OPERATIONSPOLICY AND COUNTRYSERVICES WORLDBANK June 24,2005 ABBREVIATIONS ACRONYMS AND AAA Analytical and Advisory Activities ADB Asian Development Bank AfDB African Development Bank APR Annual Progress Report BP Bank Procedures (Operational Manual statement) CAFTA Central American Free Trade Agreement CAS Country Assistance Strategy CDF Comprehensive Development Framework CEM Country Economic Memorandum CFAA Country Financial Accountability Assessment CIDA Canadian International Development Agency CMU Country Management Unit CODE Committee on Development Effectiveness CPAR Country Procurement Assessment Report CPIA Country Policy and Institutional Assessment DANIDA Danish International Development Agency DFID Department for International Development DPL Development Policy Lending DPR Development Policy Review EC European Commission ESAF Enhanced Structural Adjustment Facility ESSD Environmental and Socially Sustainable Development Network ESW Economic and Sector work EC European Commission EU European Union FY Fiscal Year HD Human Development Network HIPC Highly Indebted Poor Country ICA Investment Climate Assessment ICR Implementation Completion Report IDA International Development Association IDB Inter-American Development Bank IMF International Monetary Fund I-PRSP Interim Poverty Reduction Strategy Paper ISN Interim Strategy Note JSA Joint Staff Assessment JSAN Joint Staff Advisory Note KfW Kreditanstalt fir Wiederaufbau (German Aid Agency) LCR Latin America and the Caribbean Region LICs Lower-income countries MCA Millennium Challenge Account MDBs Multilateral Development Banks MDGs Millennium Development Goals M&E Monitoring and Evaluation MOU Memorandum of Understanding MTEF Medium Term Expenditure Framework ODA Overseas Development Assistance OED Operations Evaluation Department OP Operational Policy (Operational Manual statement) OPCCE Country Economics Unit in OPCS OPCS Operations Policy and Country Services PA Poverty Assessment PAF Performance Assessment Framework PBA Program Based Approach PEAP Poverty Eradication Action Plan PEM Public Expenditure Management PER Public Expenditure Review PFM Public Financial Management PREM Poverty Reduction and Economic Management Network PRGF Poverty Reduction Growth Facility PRSC Poverty Reduction Support Credit PRSP Poverty Reduction Strategy Paper PSD Private Sector Development PSIA Poverty and Social Impact Analysis SIDA Swedish International Development Agency SPA Strategic Partnership with Africa SWAps Sector Wide Approaches TA Technical Assistance TAL Technical Assistance Loan UNDP United Nations Development Programme WDR World Development Report Executive Summary ............................................................................................................ix I Introduction . ......................................................................................................................1 11 PRSC Objectives and Trends . ........................................................................................3 11. PRSC Approach.............................................................................................................3 B. PRSC Trends ................................................................................................................. 5 1. Share and Number ...................................................................................................... 5 2. Country Selectivity .................................................................................................... 7 3. Links to CAS..................................................................................................... 11 4. Analytic Underpinnings .......................................................................................... 15 5. Environmental Aspects ......................................................................................17 6. Cost Considerations .................................................................................................18 7. Conclusiot~............................................................................................................... 19 I11 PRSP Alignment: Content and Coverage . .................................................................19 A . Operationalizing the PRSP.......................................................................................... 20 1. Elaborating Strategies ..............................................................................................20 2. Policy Matrices as Performance Frameworks..........................................................21 3. Improving Strategic Budgeting................................................................................ 27 4. Monitoring and Evaluating PRSP Results ................................................................ 29 B. Supporting the PRSP'S Corr1pl-e17ensil!eProg1.a~....................................................... 33 1. General Trends.........................................................................................................33 2. Public Financial Management..................................................................................36 3. Service Delivery in Social Sectors........................................................................... 38 4. Selectivity and Sequencing of Coverage ................................................................. 40 5. Implications of Multisectoral Approach ..................................................................43 IV Harmonization and Predictability . .............................................................................44 A. Donor Coordination and Harmonization .................................................................... 45 1. Coordination and Harmonization Modalities........................................................... 45 2. Approaches to Harmonization .................................................................................47 3. Risks and Challenges in Harmonization .................................................................. 50 B. Inlproving Predictability ............................................................................................. 51 1 . Programmatic Alignment with the PRSP ................................................................ 51 2. The Record on Predictability ...................................................................................54 3. Design Options to Further Improve Predictability...................................................56 4. Timing of Disbursements and Aligning with Country Processes ............................58 V Conclusions . .....................................................................................................................60 Box 1. Definitions.................................................................................................................... 4 Box 2. Lessons Learned from Previous Policy-Based Lending .............................................. 9 Box 3. Linking PRSCs to Results-Based CASs..................................................................... 12 Box4. BeyondProjects......................................................................................................... 13 Box 5. Analytic Underpinnings of Honduras PRSC......................................................... 16 Box 6. Poverty and Social Impact Analyses in Madagascar ................................................. 17 Box 7. Ethiopia's Country-Driven Policy Matrix.................................................................. 25 Box 8. Trends in PRSC Policy Matrices................................................................................ 26 Box 9. Issues in Adopting Outcomes-Based Conditionality ................................................. 32 Box 10. Public Financial Management Reforms in Burkina Faso...........................................38 Box 11. Selectivity and Sequencing of Sectoral Programs in the Benin PRSC ......................41 Box 12. Coverage of Rural Development and the Growth Agenda in PRSCs........................ 42 Box 13. Donor Harmonization in Burkina Faso ...................................................................... 49 Box 14. Different Approaches to Triggers .............................................................................. 52 Box 15. Nepal PRSC Program in "Wait and See Mode" ........................................................ 53 Figure 1. Number of IDA Operations. by Instrument................................................................ 6 Figure 2. IDA Commitment Volumes ($million). by Instrument.............................................. 7 Figure 3. PRSC Share in IDA DPL ........................................................................................... 7 Figure 4. Share of DPL in IDA .................................................................................................. 7 Figure 5. Average CPIA Ratings ............................................................................................. 11 Figure 6. Country Business Model for Low-Income Countries .............................................. 13 Figure 7. PRSC Share of New IDA Commitments.................................................................. 15 Figure 8. Preparation and Supervision Costs ($thousands) ..................................................... 18 Figure 9. Commitment per Cost Ratio..................................................................................... 19 Figure 10.SectoralContent Among PRSCs.............................................................................. 34 Figure 11.Thematic Content Among PRSCs ............................................................................ 34 Figure 12.SectoralDistribution of Binding Measures in PRSCs and Other DPL Operations..35 Figure 13.ThematicDistribution of Binding Measures in PRSCs and Other DPL Operations 35 Figure 14.Increased Predictability and Volume of Bank Disbursements to PRSC Countries..................................................................................................... 55 Table 1. What is the PRSC Stocktaking Paper Assessing? ...................................................... 2 Table 2. PRSC Countries and Commitments ($million). FYO1-05..........................................6 Table 3. CPIA Quintiles and PRSC Countries ....................................................................... 11 Table 4. Drawing on Analytic Work to Design Reform Measures in Honduras ...................23 Table 5. Examples of Linking Actions to Outcomes ............................................................. 31 Table 6. Evolution of Sectoral Content of PRSC Binding and Non-Binding Measures ........36 Table 7. Share of PFM Conditions in PRSCs......................................................................... 37 Table 8. Harmonized Budget Support in PRSC Countries..................................................... 48 Table 9. Commitments and Actual Disbursements ................................................................ 54 Table 10. Regularity of PRSC Operations ............................................................................... 56 Table 11. Example of Synchronized Cycle for Budget, PRSP Review, and Budget Support .59 This report was prepared by John Factora with key contributions from Mozarnrnal Hoque, Ken Green, Zhanar Abdildina, and Gero Verheyen. Linda Van Gelder and Satu Kahkonen were peer reviewers and Allison Berg, Harold Bedoya, Jaime Jaramillo, Janet Entwistle, and Jan Walliser provided feedback throughout the preparation. It benefited from the insights of the Bank's PRSC teams, including Julie Van Domelen, Helene Grandvoinnet, Jean van Houtte, Suzane Otis, Manuela Ferro, Rachidi Radji, Jacques Morisset, Jeeva Perumalpillai-Essex, Rachidi Radji, Carlos Cavalcanti, Kees Kostermans, Trina Haque, Robert Utz, Neeta Sirur, Ulrich Lachler, as well as from those who commented on earlier drafts. Pansy Chintha and Philomene Koya assisted with the document processing and Patricia Rogers provided editorial support. The report was prepared under the supervision of Stefan Koeberle, Manager of External (non-Bank) readers should be aware that publicly available documents that are referenced in this paper may be accessed on the World Bank's external website- www.wor1dbank.org-by searching at Publications and Documents; or obtained through the World Bank Infoshop in Washington, DC, or its Public Information Centers in many member countries. 1. The World Bank introduced poverty reduction support credits (PRSCs) in May 2001 as one of IDA'S main vehicles to support low-income countries in implementing their Poverty Reduction Strategy Papers (PRSPs).' The PRSC is a programmatic approach to development policy lending (DPL) in low-income countries, which typically consists of three or four annual, single-tranche operations, phased to support the government's medium-term development objectives. It embodies the Bank's contribution to the emerging experience with budget support, which is commonly understood to mean the provision of financial assistance directly to a partner country's budget on a regular basis, using the country's own financial management systems and budget procedures. The PRSC's overarching goal of bolstering the PRSP approach, and particularly of reinforcing the country ownership embedded in the initiative, was to be achieved through several mutually reinforcing objectives: Operationalize and Implement PRSPs. To help implement a medium-term program that builds upon and draws from the PRSP's priorities and objectives. Resource Predictability. To improve resource predictability through medium-term annual commitments that are disbursed in line with domestic planning, budgeting, and review processes. Donor Harmonization. To provide a framework for coordination with development partners, particularly with the IMF through its Poverty Reduction Growth Facility (PRGF) and other budget support donors. 2. Purpose of the Retrospective. This retrospective reviews the extent to which the PRSC approach is meeting these objectives. Although only a few years have elapsed since the PRSC was introduced, the paper takes stock of the Bank's emerging experience with PRSC implementation, highlighting operational challenges and evolving good practice. The stocktaking builds on preliminary findings of the Programmatic Adjustment Lending Retrospective as well as recent country-specific assessment^.^ It responds to the interest by various stakeholders in the PRSC experience expressed during the recent introduction of the Bank's operational policy on development policy lending,' internal and external reviews of the PRSP process, and assessments of budget support by various stakeholders. Its timing is also coordinated to provide a contribution to the Bank's ongoing review of ~onditionality.~While much has already been learned within the Bank about how PRSCs can best support PRSP implementation, the approach is still relatively new, and operational lessons will continue to evolve as the Bank gains further operational experience. I Interim Guidelines,for PoverQ Reduction Support Credits (PRSC): Discussion Draft, (IDAlSecM2001-025l), World Bank, April 10,2001. Programmatic Adjustment Lending Retrospective, OPCS, World Bank, January 2004; and Poverty Reduction Support Credit in Uganda:Results oj'a Stocktaking Study, World Bank, June 2004. OPIBP 8.60, Development Policy Lending, was approved by the Executive Directors in August 2004. It replaced OD 8.60, Adjustment Lending, incorporated the interim guidelines for PRSCs and no longer recognizes distinct instrument types for policy-based lending. The PRSC is covered by the new policy, and should be understood as a programmatic approach to development policy lending in IDA-eligible countries. 4 See Review of WorldBunk Conditionality:Issues Note (CODE-2005-0002), World Bank, January 11,2005. 