: World Bank. World Bank and UNICEF (United Nation’s Children’s Fund). 2009. Romania—Rapid Assessment of the Impact of the Economic Crisis on Poverty. Washington, DC: World Bank. 262