3. Overview of Findings. The stocktaking paper concludes that the PRSC has emerged as a robust and effective lending instrument that is broadly delivering on its promises and meeting original expectations. It has evolved as an integral part of the PRSP process, helping to improve the operational relevance of the PRSP and providing key support to the government's policy reform agenda. As PRSPs necessarily reflect individual country priorities, PRSCs have varied in respect to content, scope and coverage, and their role within country portfolios. They share a strategic selectivity in favor of good performers, strong analytic underpinnings and a sustained, yet flexible support for complex medium-term institutional and policy improvements. Most PRSC programs have embraced the harmonization agenda, providing an effective vehicle for the Bank to align with other budget support programs around a single country-owned framework. The PRSC's flexible design has established a strong record of stabilizing and improving resource predictability while maintaining a clear performance orientation. While most PRSC programs to date have been broadly implemented as anticipated, the Bank has been able to respond to uneven country performance in a graduated manner by adapting prior actions, delaying operations or, more rarely, reducing the amount. Moving forward, as countries enter the second generation of PRSPs, PRSCs will continue to be important in strengthening domestic accountability by helping to further operationalize the PRSP; shifting the policy dialogue away from traditional conditionality to mutual accountability for development results; improving alignment with the annual PRSP, budget, and planning cycles; and strengthening the underlying systems, institutions, and local capacity needed to enhance medium-term strategic planning, public expenditure management, program implementation, and results monitoring and evaluation. A. General Trends 4. The introduction of the PRSCs has been gradual, in line with initial expectations. From FYOl to February 2005, the Board approved a total of 33 PRSCs for $3.0 billion. The 20 countries that currently have PRSCs represent less than half of the 45 countries that have submitted full PRSPs to the Board. Within IDA, the PRSC has emerged as a significant share of new development policy lending operations and commitments-in FY05, 50 percent and 55 percent, respectively. While the number of PRSC programs has grown, the overall weight of Bank investment lending has not been affected significantly. Since PRSCs were introduced, the share of IDA development policy lending has fluctuated at about 27 percent, an increase from the 16 percent recorded in FYOO, but matching the average share of development policy lending seen in the early 1990s. 5. Country Selectivity. Key to the effectiveness of the PRSC approach has been strategic selectivity based on known success factors. In line with initial expectations, PRSCs have been used in countries that show solid country performance, commitment to an effective reform program, and readiness to improve fiduciary and public financial management arrangements. All but three of the 20 current PRSC countries undertook Bank-supported policy-based lending within the three years leading up to the PRSC program. Of those pre-PRSC adjustment operations that were evaluated by OED, 95 percent were rated as satisfactory or highly satisfactory. PRSC countries have typically been above-average performers that receive relatively high rankings in the Country Policy and Institutional Assessment (CPIA), reflecting their progress in maintaining macroeconomic stability and implementing structural reforms, and particularly in improving their planning and budget management systems5 6. ESW Underpinnings. All PRSCs have been underpinned by a set of integrative economic and sector work (ESW) that assesses public financial management systems and key social, structural, and sectoral policies. These include "core" ESW products, namely the Country Financial Accountability Assessment (CFAA), the Public Expenditure Review (PER), Country Procurement Assessment Report (CPAR), and the Country Economic Memorandum (CEM) or Development Policy Review (DPR). In addition, PRSCs frequently draw on Poverty Assessments (PA), Poverty and Social Impact Analysis (PSIA), and environmental and sector- specific analyses. These studies have typically been undertaken by the Bank, but also by the borrower, or development partners, either on their own or collaboratively. While the PRSP provides the overall framework and policy direction for Bank assistance, the ESW informs the policy dialogue that guides the design of the PRSC and helps policy makers formulate a clear agenda to achieve the country's development objectives. B. PRSP Alignment: Content and Coverage 7. The breadth of policy areas covered by PRSCs testifies to the instrument's ability to address both sectoral constraints and cross-sectoral issues. On average, PRSCs cover eight sectors and seven themes, compared to four sectors and four themes for other DPLs. All PRSCs focus on public sector reforms; close to half of all prior actions for PRSCs are public sector-related, and nearly 65 percent of those focus specifically on improving public financial management. Public expenditure and budget management, including their links to PRSP priorities, has been a particular focus of the PRSC's policy dialogue. Sector specific issues in education, health, agriculture, and water and sanitation have been well covered in PRSCs, comprising nearly 35 percent of prior actions for PRSCs; service delivery improvements in these sectors have also been supported by PRSC financing and monitoring and evaluation. 8. Alignment with PRSPs. In general, PRSCs are closely aligned with the country's PRSP, with some variation depending particularly on how well the PRSP prioritizes and operationalizes the country's medium-term development objectives. Factors that have been found to determine the specific scope of any particular PRSC operation include the Bank's comparative advantage and the potential for synergy with other donors; the operation's consistency with the Bank's CAS; the readiness of the specific sector for budget support based on institutional capacity; the robustness of the analytic work in the sector; the extent to which the government has articulated a results framework within a sectoral strategy or medium-term expenditure framework; and the sequencing of sector-specific activities over time. As a PRSC program advances with subsequent operations, the coverage often expands into new areas while deepening the policy dialogue in existing areas. 9. Operationalizing PRSPs. The PRSP process is intended to allow governments to formulate the country's developmental objectives and to steer donor support around them. However, many PRSPs are still not quite at the stage where they can provide the level of ' The CPlA is an aggregate of performance ratings on public sector management, economic management, and social and econon~icpolicies. The CPIA ratings for the year prior to the calendar year of PRSC approval were used (e.g. for projects approved in CY 2004, scores for CPIA 2003) since the ratings are usually finalized by end of the year and become the operational ratings for the next calendar year. informed operational and policy details necessary for a medium-term program of annual reforms. Through a results-based policy dialogue and coordinated budget support program, including the PRSC, the Bank and donors help governments identify or refine specific priority actions that support the main objectives of the PRSP. However, this approach raises a risk that donors may establish parallel mechanisms for aid delivery to fill capacity gaps, thus not fully leveraging the country ownership embedded in the PRSP process and possibly even displacing domestic processes. In some African countries, notably Ethiopia and Mozambique, governments have driven a country-owned process of identifying priority areas for reform and then articulated sequenced measures with monitorable indicators in a policy matrix linked to the PRSP, with individual donors supporting a subset of the government's policy matrix. The approach, while perhaps not feasible in countries with greater capacity constraints, nonetheless provides a noteworthy model of country ownership and commitment. C. Donor Harmonization and Resource Predictability 10. Aid effectiveness in low-income countries demands that donor support be aligned around the country's own processes. Experience has shown that PRSCs can make a substantial contribution toward greater alignment across different dimensions. Particularly in countries where donor budget support represents a significant share of total budgetary resources, it is important for the PRSC to be aligned with other donor budget support programs, and for budget support to be aligned with the government's annual PRSP, budget, and planning cycles. The PRSC has emerged as an important tool for mainstreaming the implementation of the PRSP in order to make it a more relevant framework for government planning. In countries that have advanced most toward donor alignment around country processes, such as Mozambique and Tanzania, the tangible result is donor provision of timely and predictable multi-sectoral budget support that is informed and guided by the annual PRSP process and improves coordination across ministries and levels of government. 11. Donor Harmonization. In almost all PRSC countries, but particularly in Africa, the PRSC has become a useful platform to facilitate donor coordination and harmonization around the common framework provided by countries' PRSPs and annual progress reports. The aim is for governments to negotiate effectively one broad comprehensive reform program, with lower costs in terms of time and effort, preparation, reporting and monitoring. The content of all donor programs should be consistent with the PRSP priorities and with each other, and streamlined. Budget support donors in African countries have gone the furthest in systematically establishing formal "budget support groups" with Memoranda of Understandings (MOUs) signed by the government and donor partners (although coordination differs with country circumstances). MOUs aim to clarify the rules of the game (including mechanisms for the resolution of differences between donors or between donors and the government) and encourage donors to comply with their harmonization and alignment commitments, for instance by establishing a framework for mutual accountability and peer pressure. On balance, governments tend to favor greater harmonization that reduces transaction costs with individual donors. However, some governments prefer to diversify funding sources which may not be feasible if conditionality is perfectly harmonized. 12. Predictability. The PRSC approach reflects an evolution in the design of policy-based lending. Drawing on lessons learned during the past two decades, the PRSC approach aims for a more predictable, medium-term flow of resources; a more flexible, step-by-step approach that is based on completed (as opposed to promised future) actions and that takes into account changing circumstances; limited binding measures related to high-priority actions; a strong results orientation; and a move from intrusive conditionality to a set of progress benchmarks based on country priorities. The evidence suggests that the PRSC has done well in delivering on these promises. At the same time, the objective of resource predictability has been balanced against a regular assessment of performance: PRSC programs have been suspended when country circumstances did not warrant a continuation (as in Guyana and Nepal). 13. Programmatic Design. PRSCs have followed a programmatic design with a flexible, step-by-step approach that takes into account changing circumstances. The PRSC approach moves beyond traditional ex-ante conditionality based on promises, which has been criticized as ineffective and corrosive of government ownership: all but two PRSCs have been single- tranche operations based on a limited set of up-front, completed actions. On balance, this design has helped ensure an annual cycle that is aligned to the PRSP process and provides a more predictable flow of resources that improves budget planning and execution. 14. Policy Matrix. Policy matrices typically go well beyond simply listing conditions for disbursement, often because they are used as a reference framework for a subset of government policies supported by the Bank, and to help to spell out implementation steps for achieving the program objectives. This is borne out by the evidence on differences between prior actions, triggers and indicative benchmarks. Even though the number of prior actions has fallen in the most advanced PRSC countries, such as Uganda and Burkina Faso, the overall length of the policy matrix has sometimes increased since the number of indicative benchmarks, or nonbinding actions, has grown. This has partly been explained by the fact that the sectoral coverage of PRSCs, and therefore the scope of the policy program, typically broadens as the program matures. In many countries, government officials have also found a more specific implementation framework helpful, particularly in addressing cross-sectoral issues. However, in view of capacity constraints, the challenge is to avoid overloading the policy matrix, particularly when it is driven by efforts to include specific donor preferences in a harmonized framework. 15. Triggers. For PRSCs, like other programmatic operations, triggers (or expected prior actions) have emerged as an important design feature to reconcile the tension between flexibility and discipline in multiyear programs. Triggers allow the Bank to make an overall bottom-line assessment of whether sufficient progress has been made to move to the subsequent operation, and typically provide the basis for prior actions in the subsequent operation. The formulation of triggers varies from country to country. Some triggers are defined in an open-ended manner (e.g., "satisfactory progress in implementing public procurement reforms"), while others are specific measures that eventually constitute the prior actions for subsequent operations. The average number of triggers has been reduced from 15 in FYO1 to 1 1 in FY04. 16. Graduated Response. In most PRSCs, triggers were converted into prior actions and met before approval of the next operation, indicating that the programs were progressing as intended. In the few cases in which implementation deviated significantly from expectations (five percent of triggers were not met at the time of Board approval), the Bank responded in a graduated manner. In principle, failure to meet triggers could result in a reduction of the commitment amount or a delay in the next operation. However, in these cases, the Bank determined that enough progress had been made in other areas to justify moving to the next operation. 17. Budget and Planning Cycles. PRSC programs are typically expected to be closely aligned with the government's budget and planning systems and timetables. The aim of facilitating the government's ability to plan and execute the budget could be met by confirming commitment amounts at a time when the government is finalizing the budget, and then disbursing at the beginning of the fiscal year to make budget resources available for program implementation. A key step is for the government to align these planning processes with the APR to improve links between program priorities and budget allocation. The ideal sequencing of performance reviews, budget preparation, donor confirmation of commitments, donor disbursement, and budget reporting has not yet been put into practice. For the Bank, disbursement delays have generally been caused by government delays in meeting policy actions, but these delays have always stayed within the limits of the government's fiscal year. 18. Selectivity. PRSCs have been extended to a limited number of well-performing countries that are likely to make the most effective use of regular budget support. There is a broad consensus that the PRSC "label" should remain a selective approach that is reserved for good performers with a proven track record of program implementation. These programs would build on a demonstrated commitment to a nationally-owned strategy that draws on a robust set of analytic and diagnostic work. The Bank of course has the discretion to support genuine turn-around cases through a PRSC, a decision that should be made on a country by country basis, without relying on any single performance indicator. There are also opportunities in weaker performers to improve institutional environments and strengthen country performance with development policy lending operations that could be more narrowly focused and potentially less frequent than typical PRSCS.~ See Good Practice Note on Development Policy Operations and Program Conditionality in Fragile States, OPCS, World Bank, June 2005. I . The World Bank introduced poverty reduction support credits (PRSCs) in May 2001 as one vehicle to help countries implement their Poverty Reduction Strategy Papers (PRSPs). This paper takes stock of the Bank's implementation experience and assesses the extent to which PRSCs are fulfilling their objectives. 2. Background. The last decade has seen a fundamental change in the paradigm on development assistance: recognition of the importance of country leadership and ownership of the development process, the significance of a good policy environment for aid effectiveness; stronger partnerships among all stakeholders, and a focus on results as a way to measure development impact, including the achievement of targets set out in the Millennium Development Goals (MDGs).' These principles, articulated in the Comprehensive Development Framework (CDF). helped bring about a rethinking of Bank operational work in the late 1990s. The Bank's "country program model" sought to move away from project-centered approaches to one in which the country was the critical unit of account and the focus of donor collaboration on opcrational work such as diagnostic and analytic assessments; monitoring and evaluation; and nontraditional, medium-term, program-based approaches such as sectonvide approaches (SWAps) and programmatic policy-based lending.' Coupled with a growing body of empirical evidcnce and lessons of experience, the increased emphasis on country ownership also helped trigger discussion on, and a change in, conditionality in policy-based lending, from its traditional role of inducing and leveraging reforms to a means of reaching mutual understanding with governments on reform priorities.' 3. PRSP Initiative. The PRSP initiative, launched in 1999, was seen to embody a "new way of doing business" in low-income countries. The main objective was to assist countries to improve the poverty impact of public policy and expenditure by developing and implementing more effective ~trategies.~This process requires adhesion to four principles-country ownership, comprehensiveness and long-term perspective, results orientation, and partnership. Governments of low-income countries are expected to take a lead role in setting out their country's vision and defining its policies and programs to reduce poverty, working collaboratively with domestic stakeholders and donor partners to tailor the PRSP to country conditions and circumstances. The PRSP initiative has become the donor community's central framework for development assistance and policy engagement in low-income countries.' 4. Budget Support for PRSPs. For many donors, coordinated budget support is seen as an effective way to support country-owned PRSPs. In current usage, it typically refers to policy- based, non-project financing that is predictable, annual, with a medium-term timeframe, and based on progress in achieving the objectives of the government's PRSP. For the Bank, budget I Information on the CDF and MDGs can be obtained using the search function of the World Bank website. S~rppot.ting('ountrj Developnzent: WorldBank Role and Instrunzents in Low- and Middle-Incotne Countries (DC/2000-19).pg. 2. .' Stefan Koeberle. "Should Policy-Based Lending Still Involve Conditionality?" World Bank Research Observer, 2003. Go to ~y~v\r..~~orldbank.ore/povertyfor more information on the PRSP process. ' Sec Poi~erh.Redzrction Strates Papers -Detailed Analysis of Progress in Ir~zplementation,a series of annual reports by thc World Bank since 2001. support comes in the form of PRSCs, an approach to the development policy lending (DPL) instrument. 5. Purpose and Scope of Paper. This paper takes stock of the Bank's experience with the PRSC and highlights lessons learned, key issues, emerging practices, and recommendations. Because very few countries have completed their first series of PRSC operations, the general focus of the paper is on design and implementation issues (in the PRSC causality chain presented in Table 1, the inputs and outputs of the PRSC). The paper does not attempt to assess development impact, although it highlights preliminary medium-term outcomes from the more mature PRSC programs. The paper draws on country specific assessments; Implementation Completion Reports (ICRs); Bank reports and notes, including the Programmatic Adjustment Lending Retro~pective;~and recent good practice notes on the design of DPLs. Internal and external workshops, and discussions with staff, clients, and partners, were also important sources of information. The PRSC approach is still relatively new; thus "lessons learned" are moving targets, subject to change with further operational experience. Table 1. What is the PRSC Stocktaking Paper Assessing? Inputs Outputs Outcomes Impact The design and preparation of the The PRSC reinforces the Intermediate PRSC entails: PRSP process and country Fulfillment of medium-term Helping to improve the ownership by: PRSP objectives, such as operational relevance of PRSP improved: (specify, prioritize, monitor, and Helping to implement a Service delivery Sustainable improve budgetary linkages) strategic subset of the Climate for growth growth and PRSP through financing, Public administration poverty Aligning with and drawing from policy dialogue, and alleviation PRSP content monitoring Lonn Term Consistency with CAS's multiyear commitments and 2015 MDG targets reached in: Improving resource priorities, and complementarity predictability through Primary education with other operations medium-term Maternal health Adequate analytic underpinning commitments disbursed HIV. other disease Child mortality Medium-term programmatic in line with domestic Extreme poverty approach that is flexible and planning, budgeting, and Gender equality based on ex post conditionality review processes Alignment with domestic Ensuring harmonized and processes and timetables coordinated donor budget Donor collaboration support programs Scope of the stocktakingpaper Note: This causalitychain is based on a similarpresentation in OED's Poverty Reduction StrategyInitiative: An Independent Evaluation of'the WorldBar~li's Support Through2003. 6. Timing. The timing of the paper reflects a critical juncture not only in the PRSP initiative, but also in Bank policy-based lending. Several early PRSP countries have completed their first-generation PRSP and associated PRSC series, and are taking stock of implementation ' Programmatic Adjustment Lending Retrospective (PALR) (SecM2003-333), World Bank, March 2004. experience to guide their next medium-term strategy. The Bank is also looking to initiate PRSC programs in new PRSP countries, including those where the potential for donor harmonization around budget support programs is promising. The growing interest in PRSP implementation experience is manifest in the recent internal and external reviews of the PRSP process, and in various stakeholders' ongoing assessments of budget upp port.^ The stocktaking is also a critical input in the Bank's ongoing review of conditionality.' 7 . Organization. Following this introductory section, Chapter I1 summarizes the origins and conceptual framework of the PRSC, and presents general trends. Chapter I11 discusses the extent to which PRSCs have helped to operationalize and implement PRSPs, and Chapter IV looks at their record in fostering donor harmonization and resource predictability. Chapter V presents concluding thoughts. 11. PRSC OBJECTIVES AND TRENDS 8. The PRSC is a specialized form of development policy lending (what the Bank formerly termed adjustment lending) designed to reflect many of the lessons of development experience while specifically supporting implementation of a PRSP (Box 1 provides definitions for development policy lending and other lending modalities that are discussed in this paper). This section explains the objectives of the PRSC and explains how it has fulfilled the assumptions under which it was launched. A. PRSC Approach 9. The overarching goal of the PRSC is to bolster the PRSP process, contributing to stronger country ownership and to improved development outcomes and achievement of Millennium Development Goals (MDGs) targets. The PRSC has three mutually reinforcing objectives: Operationalize and Implement PRSPs. To help implement a medium-term program that builds upon and draws from the priorities and objectives of the PRSP through budget support financing and policy dialogue. Resource Predictability. To improve resource predictability through medium-term commitments that are disbursed in line with domestic planning, budgeting, and review processes. 7 CIDA, CIDA's Action Plan to Promote Harmonization (2004-2005), September 2004; DFID and OECDIDAC, Evaluation Framework of General Budget Support, 2004; DFID, General Budget Support Evaluability Study: Country Case Studies: Mozambique & Andra Pradesh, Volume 11, Oxford Policy Management and Overseas Development Institute, 2004; DFID, Poverty Reduction Budget Support, May 2004; European Commission, EC Budget Support: An Innovative Approach to Conditionality, European Commission DG Development, February 2005; S. Jones, and A. Lawson, Movingfrom Projects to Programmatic Aid, OED Working Paper Series No. 5, 2000; A. Lawson, and D. Booth, Evaluation Frameworkfor General Budget Support, ODI, May 2004; OECD, Harmonizing Donor Practices for Effective Aid Delivery, Volume 2: Budget Support, Sector-wide Approaches and Capaciq Development in Public Financial Management; Strategic Partnership for Africa- Budget Support Working Group, 2005; Survey of the Alignment of Budget Support and Balance of Payments Support with National PRS Processes, February 2005; and USAID, General Budget Support: A New Approach from Other Donors. Summer Seminar Series, July 2004. Review of World Bank Conditionality: Issues Note, (CODE 2005-0002), January 2005; and Conditionality Revisited: Concepts, Experiences and Lessons, World Bank, May 2005. Donor Harmonization. To provide a framework for coordination with development partners, particularly with the IMF through its Poverty Reduction Growth Facility (PRGF) and other budget support donors. Box 1. Definitions Policy-based lending: Provision of untied donor resources directly through the government's budget, using the government's own financial management, procurement, auditing, and implementation processes and systems. This lending is based on a set of policy or institutional reforms (termed conditionality when set out ex ante in a multitranche operation, or prior actions when identified ex post in a single tranche operation). The Bank's term for policy-based lending is developmentpolicy lending (DPL). Budget support: A programmatic form of policy-based lending-that is, predictable, annual, medium- term disbursement of untied resources to the budget. Budget support is typically based on an agreed set of performance indicators in the form of institutional or policy reform measures, or output or outcome indicators. It is increasingly aligned with national strategies, such as the PRSP and its annual cycle of performance review, and with domestic planning and budgeting processes. Poverty reduction support credit (PRSC): A form of budget support used by the World Bank in low- income countries that is closely tied to the PRSP and typically consists of a series of three or four single- tranche operations. 1 Program-based approach (PBA): A general term that refers to coordinated donor support to implement a comprehensive program (e.g., PRSP) or a specific sectorlthematic strategy (e.g., in education, environment) that is country-led and relies on a single budget framework and domestic processes. PBAs include budget support and SWAps. Sectonvide approach (SWAP): An approach that involves financing of a specific sector through various financing modalities, including budget support and investment financing. 10. Comparison with other DPL Operations. While the PRSC is not a distinct financing instrument, its objectives represent clear expectations beyond other DPL operations. DPL operations typically provide targeted, focused support within a well-defined timeframe, without necessarily involving the promise of regular annual disbursements. The PRSC, with its medium- term timeframe, flexible step-by-step approach, multisectoral scope, and alignment with country systems and processes, is designed to bolster the PRSP process, complement other budget support programs, and improve the predictability of Bank engagement through policy dialogue and resource flows, ideally on an annual basis. This in turn is expected to improve country performance, the stability of the policy framework, and the sustainability of reforms. PRSCs also provide a suitable vehicle for transferring relatively large medium- to long-term resources to consistently well-performing countries, reflecting an emerging consensus on aid effectiveness to achieve MDG targets. 11. Programmatic Design. The PRSC's programmatic approach also differentiates it from other policy-based lending. Bank adjustment lending operations in the 1980s and 1990s were typically multitranche, multiyear operations based on ex-ante conditionality that often proved insufficiently flexible to reflect changing circumstances. Traditional conditionality was also viewed as donor imposed, insufficiently tailored to specific country circumstances, and overlooking important considerations of political economy and government capacity and commitment. The PRSC was designed to avoid these problems. The programmatic design reinforces country ownership as each operation is based on actions already completed (rather than promised future actions). Traditional conditionality has been replaced by a performance framework that reflects an understanding between the Bank and government on the priorities of the country's reform program and the financing required. The programmatic design allows the Bank to show tangible evidence of progress in achieving the objectives of the medium-term program. Each PRSC is linked to previous and subsequent operations in the series, with actions and measures in subsequent PRSCs building on those of the preceding operation. The program's policy matrix lays out the sequencing and progression of implementation benchmarks from one operation to the next, with each set of reforms linked to results-focused indicators. 12. Flexibility. The medium-term program consists of a series of three or four annual operations, each approved by the Board once the country completes a limited and prioritized set of actions that are set out and summarized in the operation's policy matrix. Each subsequent PRSC builds on the previous one. Requiring satisfactory progress on expected policy actions, or "triggers," allows for a bottom-line assessment of a country's progress toward its reform objectives. If the Bank considers that progress has been satisfactory, it can move on to the subsequent operation. If some of the triggers are not met because local circumstances have changed (e.g., additional analytical evidence and change in reform strategy), the reform program may need to be modified. However, if insufficient progress is due to inadequate country performance, the Bank may reduce the loan amount or postpone the operation until satisfactory progress is demonstrated. B. PRSC Trends 13. When the PRSC was launched, there were several assumptions about how its role and use would evolve. The share and number of PRSCs was expected to grow over time; the instrument would be used mainly in better-performing countries; each PRSC would be strongly linked to the country's CAS; each one would be based on adequate analytic underpinnings; and the instrument would be cost-effective. This section examines the extent to which these assumptions have been borne out. 1. Share and Number 14. As expected, the share and number of PRSCs have grown gradually as countries put in place core diagnostic work and medium-term policy frameworks, and improved their institutional capacity. From FYOl to March 2005, the Board has approved 33 PRSCs to 20 countries, committing approximately US$3.0 billion to support PRSP implementation. About 60 percent of PRSC operations and commitments, and half of all PRSC countries, have been in the Africa Region. Uganda and Burkina Faso, which have both reached their fourth operation, are the furthest along in the program (see Table 2). With the exception of Guyana and possibly Sri Lanka and Nepal, all PRSC countries have advanced or are expected to advance to subsequent operations on a near-annual basis. Table 2. PRSC Countries and Commitments (%miilionj.FY01-05 ,, Region I Country (PRSC 1 / PRSC 2 ) PRSC 3 PRSC 4 I Total 1 % of Total Fenin 2 0 20 1 1% --- ~ - - ~ ---- - -- - - - -- Rwanda - 30 -- I 1 3 + lzl- 282 9% 1Uganda 150 1 150 150 150 600 20% Sub-total AFR 887 ) 590 200 210 1,887 62% IEAP 1 Vietnam 250 1 100 1 100 450 -~ -- - - -. - . ~ - - - .- 1- -So/, .., ~ Sub-total EAP 250 / 100 1 100 j 450 15% --. l~lbania i 18 1 10 ! 48 2%) r I bLA - -. 20 -- I- - - - Armenia 20 20 1% Sub-total ECA I 40 1 18 1 10 68 1 20/;; 1 ~ u v a n a 12 I I 17 - 1 0 4'4 - - -- - -- LCR Honduras 59 - 59 2YO -- - - - - Nicaragua ---- 70 70 2% Sub-total LCR 141 , 141 5% 70 70 29.0 - -- 300 1 300 10% - 1 Sri Lanka - 125 I 125 4% Sub-total SAR 495 495 16% Grand Total 1,813 708 1 310 1 210 3,041 100% Source: SAP Business Warehouse Note: Approvals as of end-February 2005 15. PRSC in IDA Context. The growth in PRSC programs must be viewed in the broader context of increased IDA lending. From FY90-00, the Bank approved an average of 101 investment and 15 policy-based operations a year for IDA. Since FYO1, those numbers have increased to 120 and 21, respectively (see Figure 1). IDA lending volumes have also increased in the past five years over those of the 1990s. Within these gradual increases, the volume of policy-based lending has fluctuated, and Figure 1. Number of IDA Operations, by Instrument PRSCs have accounted for a growing share of the overall volume of IDA DPL (see Figure 2). The average PRSC 1 DPL operations commitment size has been U S 9 2 million, slighily higher than the US$84 million average commitment size of other DPL operations approved during the same period. Actual PRSC commitments range widely, from US$300 million in Pakistan to US$12 million in Guyana. Of the 33 PRSCs approved, 10 have included IDA grant financing: four were fully funded by IDA grants and the remaining six were partly 1q90-94 .iig 1')95~00a\g 3001 11102 21101 2004 1 Source: SAP Business Warehouse funded by IDA grants, ranging from 24 percent to 77 percent of total commitments.' 16. Share of Commitments. While Figure 2. IDA Commitment Volumes ($million), by Instrument -- -- - the share and number of PRSCs has 1 grown, they have not displaced PRSC investment lending. The share of Other DPLoperalionr PRSCs in IDA DPL has increased (see Blnvestmnl operations Figure 3), but the overall share of Bank investment lending in IDA lending has not changed significantly. Since PRSCs were introduced, the share of IDA policy-based lending has hovered at about 27 percent, an increase from the historically low 16 percent recorded in FYOO, but matching the average share of DPL of the early 1990s (see Figure 4). The latest 1990-94avg 1995-00avg. 2001 2002 Z(M3 corporate lending projections are that Source: SAP Business Warehouse the share of IDA policy-based lending should remain steady for the next three y Figure 3. PRSC Share in IDA DPL Figure 4. Share of DPL in IDA IDA DPL Opcra~ions IDA DPLC'ummitments Source: SAP Busincss Warehouse Sozrrce: SAP Business Warehouse Note: Approvals as of end-February 2005 Note: Appro\ als as of end-February 2005 2. Country Selectivity 17. The I'RSC was expected to be used in better-performing countries-that is, countries with a stable macroeconomic framework, adequate financial management systems and the commitment to improve these systems further, and demonstrated commitment to a robust program of structural and social reforms. The PRSC is equivalent to a vote of confidence, with the Bank trusting countries with proven track records to allocate untied resources consistent with ' The four PRSCs that were fully funded by IDA grants were Ethiopia PRSC 1, Burkina Faso PRSC 3, and Uganda PRSC 3 and 4. 1 0FY05-07 0z11look,/i)i. the Bankwide Slitrre of Developn~rnlPolicy Lending, First Annual Report (SecM2005- 0128). April 2005. PRSP priorities. The Strategic Framework for IDA'S Assistance to Africa makes this explicit, suggesting that larger resource flows could be provided to support programs through countries' budget and planning processes, if the countries had improved capacity and a strong track record." This section examines the extent to which this expectation was borne out. 18. Macroeconomic Framework. Like all development policy lending, PRSCs require an adequate macroeconomic policy framework. When the Bank assesses the adequacy of the macroeconomic policy framework, it takes IMF views, often expressed in the context of a PRGF arrangement, as an important input into this determinati~n.'~All but two PRSCs have proceeded in parallel with a PRGF: in Benin and Vietnam, the PRGFs had elapsed just prior to PRSC approval.13 In Mauritania, the Bank has put the PRSC on hold because of Bank and Fund concerns about the adequacy of the country's macroeconomic policies. In the future it may not always be possible for the two instruments to proceed together-while macroeconomic reforms are typically short-term, social and structural reforms are long-term and hence require more time to put in place. Moreover, the PRSC is limited almost exclusively to above-average performers with a full-PRSP and strong analytic underpinnings. By contrast, the IMF7sPRGF program, which replaced the earlier Enhanced Structural Adjustment Facility (ESAF), has included countries with an interim poverty reduction strategy (I-PRSP).14 For some countries, after a PRGF has supported effective macroeconomic stabilization, future PRSCs may support continuing structural and social reforms without a parallel PRGF, but with continued IMF monitoring and technical assistance. 19. Financial Management Systems. The record among PRSC countries indicates general progress in their financial management reform agendas leading up to the PRSC. In most cases, these reforms were supported by Bank lending, analytic work, and technical assistance. In Transparency International's latest Corruption Perception Index, 94 percent of the PRSC countries surveyed were ranked in the top three quintiles, with close to 70 percent in the top two quintiles, relative to other IDA countries.I5 Fiduciary assessments have usually determined that while fiduciary risks exist, countries are continuing to make progress. The latest preliminary IMF/World Bank HIPC assessment shows most of the participating PRSC countries with above average performance: 10 having met at least seven of 15 PEM benchmarks, compared to an l1 Strategic Frameworkfor IDA'SAssistance to Africa, World Bank, July 2003. l2 If there is no IMF program, Bank staff ascertain whether the IMF has major outstanding concerns about the adequacy of the country's macroeconomic policies. This practice reflects the division of labor between the two institutions: the Fund has led the dialogue with country authorities on macroeconomic issues and the Bank the dialogue on social and structural issues. The Fund's core areas are macroeconomic policies (monetary, fiscal, and exchange policies) and directly related structural reforms. The Bank's core areas are design of poverty reduction strategies (including poverty diagnostics, sectoral strategies, effectiveness of public expenditures, and social safety nets) and structural, social, and governance reforms. Overlapping areas are the environment for private sector growth, trade, financial sector, tax and customs policy administration, and issues related to public expenditure management, budget execution and monitoring, and fiscal transparency. See Strengthening IMF- WorldBank Collaboration on Country Programs and Conditionality (SecM2001-0461), August 23,2001. l3 In Vietnam, the third PRGF review was eventually cancelled because of an impasse between the government and the Fund over safeguard policies on audit and publication of the state bank's foreign reserves position. The Government considered measures expected under the IMF's safeguard rules incompatible with the country's legal framework. It was therefore agreed that the PRGF would expire without additional disbursements. 14 Sanjeev Gupta et. al., Is the PRGF living up to Expectations? An Assessment of Program Design, IMF, September 26,2002; and IMF Lending to Poor Countries-How does the PRGF differfrom the ESAF? IMF, April 2001. l 5 The survey ranked 146 countries in terms of the degree to which corruption is perceived to exist among public officials and politicians. There were 54 IDA countries surveyed and all but four PRSC countries were ranked. average of five benchmarks met by non-PRSC countries.16 In addition, 17 of the 20 PRSC countries demonstrated commitment to reform by undertaking DPL operations-most supporting reforms related to public expenditure and financial management-in the three years leading up to the PRSC program.17 Of the DPL operations that OED evaluated, 88percent of the policy-based operations were rated as having satisfactory or highly satisfactory outcomes; by comparison, in non-PRSC countries 70 percent of policy-based operations had satisfactory outcome ratings during the FYOI-05 period (Box 2 highlights lessons from DPL implementation that influenced PRSC design). Box 2. Lessons Learned from Previous Policy-Based Lending Successful implementation of policy-based lending is probably one of the best indicators of future success with PRSCs. Lessons learned have typically influenced the design of the PRSC program. Armenia: The Bank supported a decade of structural reforms through six policy-based credits prior to the PRSC, all but one of which had satisfactory outcomes, according to OED. Based on lessons learned, the PRSC program seeks to (1) strengthen the transmission channels of growth to poverty reduction, underpinned by well-sequenced ESW; (2) leverage trust funds and other donor support, notably PHRD grants and Millennium Challenge Account (MCA); (3) focus strongly on improving the business climate and public sector management as part of a strategy to eventually transition to IBRD lending. Ethiopia: Two policy-based credits preceding the PRSC helped improve pro-poor public spending and launched the public and private sector reform agendas. Both had satisfactory outcome ratings and influenced the design of the PRSC based on lessons learned, including: (1) technical assistance (such as the PSCAP that supports decentralized service delivery) should accompany policy-based lending to ensure quality implementation; (2) a medium-term framework which takes into account the varying pace of reform in different policy areas can balance the need to reward good performance, prod better progress, and meet urgent budgetary financing requirements: (3) in more challenging policy areas such as private sector development, the Bank should complement policy-based lending with a wide range of analytical work, persistent policy dialogue, and strong monitoring and supervision of reforms. Pakistan: Since FYO1, the Bank approved six policy-based credits, several of which were subnational operations. Of the four that have been rated, all had satisfactory outcomes. Lessons learned includes: (1) given the decentralized nature of service delivery, health and education reforms were more appropriately dealt with through subnational policy-based operations; (2) successive single tranche operations kept pace with the government's medium-term reform implementation; as reforms slowed during the political transition, so did Bank financial support; (3) given widespread poverty and scarce administrative resources, rapid, broad-based growth is more likely to be effective in reducing poverty than targeted programs. Benin: The Public Expenditure Reform Adjustment Credit (PERAC) supported a comprehensive public expenditure management reform in 2001 and extended with a supplemental credit. Satisfactory outcome is reflected in the fact that the country met eight of 15 HIPC benchmarks at the close of the operation. Key lessons upon which the PRSC builds include: (1) the momentum and framework created by the long-term vision of the PEM reform facilitate learning and capacity building as part of regular activities rather than separate from these; (2) it is important to clarify the role of the state in a given sector and the mandate of the line ministry before embarking on program-budget design, e.g., the organization of the rural water agency was well aligned with the rural water and sanitation program and subsequently performed well, while the breakup of the education sector into three separate ministries generated a host of problems; (3) good performance in ministries and agencies need to be rewarded in the form of increased budget allocation as an incentive. 20. Structural and Social Reforms. PRSC countries tend to receive relatively high quintile rankings on the Bank's Country Policy and Institutional Assessment (CPIA),18 as Figure 5 16 Update on the Assessments und Implementation of Action Plans to Strengthen Capacity of HIPCs to Track Poverty-reducing Public Spending, IMFIWorld Bank, February 2004, draft. 17 The countries that did not implement DPL operations within three years before PRSC approval were Guyana, Nepal, and Sri Lanka. I X The CPIA is an aggregate of performance ratings on public sector management, economic management, and social and economic policies. The CPIA ratings for the year prior to the calendar year of PRSC approval were used (e.g. for projects approved in CY 2004, scores for CPIA 2003) since the ratings are usually finalized by end of the year and become the operational ratings for the next calendar year. shows. All countries for which PRSCs have been approved to date-except Lao PDR-fall into the top three quintiles of this rating, with most falling in the top two (see Table 3). All 10 countries for which new PRSCs are in the pipeline for the next two fiscal years are also ranked in the top three quintiles, with seven of them ranked in the top two. Looked at from a different angle, all but four of the 23 PRSP countries in the top two CPIA quintiles are either currently implementing PRSCs or have a new PRSC in the pipeline. By comparison, only five of the 21 countries in the lowest three CPIA quintiles are or will be implementing PRSCs. While neither CPIA rating nor or any other single indicator determines eligibility, there is an expectation that PRSC countries would continue to represent the high end of the CPIA spectrum. 21. Comparison with Other Donors. The Bank's selectivity is strongly in line with that of other major donors. The U.S. Millennium Challenge Account (MCA), for example, is expected to provide significant resources to support economic growth and poverty reduction in a number of low-income countries that have demonstrated commitment to governance, investment in human capital, and market-oriented growth.19 To the maximum extent possible, country selection is based on objective and quantifiable indicators of a country's commitment to these principles. Of the 16 countries eligible to apply for MCA assistance in 2004, 10 were PRSC ~ountries.~'The European Commission and the UK's Department for International Development (DFID) provide a significant amount of their development assistance in the form of budget support, for which certain eligibility criteria are also applied. For DFID, eligibility is based on the government's planned budget priorities to support poverty reduction; its commitment to reliable administrative, technical, and financial systems; and the value added of budget support vis-a-vis other financing instruments (e.g., through reduced transaction costs).21DFID currently provides budget support to 20 countries, of which eight are also PRSC countries.22To be eligible for EC budget support, a country must have a macroeconomic reform program in place, a proven commitment to reform and improve public financial management, and a PRSP or an equivalent national development strategy under implementation; and it must use performance indicators to assess progress.23 Between 2001 and 2004, the EC provided budget support to 19 countries, of which 10are currently implementing PRSC-supported programs.*" l9 More information on the MCA is available at www.mca.gov. 20 The 10 PRSC countries are Armenia, Benin, Cape Verde, Ghana, Honduras, Madagascar, Mozambique, Nicaragua, Senegal, and Sri Lanka. Madagascar and Honduras became the first countries to sign multi-year compacts. 2' Poverty Reduction Budget Support, DFID Policy Paper, May 2004. 22 DFID provides budget support to the following countries: Kenya, Malawi, Sierra Leone, St. Helena, Zambia, Afghanistan, East Timor, India, Bolivia, Montserrat, Macedonia, Serbia, Ethiopia, Ghana, Mozambique, Rwanda, Tanzania, Uganda, Vietnam, and Pakistan. The latter eight are also PRSC countries. 23 Budget Support for Poverty Reduction in ACP Countries, Note to the Members of the EDF Committee, European Commission, March 2004. 24 The 19 countries are Benin, Burkina Faso, Cameroon, Chad, Ethiopia, Ghana, Lesotho, Madagascar, Malawi, Mali, Mauritania, Mozambique, Niger, Rwanda, Sao Tome, Senegal, Tanzania, Uganda, and Zambia. Figure 5. Average CPIA Ratings Table 3. CPIA Quintiles and PRSC Countries CPIA NO.of Countries Countries W PRSC L'ountnes BAII IDA Borrower: quintile countries with PRSPs with PRSC* 1 17 11 9 2 14 12 7 3 15 13 4 4 15 7 5 15 1 1 TOTAL 76 44 21 Source: SAP Business Warehouse, 2003 Country Policy 200 1 Institutional Assessments Note: Includes approved PRSCs as of April 2005. Sozirce: World Bank, CPIAs, 2001-2004 22. Extending Country Coverage of the PRSC. The challenge for the future is to ensure that PRSCs will continue to reflect selectivity based on adequate country performance. While the recent Africa Regional strategy explicitly states that PRSCs are intended for good performers, other Regions have not set out such guidance. This leaves IDA the option of flexibly using the PRSC to support potential "turnaround countries such as Georgia, or even countries such as Lao PDR, where the track record is more uneven and the policy environment and consensus for reform are still evolving, but where there is potential to leverage future progress. Where the country context poses significant risk of program disruption, investment projects and sectonvide approaches are likely to provide more appropriate implementation arrangements for establishing a track record of reform and a growing degree of confidence to move to a PRSC. For some PRSP countries, notably Bangladesh, Bolivia, and Kenya, bridging policy-based operations are preparing the necessary groundwork and track record particularly in public financial management, for a PRSC program. For non-PRSP countries, which have severe capacity constraints, Bank support in the form of advice and capacity building is typically more effective than policy-based lending in supporting reforms in their early stages.25 3. Links to CAS 23. In the Bank's country business the PRSP formally underpins the Bank's Country Assistance Strategy (CAS), which determines how the Bank expects to use its lending and nonlending instruments to support the priorities set out in the PRSP. All PRSCs are closely linked to the CAS, but in different ways that reflect the country's specific circumstances and needs. 24. Sequencing. For the PRSC, the CAS sets out the Bank's comparative advantage in the country, the scope and size of the PRSC program, the mix and complementarity with other operations and instruments, and the timing and sequencing of each operation (see Box 3). Thus, in the steady-state scenario, the Board would first discuss the PRSP and Joint Staff Advisory Note (JSAN),27followed by the CAS, and then the PRSC (see Figure 6). However, this ideal sequencing has not always been realized-for example, in Ghana, Pakistan, and Vietnam, where 25 Good Practice Note Development Policy Operations and Program Conditionali& in Fragile States, OPCS, World Bank, June 2005. 26 Supporting Cozintr? Development: World Bank Role and Instruments in Low- and Middle-Income Countries, World Bank, (DCl2000-19), September 2000. 27 The Joint Staff Assessment (JSA), which reflected Bank and Fund staffs summary assessment of the PRSP's adequacy. has been replaced by the Joint Staff Advisory Note (JSAN), which offers advice on how the government could strengthen the strategy and its implementation. sufficient information was available on which to base the PRSC, or where coordination with the government and other donors demanded a different timing, the PRSC was approved before the completion of the CAS or final PRSP. Box 3. Linking PRSCs to Results-Based CASs The CAS is the Bank's central strategic tool for aligning its support to country's development and poverty reduction goals, as set out in the PRSP. The new Results-based CAS is guided by the four principles of the Bank's Country Business Model-vision, diagnosis, programming, and results. It takes the country's development vision as set out in the PRSP as the starting point for identifying country level development and poverty reduction goals that the Bank is asked to contribute to, and has the comparative advantage and resources for doing so (strategic selectivity). Drawing on the available analytical work, the CAS identifies the key development challenges that need to be addressed for achieving country's goals. This diagnosis work sets the framework for designing CAS outcomes that the Bank's is likely to influence through its lending and non- lending activities. Finally, the CAS identifies the most appropriate mix of instruments, including the PRSC, for delivering these outcomes, taking into account programs supported by other development partners of the country. This results framework establishes clearly defined causal chains between the PRSP goals and priorities and the Bank's interventions, and, hence, sets the development policy agenda for the PRSC. Ideally, for a country with a results-based CAS, the PRSC will essentially be an operationalization of CAS outcomes through defining policy actions, institutional changes, and resource allocation decisions to be supported by the Bank. However, the results-based CAS approach is fairly new both for the Bank, and, more importantly, for the client countries. It has been piloted during the last two years, and is being mainstreamed in FY05. The Bank teams would need to gain further experience and acquire skills, and transfer the know-how to the client countries, before it becomes an effective tool for hlly aligning PRSC, and other Bank interventions, with the PRSP priorities. - Source: Results Focus in Country Assistance Strategies: A Stocktaking of Results-Based CASs, World Bank, February 2005. 25. CAS Lending Case. In outlining the contours of the PRSC, the CAS suggests a notional medium-term commitment for the PRSC program, which it links to the government's medium- term expenditure framework (the commitment amount is subject to IDA'S annual performance- related allocation). In most cases, the CAS pegs the PRSC to the base-case lending ~cenario,~' reflecting the central role that PRSCs play in supporting PRSP implementation and anchoring the Bank's policy dialogue in the country (a departure from the past practice of undertaking policy- based lending only when country performance exceeds expectation). However, in several countries, including Tanzania and Madagascar, the PRSC was part of the high case; and in Armenia and Nicaragua, the move to the high case triggered additional IDA resources for the PRSC program. 28 With the full alignment of CAS envelopes with the performance-based IDA allocation system under IDA-14, CAS scenarios are now used for the implementation of the strategy rather than the modulation of financial support. CASs typically contain two or three lending scenarios. The "base-case" scenario sets out the most likely levels of Bank lending and nonlending activities over the CAS period, as determined by the country's expected macro/fiscal and policy performance. In addition, depending on how country circumstances are expected to change over the CAS period, the CAS may contain a "high-case" and a "low-case" scenario. Scenarios, and the triggers for moving between them, are tailored to country circumstances and the nature of the Bank's relationship with the country. Figure 6. Country Business Model for Low-Income Countries 26. The Role of the PRSC. The CAS also determines the scope of the PRSC, and specifically the role it plays in the country portfolio. In most cases, the PRSC anchors the Bank's policy dialogue on a broad range of cross-cutting and sectoral areas. Some PRSCs go further, representing the Bank's main financing instrument; in these countries, the Bank's lending portfolio is shifting away from multiple stand-alone projects and toward a consolidated, comprehensive, and programmatic approach that is aligned with government priorities and responds to the government's call for better streamlining and rationalization, and greater control over donor lending modalities (see Box 4). The different roles that PRSCs play in the country portfolio are reflected in the varying weight of PRSC commitments relative to total Bank lending commitments to the country (see Figure 7). Box 4. Beyond Projects The current development literature has recognized budget support as a response to the disadvantages inherent in a traditional donor-driven project-based approach: high transaction costs, unpredictable resource flows, and lack of integration with government systems and budget processes, and hence uncertain chances for sustainabi~it~.~~ In several countries, the PRSC program has been welcomed as a new approach to respond to these limitations. The Benin PRSC program document notes that the transition toward a consolidated programmatic approach through the PRSC responds to "limited effectiveness of project lending...while supporting greater government ownership in formulating and implementing development programs." The program document for the Mozambique PRSC similarly notes that the "shift in lending modalities is needed because investment projects have been characterized by variable local ownership and poor integration with government-funded activities, with the result that impact and sustainability has been low." However, projects still play an important role in development. As the Strategic Frameworkfor Investment in Africa (SFIA) notes, "IDA remains committed to traditional interventions where they are most appropriate." Even the Benin PRSC program document notes the use of traditional self-standing operations would continue "where the benefits of separate implementation and financial arrangements appear to outweigh the advantages of a more integrated approach to public financial resource transfers." 27. The Uganda Model. The Uganda PRSC represents about 45 percent of IDA commitments to the country since FYO1. It was designed to give the government more control over the reform process and reduce the high transaction costs of managing a highly fragmented portfolio of donor projects. It folds into one operation the sectoral investment programs that had helped finance the government's core service delivery in the education, health, water and sanitation, and rural sectors. Subsequent operations deepened the policy dialogue in these sectors while adding specific policy issues, such as rural finance and pensions. The investment 29 A. Larson, and D. Booth, EvaluationFrameworkfor GeneralBudget Support, ODI, 2004, pg. 25. portfolio is still robust, with interventions continuing where separate implementation and financial arrangements make more sense, as in capacity-building and large infrastructure projects with international bidding and procurement. 28. PRSCs in Consolidated Portfolios. Other African countries also use PRSCs to consolidate their portfolios, without going as far as Uganda. The Benin PRSC has a consolidated approach similar to Uganda's, covering education, health, and rural water, with the intention of folding in other sectors in subsequent operations. In Ethiopia, education and health financing were folded into the PRSC after the sectorwide programs ended; in Burkina Faso and Cape Verde, only the health sector is exclusively financed by the PRSC. In Mozambique, the PRSC program has primarily focused on cross-cutting issues, with the expectation that subsequent operations will incorporate the policy dialogue and financing of sectoral investments once the current SWAps and sectoral DPLs come to a close. 29. Small Cross-Cutting Focus. For other countries, the PRSC represents a much smaller share of the portfolio. The Albania PRSC program, for example, represents about 20 percent of total lending, focusing on cross-cutting issues in public sector, policy development and performance monitoring, public expenditure management, and improved service delivery in social sectors.'O Through these cross-cutting reforms, the program provides the umbrella for complementary support through investment operations. For instance, the PRSC refers to policy reform in the power and financial sectors, but the specific investments are supported by individual projects. The decision to maintain a large investment portfolio also reflects donors' coordinated effort to avoid duplication, with the Bank filling a void left by the EC's shift away from infrastructure projects. As Albania prepares its second PRSC program, there is discussion, triggered by an OED recommendation, about reshaping the lending portfolio to fewer but larger and more strategic programs, with a more prominent role and funding for PRSCs that could leverage greater structural and institutional reforms to increase the sustainability of economic growth and poverty red~ction.~' 30 Of the US$189 million in Bank commitments approved from FY02-04, US$38 million was for PRSCs. 3 1Albania CountryAssistance Evaluation, OED, World Bank, January 2005 30. Sectoral Policy Dialogue. In other Figure 7. PRSC Share of New IDA Commitments countries, PRSCs haie focused on cross- 1- Cape Verde cutting reforms but have also brought in I Armenia 50% specific measures to complement the sectoral Madagucar 49% financing of investment programs. In Rwanda 48% Vietnam, where the PRSC represents about 20 Nicaragua 48% percent of total IDA lending and focuses on Sri Lanka 47% the private and financial sectors, SOE reform, Burkia Faso 47% public expenditure management, and sectoral u~ d a 45% policy and institutional reforms aimed at Benin 42% accelerating PRSP implementation, the Ghana 38% government preferred to maintain SWAps in i Senegl 37% sectors ranging from education to rural Nepal 37% transport. In Ghana, the PRSC program has Pakistan 36% identified targets and focused on budget Mozambique 35% execution to fund sectoral reforms that Ethiopia 35% complement new and ongoing SWAps. In Sri Tanzania 27% Lanka, the PRSC program consists of four Vietnam 20% operations, each expanding the program's Alban~a 20% coverage with such "themes" as rigidities in Honduras 19% labor and land markets. reforms of the welfare 0% 10% 20% system, and reinvigorating the rural 30% 40% 50% 60% 70?6 economy.32 In Senegal, the Bank preferred to -- -- Source: SAP Business Warehouse consolidate progress on PEM before folding Note: PRSC shares in Uganda, Burkina Faso, Albania, in sectoral financing under the PRSC. and Vietnam are based on actual approvals. Those for other countries are based on CAS base case lending 4. Analytic Underpinnings scenarios. 3 1. In the Bank's country program model, a cycle of analytic and diagnostic work informs and underpins the policy choices of the PRSP, the Bank's CAS, and the Bank's lending program. While the PRSP provides the overall framework and policy direction for Bank assistance, the analytic work helps inform the policy dialogue that guides the design of the PRSC and establishes clear actions to achieve the government's objectives. This analytic work may be undertaken by the Bank, the borrower, or development partners, either independently or in collaboration with one another. 32. Core Analytic Work From 2000 to 2004, to provide a minimum analytic base, Bank management committed to have in place core economic and sector work (ESW) for every active country. Now that this objective has been largely achieved, the CAS determines the core analytic work. It typically assesses public financial management systems, specifically fiscal transparency and financial accountability, and key social, structural, and sectoral policies and institutions aimed at stimulating investment, creating employment, and accelerating and sustaining growth. OP 8.60 requires all DPLs-including PRSCs-to draw on relevant analytic work on the country. The program document describes the main pieces of analytic work on the country and indicates how gaps will be addressed. These specific products include the Country Financial Accountability Assessment (CFAA), Public Expenditure Review (PER), Country 32 The approach laid out in the Sri Lanka PRSC 1 has been disrupted following the change in government. The PRSC program may change under a revised PRSP. Procurement Assessment Report (CPAR), Poverty Assessment (PA), and Country Economic Memorandum (CEM) or Development Policy Review (DPR). PRSC measures are drawn from or shapedby the recommendationsprovided in the diagnostic work (see Box 5). Box 5. Analytic Underpinningsof Honduras PRSC The reform agenda supported by the PRSC is drawn from the government's PRSP and buttressed by a wide range of formal and informal analytical sources. The CFAA and CPAR werejoint undertakings by the Bank and the Inter-American Development Bank (IDB) that strongly influenced the PRSC, pointing to measures that could improve the country's public financial management systems: legislation to strengthen the legal and accountability framework, and capacity building of internal procurement and external auditing agencies. Moreover, the DPR and an Investment Climate Assessment (ICA), undertaken in parallel with the PRSC and in close collaboration with the government, made recommendations that were taken into account in the design of the PRSC. The DPR, for example, recommended specific measures to support the government's objectives of improving education achievements and expanding public infrastructure, while the ICA recommended measures to decrease red tape and corruption. Several sector-specific studies also underpinned the actions in the PRSC: assessments of infrastructure, nutrition, education, rural growth, and trade, all of which helped the government and Bank reach agreement on PRSC priorities and benchmarks. A set of studies on the implications of the Central American Free Trade Agreement (CAFTA) for the region and their policy recommendations concerning ways to optimize the Agreement's potential economic and social benefits influenced the design of the competitiveness actions included in the PRC policy matrix. Lastly, a PSIA on tax reform found that tax measures undertaken by the government within the PRSP framework have been progressive overall. Thus, improvements in tax administration to reduce tax evasion and increase revenue collection will be key to materialize the distributive gains of the reform. I I Source: Program Document for Honduras PRSC 33. Analpic Coverage in PRSC Countries. The PRSC and other donor budget support programs have helped governments address analytic gaps. PRSC countries are more broadly covered by key ESW products than non-PRSC countries (Figure 8). Where gaps exist in Bank ESW, other similar analytic and diagnostic work by the government or by developmentpartners has substituted. For example, in lieu of Country Economic Memoranda in Burkina Faso, the Bank and government carried out a Growth and Competitiveness Study to bolster the growth strategy supported by the PRSC; and in Burkina Faso and Rwanda, the Bank and development partners used the governments' assessments of their financial management systems as the basis for addressing key fiduciary and financial management issues.33 In Sri Lanka, the Bank and Government are jointly undertaking a program of annual PER work, focusing on different themes each year, with the results feeding into the PRSP process and the PRSCs. 34. Sectoral Analytic Work. Compared to other IDA countries, PRSC countries have also been subject to a higher average number of total ESW, apart from the core products already mentioned.34These include reports that address sector-specificor thematic issues, such as health and education sector reviews, and gender and environmental analyses. Like the core ESW products, these reports have provided an analytic basis for the PRSP, directly feeding into the design, content, and coverage of PRSC and other lending operations. 33 The government assessment in Burkina Faso built directly a multidonor CFAA that was completed shortly before the second PRSC operation. 34 There are four types of ESW: Core Diagnostic ESW; Other Diagnostic ESW, such as sector and theme specific assessments and reviews; Country Advisory Reports, consisting of more specialized products relevant to specific clients and specific topics; and other Non-Formal or Non-Assigned ESW, such as policy notes, workshops, and conferences. 35. PSIA. With Poverty and Social Impact Analysis (PSIA)," Bank teams analyze the distributional impact of policy reforms on the welfare of different stakeholder groups, focusing particularly on poor and vulnerable people. Nearly 60 percent of the PSIAs completed for IDA countries have targeted countries currently undertaking or preparing PRSC programs,36and nearly half of the current PRSC programs have benefited from PSIA-type analyses (see Box 6). Formalizing the link between policies and analytical tools on their impact is contributing to a - deeper understanding of the impact of PRSC-supported policy reforms. Box 6. Poverty and Social Impact Analyses in Madagascar To inform the content of the PRSC, ex-ante PSIA was undertaken during the preparation of the PRSC. The analysis included: The distributional impact of changing rice tar@, agricultural input subsidies and reducing transactions costs. Based on the results of the studies, rice tariffs were maintained at 15 percent so as to ensure a producer floor price that would enable most farmers to recuperate production costs, while ensuring a reasonable retail price for urban consumers. Leakage in the education sector through apilot expenditure tracking survey. The study found that there are two primary types of leakages: direct resource leakages that occur when equipment and funds that are intended for schools do not arrive; and leakages through overpricing, when school supplies are purchased at a higher price than what is available in the market. Analysis of the accountability mechanisms and institutional constraints at the district level in education and health. The study found that remote areas receive less supervision, fewer resources, and rely to a greater degree on community participation, even to pay staff with particularly acute gaps in areas that were as little as 20km from district capitals with increasing gaps as distance from the capitals increased. 1In addition, two primary areas where PSIA are planned: I Service quality at the district and facility level in the health sector and analyzing the impact of cost recovery in the health sector and evaluating alternate options to mitigate it. Impact of rural roads, which is critical in the Madagascar PRSP, and the evaluation will aim to quantify the benefits of roads, and assess the distributional impact on different groups of households. Source: Program Document for Madagascar PRSC 5. Environmental Aspects 36. Before the update of Operational Directive (OD) 8.60, Structural Adjustment, all PRSCs were subject to review by a Bankwide working group to determine the applicability of safeguard policies. The 21 PRSCs that were reviewed were evenly split between structural adjustment credits (SACs) and sectoral adjustment credits (SECACs). For SACs, OD 8.60 recommended as good practice that environmental policies, practices, and institutions of the country be reviewed as an integral part of loan preparation; and SECACs supported policy and institutional reform measures whose potential environmental effects needed to be addressed and documented as part of the program document. Hence, the nature of the operation to some extent determined the level of attention to environmental issues. Since the introduction of OP 8.60, all PRSCs have been subject to the environmental provisions of the new operational policy, requiring the determination of any significant effects of specific actions supported by the PRSC program on the country's environment, forestry, or other natural resources. 3' PSIA is a systematic analytic approach, not a separate ESW product. See www.worldbank.or~/~siafor further '' information. The World Bank is enhancing its efforts to mainstream the use of PSIA techniques by providing supplemental budget resources to teams undertaking PSIA analysis. The data cited refer only to activities receiving supplemental budget resources. They include Albania, Armenia, Benin, Burkina Faso, Ethiopia, Ghana, Guyana, Honduras, Madagascar, Mozambique, Nicaragua, Rwanda, Uganda, and Vietnam. PREM is currently undertaking a compilation of analytic work that could be classified as PSIA. 6. Cost Considerations 37. Programmatic approaches such as SWAps and PRSCs were expected to be cost-effective vehicles for supporting the policy dialogue on the social and structural agenda.j7 By consolidating the project portfolio and moving away from self-standing projects, the PRSC would in theory benefit from economies of scale that would cut the costs associated with stand- alone project preparation. Since only a handful of countries have completed their first series of PRSCs, PRSC cost assessments must be seen as highly preliminary; however, the trends indicate that PRSCs are generally cost-effective. 38. Preparation and Supervision Costs. On average, PRSCs are more expensive to prepare than other DPLs and investment projects: an estimated $515,000 during the FY01-05 period, compared to $350,000 for other DPLs and $360,000 for specific investment loans and adaptable program loans (SILIAPLs). Three of the four countries that are most advanced in their PRSC programs have shown a general decline in preparation costs over time, reflecting the expectation that costs should lower as the program matures (see Figure 8). Uganda's program costs, however, have continued to rise, perhaps because of the comprehensive nature, progressively broader coverage, and larger number of donors that characterize the Uganda program compared to the others. As single-tranche instruments, PRSCs incur very low supervision costs since supervision coincides with the preparation of the next operation: average supervision costs for all PRSCs are about $126,000, compared to $700,000 for SILsIAPLs. Among the four advanced PRSC countries, the first PRSCs for Albania and Vietnam stand out as having incurred substantial supervision costs, and only Burkina Faso has PRSC completion costs consistently falling below those of SILs and APLs. Figure 8. Preparation and Supervision Costs ($thousands) I 1 -- 1 %pervirion 1,600 Preparation 1,482 1.359 UGANDA BURKINA FASO VlETNAiM ALBANlA I I Source: SAP Business Warehouse Note: Costs for SILJAPL are average costs of closed operations from FY95-05 39. Cost-Effectiveness. Relative to commitment volumes, PRSCs have generally been cost- effective compared to SILsIAPLs: every dollar spent to prepare and supervise a PRSC translates to about $143 in commitments, compared to $36 in commitments for SILsIAPLs (see Figure 9). Among the PRSC countries that have moved to subsequent operations, Tanzania has the highest cost-effectiveness ratio and Albania the lowest. The disbursement-to-cost ratio also follows a 37 Support Country Development: World Bank Role and Instruments in Low- and Middle-Income Countries, World Bank (DCJ2000-19), September, 2000. very similar trend, with PRSCs disbursing an average of $143 for every dollar spent preparing and supervising the PRSC, and APLsISILs an average of $32. Figure 9. Commitment per Cost Ratio 7. .PRSC f3SIUAPL Tanzania Calana Eth~opia qetnam Uganda Burkina Albania All PRSC Faso Countries Source: SAP Business Warehouse Note: Figures are based on average costs Conclusion 40. As discussed in this section, many of the early assumptions about PRSCs have been met: their share and number have increased; they are used mainly in better-performing countries; each PRSC is strongly linked to the country's CAS and based on adequate analytic underpinnings; and the instrument appears to be cost-effective. 4 1. Selectivity. PRSCs have been extended to a limited number of well-performing countries that are likely to make the most effective use of regular budget support. There is a broad consensus that the PRSC "label" should remain a selective approach that is reserved for good performers with a proven track record of program implementation. These programs would build on a demonstrated commitment to a nationally-owned strategy that draws on a robust set of analytic and diagnostic work. The Bank has the discretion to support genuine turn-around cases through a PRSC, a decision that should be made on a country by country basis, without relying on any single performance indicator. There are also opportunities in weaker performers to improve institutional environments and strengthen country performance with development policy lending operations that could be more narrowly focused and potentially less frequent than typical PRSCs." 111. PRSP ALIGNMENT:CONTENTAND COVERAGE 42. This section will assess to which extent the PRSC is meeting the first objective, to help implement a medium-term program that builds upon and draws from the priorities and objectives of the PRSP. PRSCs could meet this objective in two ways-by helping to operationalize the PRSP and by supporting the PRSP's comprehensive program. Good Practice Note on Development Policy Operations and Program Conditionality in Fragile States, OPCS, World Bank, June 2005. A. Operationalizing the PRSP 43. PRSPs are comprehensive documents that attempt to diagnose constraints to, and outline strategies for, sustainable growth and poverty alleviation. They provide the central framework for government and donor partnerships in low-income countries, but they often have major weaknesses that make implementation difficult: for example, they may not develop sector and thematic strategies sufficiently, they may lack an operational blueprint for implementation, or they may not be well linked to underlying analytic work, budget priorities, and broader objective^.^' Nearly 67 percent of JSAs/JSANs for PRSC countries note the need for greater elaboration of underlying strategies; 72 percent for costing and financing scenarios; and 85 percent for greater specificity, prioritization, and sequencing of time-bound actions. However, the PRSP process is dynamic and allows changes to improve the operational framework of the strategy. PRSCs have proven to be an integral part of this process, focusing the policy dialogue to address these shortcomings and, in effect, making PRSPs more hlly operational. I. Elaborating Strategies 44. Gathering sector programs and strategies under one strategic framework allows stakeholders to understand the multidimensional constraints to development. However, formulating these strategies is a challenging task for many capacity constrained countries. While some PRSPs build upon strong sectonvide approaches, particularly in education and health, strategies for other areas have been more uneven and require greater elab~ration.~~Nearly 70 percent of JSAs/JSANs in PRSC countries note the need to strengthen and elaborate on specific sector and thematic strategies. In helping to address this shortcoming, the Bank provides technical support and analytic work. The policy dialogue of the PRSC has also helped by underpinning the policy dialogue with specific benchmarks and measures in the policy matrix that help focus government attention. In many cases, the aim is to accelerate the sector's readiness for future program lending. In Ethiopia, policy dialogue leading up to the PRSC focused on making progress on the government's rural development and private sector development strategies. The PRSC also helped accelerate a sectonvide program in rural water in Madagascar, a public sector reform strategy in Ghana, a rural development strategy in Mozambique, a water and sanitation strategy in Nicaragua, and a health sector development plan in Burkina Faso. Moreover, a number of specific action plans have led to greater elaboration of sector strategies, such as for the power sector in Pakistan, and private ownership of land in Sri Lanka. More broadly speaking, Bank analytic and diagnostic work, particularly in public financial management, has underpinned the reform agendas in this area for all PRSC countries, providing specific recommendations to address key issues. '' Poverty Reduction Strategy Papers-Detailed Analysis of Progress in Implementation, World Bank and IMF, September 15,2003. 40 The Poverty Reduction Strategy Initiative, An Independent Evaluation of the World Bank's Support Through 2003, OED, World Bank, October 2004. 2. Policy Matrices as Performance Frameworks 45. PRSPs have been charged with multiple expectations that have resulted in tensions and challenges for governments and donors alike. On the one hand, PRSPs are expected to be comprehensive documents that provide an overarching framework for development. On the other hand, PRSPs are expected to be operationally relevant, with specific implementation plans and links to budgets. However, they often do not meet the second expectation-they generally do not provide the level of operational detail that specify how objectives are to be achieved through policy actions. Where an implementation plan or program matrix is included, it normally takes the form of a long list of actions across a wide range of areas, with the intended focus of the strategy remaining unclear. They often fail to prioritize, sequence, and establish a timeline of actions that are vital to achieving medium-term objectives. These problems often impede the operational usefulness of the PRSP to both national authorities and development partners." 46. Filling the Gap. In the absence of such specific implementation plans, the Bank and other donors have helped to make PRSPs more operational, and several approaches have emerged in this regard. The most common means is to develop an annualized medium-term operational plan with specific policy measures4' negotiated with government by donors, and that is based on, but independent from, the PRSP itself. In Uganda, Vietnam, and Nicaragua, the PRSC policy matrix acts as the implementation plan for several budget support donors. In Cape Verde, Ethiopia, Ghana, Mozambique, Tanzania, and many other African countries, the PRSC policy matrix is identical to or a subset of the common donor matrix (often referred to as a performance assessment framework, or PAF), which itself draws upon the PRSP and its objectives. The EU and several bilateral donors, while including policy measures in public financial management, generally focus on output and outcome indicators. This experience, however, has been similar to the Bank's. Key indicators and annual targets are typically negotiated with governments, given their absence in many PRSPs. While these approaches have the clear advantage of filling the PRSP's operational gaps and clarifying criteria for budget support disbursement, they have also prompted discussion over country ownership, conditionality, and mutual accountability. In several countries, differing donor approaches to budget support have also resulted in challenges to donor harmonization. 47. PRSC Policy Matrix. The PRSC policy matrix summarizes the programmatic alignment with the PRSP. It sets out key indicative reform measures in an annual, flexible medium-term framework that is sequenced with the PRSP and linked to measurable outcomes. While there is no formal requirement for policy matrices, they have emerged as a broad reference framework and a management tool for the government and its development partners. The design of the policy matrix provides stakeholders with a transparent understanding of the scope, logical sequence, and expected results of the government's measures, and to keep track of progress It presents the set of: 41 PoverQ Reduction Strategy Papers - Detailed Analvsis oj'Progress in Implementation, World Bank, September 2003, pg. 42. " Policy measures in the PRSC include all policy and institutional reforms laid out in the policy matrix. However, a distinction is made between conditions (on which disbursements are contingent) and non-binding measures or indicative benchmarks. Given the single-tranche, programmatic approach of the PRSC, conditionality essentially consists of "prior actions," which must be met prior to Board approval. For the purpose of this paper, conditionality or prior actions will be referred only as conditions. 43 G o d Prclctice Note on the Design ofDevelopment Policy Lending, OPCS, World Bank, May 2004. binding measures upon which disbursement decisions are made other measures, or indicative benchmarks, that are not binding but comprise key components of the policy agenda triggers, or expected actions, that determine whether adequate progress has been made to warrant continuation of the program outcome indicators over the entire PRSC period. The components of the program are typically organized by broad policy area, with measures presented in a way to demonstrate reform sequence and logic from year one to year three. However, the policy matrix is not designed to be static. It is an integral part of a step-by-step, flexible process that takes account of changes in the reform program. 48. Deriving Policy Measures. The natural starting point for designing the policy matrix is the PRSP itself. This entails hard choices on the part of the government, the Bank, and other donors in terms of prioritizing, refining, elaborating on, and adding to the policy actions contained in the PRSP, based on the policy dialogue and underlying analytic and diagnostic work (see Table 4). A recent survey conducted by the Strategic Partnership for Africa (SPA) indicates that while very few binding measures in budget support programs for Africa are drawn directly from the PRSP, a majority is drawn partly from the PRSP and partly from other government documents.44 The policy dialogue with the government is critical in gauging what is realistic in terms of timing, sequencing, and government commitment, given capacity constraints, lessons learned, political considerations, and available analytic work. Particularly in the area of public financial management, Bank ESW has provided recommendations in the form of specific policy and institutional measures that are typically picked up in the PRSC policy matrix. The ideal product should be a single policy framework that builds on the PRSP, is developed by government with donor support, and promotes domestic accountability. 49. Country Ownership. Typically, the policy and institutional measures in a PRSC program are drawn from, or elaborate on, a subset of the PRSP. They typically derive from a negotiated process with government, based on a shared understanding of the actions needed to achieve the country's development objectives. In countries where the PRSC is part of a harmonized multi- donor framework, negotiations on policy measures are typically done multilaterally. The shared understanding in most cases has translated to strong ownership by many governments. However, several governments and some donors view these measures as excessively intrusive, negotiated more in line with traditional conditionality, and not reflective of government views on the pace, timing, or priorities of the reform program. Twenty years of adjustment lending conditionality has shown that country ownership plays a critical role in sustainable reform implementati~n.~~ 44 Suwey of the Alignment of Bzldget Support and Balance of Payments Support with National PRS Processes, SPA, February 2005. 45 Adjustment Lending Retrospective, OPCS, World Bank, June 2001. Table 4. Drawing on Analytic Work to Design Reform Measures in Honduras Analytic Report Recommendations PRSC reform measures Improve education achievement through the The MOE has published in its wehsite and has expansion of secondary education coverage, submitted to the INE: (i) the educational statistics improved efficiency of primary education, database from the 2003 school census, including the restoration of control over teacher the analysis of teacher and student distributions, remuneration and performance, and the and (iij the teacher payroll by department as it reduction of the share of sector expenditures appears in SIARHD. Development Policy devoted to headquarters. The government has increased access and Expand infrastructure services Broad Based efficiency in telecomn~unicationsservices through through improved regulatory and policy the following actions: (i) HONDUTEL has framework, elimination of distorting price awarded 27 commercial contracts and four controls and subsidies, greater private interconnection contracts for private operators to participation and competition, and provide market-priced, national and local call improved operational and institutional services; and (ii) CONATEL has awarded a license efficiency to a second cellular telephone operator Satisfactory implementation of the Administration Decrease red tape, corruption and crime simplification ~ a wevidenced by (i) the reduction , through the introduction of measures to in the time needed to register a typical business in streamline business registration and Tegucigalpa from 129 days in 2002 to 62 days in lnvestment Climate operating procedures. improve government the first quarter of 2004, and (ii) the creation of the Assessment (Draft financial management systems, reducing Administrative Simplification Committee (CSAE), Report) corruption in the civil sewice, and through an executive decree and a resolution promoting decentralization while adopted by the National Competitiveness maintaining accountability. Commission, to coordinate all initiatives for administrative simplification. There are complex determinants of The Executive has redefined poverty reduction heing poor increases with the number of expenditures in its first annual PRSP Progress babies and children, heing from an Report to improve linkages between poverty indigenous population and having an Poverty Diagnostic spending and PRS targets. unemployed or female head of household. Report Progress can also he in non-monetary Core funding has been provided in 2004 for the indicators: Poverty can be reduced through water and sanitation regulatory agency to pay for access to basic public infrastructure sewices key staff salaries and minimum operating costs. such as water and sanitation. Source: Broadly based on Box B of the Honduras PRSC Program Document 50. Managing Tensions. In light of this finding, the Bank has moved away from ex ante conditionality, to support reform measures that the government is committed to.4h In PRSCs specifically, the Bank envisioned that focusing on actions already implemented, or ex post conditionality, would provide a pragmatic approach to balance, and ultimately to reconcile, the tension between country ownership and donors' need for due diligence and results. In practice, the distinction between ex-ante and ex-post conditions may not always be clear, especially for subsequent PRSC operations where future policy measures are identified as 'triggers' for future support. Bank teams acknowledge that the tension between country ownership on the one hand and due diligence and results on the other, is an inherent aspect to any policy dialogue. The challenge for the Bank is to build an honest and transparent relationship that keeps this tension in check; that builds a robust reform program that is technically sound, consistent with PRSP objectives, and that pushes the envelope in terms of what the government could commit to, while reinforcing accountability toward domestic stakeholders. For the Bank and other donors, lessons learned include the need to realize the distinction between policy reforms that are feasible early on and those that are likely to be slow-moving. A balance needs to be struck between signaling 46 Stefan Koeberle, "Should Policy-Based Lending Still lnvolve Conditionality?" World Bank Research Observer, 2003: and Programmatic Aajustmei~tLending Retrospective, OPCS, World Bank, January 2004. the need for better progress and rewarding continued government commitment and the implementation of good policies. For example, in Ethiopia, progress on private sector development is expected to evolve slowly, given the tradition of mistrust between the private sector and the government, and the role of the Bank should be to support a more open dialogue between the government and the private sector with a strong and flexible partnership, wide range of analytical work, continued in-depth policy dialogue, and strong monitoring of reforms.47 5 1. Experience to Date. Experience to date has generally been positive. The more advanced PRSC programs provide evidence that strong country ownership can derive from a negotiated process. In the PRSC stocktaking in Uganda and Burkina Faso, the governments clearly claimed the reform measures in the PRSC program as reflective of their own priorities. In a budget support workshop, an official from the Ugandan Treasury noted that the term "conditionality" should not be used, in light of the government's commitment to the PRSC's reform measure^.^' Several other African government representatives expressed similar views in other f ~ r a .In~ ~ Vietnam, the government has expressed appreciation for the need to underpin the policy dialogue with a medium-term framework of specific policy actions. Newer PRSCs also provide good practice examples. In Pakistan, the preparation of the PRSC policy matrix encouraged a dialogue among various parts of the government and with nongovernmental stakeholders that helped define priority actions and expected outcomes. 52. Government-LedProcess. Nevertheless, donor-negotiated processes and instruments are always a second best option in operationalizing the PRSP. In Ethiopia and Mozambique, the governments have taken a more assertive role in leading the process of deriving priority measures from their PRSP policy matrices. The role of the Bank and other donors can be characterized more in the spirit of consultation, rather than a donor-led negotiated process. In both cases, the result is a much more prioritized set of actions directly linked to the PRSP, which all budget support donors use as a joint framework to draw disbursement triggers and to assess performance. In Mozambique, this process entailed the establishment of 19 working groups to develop specialized matrices in five key areas of the PRSP. Representatives of the government and a lead donor jointly chaired each working group. In Ethiopia, the government created four inter-ministerial working groups to develop matrices for each of the four main themes contained in the PRSP (see Box 7). While these processes have entailed a substantial amount of time and resources, they are an integral part of the PRSP, a fact that helps to reinforce strong domestic accountability. 53. Lessons Learned. These examples of country-driven processes for defining conditionality show the path ahead for other budget-support arrangements, although they are not perfect. In both Ethiopia and Mozambique, these initiatives are still evolving and experience suggests a need to further clarify the role of the policy matrix and the process for deriving them. Certain line ministries apparently wanted to be more closely consulted on specific measures in the government matrix. Identifying reform measures may also benefit from greater consultation and participation by Parliament and other key stakeholders, such as the private sector, in the case of Ethiopia. 47 E~hiopiaPRSC 2 Program Document (29772-ET), World Bank, November 2004. 4g Strategic Partnership with Africa, Budge2 Support Workshop, Kigali, Rwanda, June 2004. 49 Review of World Bank Conditionality:Summary of Consultations,June 2005. Box 7. Ethiopia's Country-DrivenPolicy Matrix I The Ethiopian government has viewed the PRSP as an important opportunity to articulate and provide leadership to its longstanding development program to both internal and external constituents. Since the PRSP's completion, the government undertook a considerable effort to refine the PRSP policy matrix and to design one M&E system for the government to track progress on the PRSP and for donors to clarify the link between input, output, and outcome indicators. In order to reduce the actions in the policy matrix to a smaller and more manageable level, the government established four inter-ministerial working groups to develop matrices for each of the four main themes contained in the PRSP: growth (private sector development, rural development, vulnerability and roads); human development outcomes (education, health, and water); public sector institutional outcomes; and democratic governance 1decentralization. The quality and pace of inputs from these groups varied; considerable time was needed, for example, to develop concrete action plans to underpin several sectoral strategies. And in some key cross-cutting areas, such as private sector development and rural development, the government required considerable time to develop the strategies themselves. By the PRSC appraisal mission, the PRSP matrix contained an unwieldy 449 indicators (40 outcomes, 289 outputs, and 120 actions). Through discussions with the Bank and other donors, the government was able to identify a subset of indicators for the PRSC and budget support purposes, reducing the policy matrix to 127 indicators (31 outcomes, 58 outputs, and 38 inputs or reform measures). The major criteria used in selecting the subset of budget support indicators were: annual measurability, attribution to government actions, and reflection of broader reform progress. The results are a much more manageable policy matrix that identifies key priority actions that are linked to monitorable output and outcome indicators and that reflect strong government commitment and ownership. 54. Future Direction. The second generation of PRSPs provides ample opportunities for governments to build upon lessons learned and to take into account emerging good practice. Donors should recognize the importance of the PRSP as a tool for domestic accountability; governments, too, should reinforce this by feeding back into the PRSP-and therefore to domestic constituents, including parliament-the reform measures agreed to with donors. In Nicaragua, the second PRSP will be an amalgamation of the first PRSP and the PRSC policy matrix, and will draw on progress reports, local development plans developed by regional and departmental development councils, and sector plans developed by new sector working Similarly, in Burkina Faso and Tanzania, the second PRSP is expected to provide both the overarching framework and specific implementation plan, building on the processes and instruments from the first PRSP, and replacing the negotiated policy matrices (the PAF, in the case of Tanzania) with matrices derived by the government from the PRSP. A similar process is underway in Albania to bolster country ownership of the reform measures supported by the second generation of PRSCs. 50 PRSP Annual Progress Reports and Joint StaffAssessments - A Review of Progress, ODI, September 2004. Box 8. Trends in PRSC Policy Matrices Good practice suggests that the policy matrix reflect key policy and institutional actions supported by the operation. In recent years, there has been an effort to prioritize and limit the triggers and prior actions (or binding measures) of PRSCs. Indeed, the average number of prior actions for PRSCs has declined substantially from 30 in FYOl to about 10 so far in FY05. The number of prior actions varies from operation to operation, ranging from a high of 44 for the first tranche of Vietnam PRSC 1, to a low of six in Tanzania PRSC 1. The number of triggers, too, has come down from an average of 15 in FYOl to 11 in FY05. Still, PRSC prior actions slightly outnumber conditions per tranche in other DPLs. This may attest to the breadth and complexity of the PRSC policy dialogue, and the multi-sector and multi-thematic nature of PRSCs. On average, PRSCs cover six sectors and seven themes, compared to four sectors and four themes for other DPLs. While the average number of PRSC prior actions has declined, the number of non-binding measures in the medium-term PRSC program, has grown. This is evident in the policy matrices of the two countries with the most advanced PRSC programs, Uganda and Burkina Faso. Between PRSC 1and PRSC 4, the total non-binding measures have risen from 12 to 50 in Burkina Faso, and from 18 to 45 in Uganda. In Uganda, this has partly been explained by the fact that government officials view cross-sectoral actions as having a much better chance of success when included in the PRSC. In education, for example, the interlinked program of teacher recruitment and deployment, textbook publishing and classroom construction, all fall within three separate ministries outside of the Ministry of Education and Sports. As a result, many ministries have attempted to have their cross-cutting goals included in the PRSC policy matrix. 1 Average Numberof Binding Measures 110 100 90 PRSCs !3OtherDPLs no 70 60 50 40 ! 30 I I 20 10 - I 1 , UGANDA B U R K I N A F A S - --, - The experience in several countries suggests the difficulty of keeping binding conditions to a minimum, given the comprehensive reform agendas and the multiple donors with differing but legitimate development priorities. In Vietnam, PRSCs have dealt with 22 ministries and line agencies on reforms covering some 15 policy areas. To support progress across such a broad front requires annual initiatives in most areas. Moreover, a dramatic narrowing in the number of actions undermines the coordinating role that PRSCs can or already play among the donor community. It is sometimes difficult to keep a donor on board without making room for its key areas of interest, which are generally quite diverse. In other countries, PRSC policy matrices have been the object of criticism precisely for their size, the number andlor relative weakness of policy measures, the level of operational specificity contained in them, and their usefulness to the government. The size of policy matrices may reflect their use as broad reference documents or a broadening of the reform agenda as new areas are covered in a multi-sector approach. A broad scope may also be explained by the inevitable complexity of government functions. Some of these issues can be addressed as a presentational issue, particularly when matrices tend to be longer as governments find them useful as management tools. To the extent that PRSPs become more operational and prioritized, the PRSC program documents also need to provide fewer details and could simply refer to the government's PRSP-based action plan. At the same time, strong commitment and ownership of the reform program has typically been expressed and demonstrated by government officials in countries where the matrix is derived through a negotiated process. 3. Improving Strategic Budgeting 55. In addition to elaborating strategies and specifying priority reform measures, governments can make the PRSP more operationally relevant by linking it to other country processes. Most notably, reform measures should be costed, and PRSPs strongly linked to the annual budget allocation and execution, as well as to a medium-term expenditure framework (MTEF) that establishes a relationship among resources projected in the medium-term and likely expenditures. Such links serve as a budgetary reality check to policymakers. Given the resources that are expected to be available for both the immediate and medium term, governments must prioritize and rationalize the PRSP's broad array of strategies, programs, and reform mehsures. However, costing and budget linkages remain weak in many countries. These weaknesses stem from a variety of causes, ranging from institutional (different institutional arrangements for PRSP and MTEFs) to external (unpredictable resource flow^).^' According to a CDF progress report, only six PRSP countries are using MTEFs to prioritize the content of their PRSPs and to link them to the budget. Four of these countries-Nepal, Rwanda, Tanzania, and Uganda-are implementing PRSCs. Nearly 40 percent are taking action to ground PRSP priorities in the budget and to the MTEF, while about 35 percent are in the planning stages of developing MTEFs. Another 10 percent of countries surveyed have taken little or no action in this area.'' All of the MTEFs are at very different stages of development and institutionalization, and the support provided by the PRSC has also varied considerably from country to country. 56. Role of PRSCs in Strategic Budgeting. Almost all PRSCs monitor resource use- through either policy dialogue or explicit measures in the matrix-to ensure consistency with PRSP priorities, to maintain fiscal discipline, and to enhance allocative and operational efficiency. In addition, many support developing and institutionalizing the MTEF for improved medium-term strategic planning. In some countries that are at the pilot stages of developing the MTEF, such as Armenia, or the planning stages, such as in Senegal, the PRSC supports these processes through complementary budget reforms, such as on budget classification and reporting, and decentralized execution in line ministries. In Nicaragua, where the MTEF was also in the planning stages, the PRSC explicitly sought to accelerate this process by including measures in the policy matrix to establish and prepare MTEFs. In countries where MTEFs have already been established, such as in Mozambique, PRSCs have helped to institutionalize them by including measures to ensure that budgets are based on and consistent with PRSPs and drawn from the MTEF; and in Burkina Faso, the adoption of the MTEF was complemented by PRSC support for training sessions on program budgeting for budget officers in several line ministries. Where the linkages between the PRSP and budget processes are more advanced, and priority sectoral programs are reflected in budget allocation decisions and their costs are integrated into the MTEFs, the PRSP is at the front end of the budget process. Uganda. PRSP objectives are linked to the MTEF and the budget through a Poverty Action Fund, which identifies expenditures in those sectors with a high impact on the poor and protects them against budget cuts. Sector ministries and national stakeholders participate in the preparation of the budget through sector working groups. To date, allocations toward pro-poor sectors such as primary healthcare, rural roads, agriculture, primary education, water and sanitation now cover 37 percent of the budget, up from 18 percent in 1998. While the MTEF preceded the PRSC and initiated by the government, the PRSC program was an integral part of strengthening this process, with policy 5 1Poverty Reduction Strateg),Papers- Progress in Implementation, World BanWIMF, September 2004. 5 2Enabling Capacity to Achieve Results- 2005 CDFProgress Report, OPCS, World Bank, June 2005. dialogue and specific measures in the policy matrix seeking specific actions, such as integrating donor projects into and ensuring consistency of the pay reform strategy with the MTEF. Tanzania. The government is moving toward an output-based budget. A participatory PER, which underpins budget formulation and informs the preparation of the MTEF, is undertaken by sector working groups consisting of sector ministries and key domestic and external stakeholders. The MTEF and PERs have allowed for significant improvements in strategic resource allocation in priority areas identified in the PRSP. Per capita public spending doubled for agriculture, education, health, HIVIAIDS, water and roads between FY99 and FY04, and is budgeted to increase further in FY05. The institutionalization of the MTEF preceded the PRSC and the added value of the PRSC has been to monitor and maintain dialogue with the government to ensure continued links between PRSP priorities and budget framework. Nepal. The annual MTEF presents normal case and alternative case scenarios, and links resource allocation to expenditure reporting and performance. As a prior action under PRSC 1, the government scaled down and prioritized the development budget that had grown sharply. Remaining projects were classified into different priority categories. Overall, the share of education, health, and potable water increased from 21 percent of public expenditures in 1998-99 to 24 percent in 2003-04. The prioritization exercise has also led to a shift towards greater pro-poor programs within sectors, so that within education, the share of the budget to primary education increased from 44 percent in 1999-00 to 55 percent in 2002-03. In health, the share devoted to preventative health care increased from 26 percent in 2001-02,to 40 percent in 2002-03. It is noteworthy that although security spending has increased in recent years, pro-poor spending has been protected. 57. Lessons Learned. While notable progress has been made, in some cases within a short period of time and with very positive results, such as in Rwanda and Albania, challenges remain. Albania failed to meet its expenditure objectives in education and health due to fiscal realignment and weak budget preparation and sector strategies in the line ministries. In Vietnam, the preparation of annual budgets is informed by the government's five-year Socio-Economic Development Plan (SEDP), which is the overarching development strategy to which the country's PRSP and sectoral development plans provide operational support. In Burkina Faso, the readjustment of budgetary ceilings following resource shortfalls and the lack of integration between the ongoing program budgeting exercise and the MTEF limited the impact of the MTEF. The resulting weaknesses with the health ministry's program budget did not allow the health budget to maintain its share of total budget vis-a-vis other priority sectors. While an in- depth analysis is beyond the scope of this paper, lessons learned from a recent study conducted by OD1 include: (a) the importance of putting the basics of budget management in place, preferably before the adoption of the MTEF; (b) a single central agency to coordinate expenditure as well as oversee PRSP implementation would greatly strengthen budget linkages to PRSP priorities; and (c) strong sector working groups that coordinate strategic approaches to budgeting helps cement links between policy, planning, and budgeting at the sector level." 53 A Review of Experience in Implementing Medium-Term Expenditure Frameworks in a PRSP Context:A Svnthesis of Eight CozmtryStudies, ODI,May 2003. 4. Monitoring and Evaluating PRSP Results 58. The PRSP has underlined the importance of strong monitoring and evaluation (M&E) frameworks to keep track of development outcomes and to assess progress toward development targets. Such a framework involves several elements: a manageable set of defined and measurable indicators; the availability and reliability of data; and the institutional arrangements to coordinate and use data for effective policy-making. The Joint Reviews of PRSP Implementation notes implementation issues in achieving these elements despite steady progress in many countries. The JSAsIJSANs of all PRSC countries almost always note the need for greater attention to M&E frameworks to support PRSP implementation, in terms of more coordinated institutional arrangements, and improving the measurability, specificity, and policy- outcome linkages of indicators. PRSCs have helped address these shortcomings. While the coverage of M&E systems varies from country to country, there are some notable good examples where M&E forms a central part of the program and which show the path for other PRSC programs. 59. Role of PRSCs in M&E. According to the CDF progress report, three countries, Uganda, Tanzania, and Mozambique, have well-developed national-level systems linked to those within line ministries and producing reports to domestic and external stakeholder^.^^ Intermediate indicators are manageable in number and relatively easy to monitor. In other countries, M&E systems are a work in progress and the role of the PRSC in countries such as Ghana and Senegal, has been to use the policy dialogue to keep track of and maintain the momentum of reforms to collect timely data; to address resource and personnel constraints; and to clarify institutional arrangements within government, among others. In other countries, such as Burkina Faso, Cape Verde, Rwanda, Pakistan, Vietnam, and Honduras, the PRSC includes explicit actions in the matrix to accelerate implementation of reforms to improve M&E systems, such as completion of technical plans describing roles of M&E personnel; developing measurable indicators in various sectors; collecting disaggregated education data at the district level; and strengthening of statistical systems through training and improved infrastructure. The Bank has also provided technical support to complement several PRSCs through trust funds that improve statistical capacity and PRSP implementation, as well as targeted capacity building operations. 60. Good Country Examples. A few PRSCs have placed M&E front and center of the reform program. In Albania and Tanzania, where the PRSC programs are relative small and narrowly focused on cross-cutting public sector reforms, the Bank has worked with development partners to support the entirety of the government's M&E agenda through the PRSC and other budget support programs, much as sectonvide programs in education have been folded entirely into other PRSCs. Through a high profile instrument like the PRSC, the Bank signals the importance of M&E as an integral part of the poverty reduction strategy. Work on M&E in Albania has focused on improving the availability and reliability of data, as well as in institutionalizing the M&E functions across line agencies. In Tanzania, which has a sound framework in place, the PRSC is addressing the capacity constraints of core M&E entities; the linkages and coordination between sectors; and the coordination between central and local governments. These two PRSC programs have included a broad range of policy initiatives, financing, and activities: the development of a M&E strategy for the PRSP; budget allocation to fund and staff the statistical institute; identification of PRSP objectives, indicators, baseline, and targets: carrying out household survey programs; producing PRSP APRs; and strengthening and 54 Enablirzg Capacity to Achieve Results-2005 CDF Progress Report, OPCS, World Bank, June 2005. sustaining the capacity of pertinent agencies. However, challenges remain. In Albania, ministries have varying capacity to define and monitor indicators and to analyze data to make a meaninghl input into the policymaking process. The challenges in Tanzania are to Wher define and sharpen outcomes and related indicators and clarifying links between outcomes and sector strategies, and between outcomes and budget allocations. The M&E framework in Tanzania is made even more critical with the introduction of greater results orientation in the second PRSP for 2005-2010, which is organized around key outcomes in three clusters. 61. Indicators. Because PRSPs often have weak M&E frameworks, it is difficult for PRSCs to draw targets and indicators directly from the PRSP; and when they do, the indicators tend to vary widely in scope and quality." While good practice suggests aligning with the PRSP, there is a potentially difficult tradeoff if PRSP indicators are deemed unrealistic or too high-level to measure budget support impact well. Hence, many PRSCs-as well as multi-donor budget support programs-include indicators in their M&E frameworks that are agreed to with government and consistent with PRSP strategies, but are separate from the PRSP itself. The experience of the EU also reveals the difficulty in solely relying on the PRSP for outcome indicators against which to di~burse.'~Even among PRSCs that do not strictly draw indicators and targets from the PRSP, there is varying quality in the choice and formulation of indicators. In some cases, it is difficult to gauge how specific policy actions correspond to the stated outcome; in other cases, indicators are not easily measurable; and in other cases, output indicators are used in place of outcomes (see Table 5).57 These kinds of problems undermine clarity and predictability. However, there has been notable progress in building good results frameworks in PRSCs. Reflecting progress in PRSPs, recent PRSC programs have tended to outline a clearer results framework in the program document, working with government counterparts to better link specific measures to output and outcome indicators, setting quantitative targets, and providing baseline data. There is also generally a growing emphasis on providing support for M&E, ensuring not only the availability and quality of data, but also the demand for and management of its use. 55 N. Hicks, Monitoring Performance in Poverty Reduction Support Credits, A Note on Best Practice, PREM, World Bank, May 2004. 56 EC Budget Support: An Innovative Approach to Conditionality, European Commission, February 2005. 57 Good Practice Notes for Development Policy Lending, World Bank, October 2004; and Norman Hicks and Klia Bassing, Monitoring Performance in Poverty Reduction Support Credits,A Note on Best Practice, World Bank, 2004. Table 5. Examples of Linking Actions to Outcomes Input Indicator/Policy Action OutcomeIndicator Good Examples Introduce differentiation of benefits within the At least 70% of poor households receive family poverty family poverty benefit benefit; leakage from this benefit falls from 27% in 2003 to 25% in 2007 - - IInitiated policy formulation on early services cover 30% of education with wide participation. least 50% of the poorest quintile of families by 2007 ~ Adhere to the rationalization program in the Between 2003 and 2007, infant mortality to fall from I2 per hospital sector as measured by target beds ration 1,000births to approach mean OECD levels or MDGs. )and physicians ratio I I I Bad Examples 1 Review targeted areas, such as cargo handling Increased exports and stevedoring, and develop an implementation plan to improve performance Complete independent survey of regulatory and Reduce time required to register a business administrative costs on business -- - - - - - Update profile of vulnerable groups using the The social protection strategy will be recent household survey and publish results implementation and monitoring of PRSP. 62. Using Outcome-based Conditions. Outcome-based conditions have been used pragmatically when appropriate. Initial PRSCs typically provide indicators and targets related to policy actions, although in some cases indicators are given without targets, particularly where fkture performance is uncertain, or there is no goal agreed to with the government. As PRSC programs evolve over time, there is often a steady increase in the emphasis given to, and the quality of, outcome and impact indicators and targets, particularly in the social sectors. Areas that are concerned with institutional changes (such as governance and public financial management), however, are less amenable to quantifiable outcome indicators than to process indicators. In some cases, particularly where outcomes and impact are less affected by exogenous factors, results chains are clear, and indicators are available, it might be possible to move to a greater reliance on outcome and impact indicators as conditions and triggers. At the same time, governments are rightly wary of using outcome and impact targets as conditions of and triggers for credits, particularly when they are held accountable for outcomes outside their control. Most PRSCs remain conditional on inputs and outputs, and there are good reasons to remain pragmatic in the use of outcome-based conditionality (See Box 9). Box 9. Issues in Adopting Outcomes-Based Conditionality An interesting design approach is to explicitly link disbursement to outcomes. In principle, this design choice has the advantage of focusing directly on the results of the program while leaving the government greater flexibility in the choice of how to reach them. The EC has tried to apply this formula to the fixed tranche of their budget support in some African countries. Despite the merits of an outcomes focus, the mixed experience to date illustrates the limits of standardizing a formal link between outcomes and disbursements. The EC finds that around 71 percent of a variable tranche is disbursed, with significant delays in some countries and substantial variability of between 65 to 85 percent. It also identifies several problems that have emerged during implementation, including the definition and revision of targets; the fact that institutional measures such as public financial management indicators do not lend themselves to quantification; only half of the countries used their variable tranche reports to analyze the reasons for failure to meet targets; the insufficient quality of data and statistical capacity; and the choice of indicators when national strategies lack key aspects. In the absence of independent long-term evaluations of this approach, early reviews of the experience in Benin, Burkina Faso, Madagascar and Uganda highlight the risks of a shift towards intermediate indicators, which are too distant from the final impact of policies. Interviews with task managers and governments revealed a number of reasons why they were reluctant to tie conditions or triggers to outcomes or impacts-as opposed to inputs and outputs. Outcomes and impacts are not directly linked to actions. While governments can control inputs and outputs, they cannot realistically be held accountable for outcomes and impacts, which are also subject to forces outside of the control of the government. These include, among others, general economic conditions, household behavior, weather, and changes in the terms of trade and politicallsocial events. In some cases, indicators can improve despite there being no government action, but purely because of factors exogenous to government. Data lags. The gathering and processing of information often lags many years behind when it is needed. This is particularly true of impact or outcome indicators (as opposed to inputs), such as indicators like infant mortality and maternal mortality, which are part of the MDGs. In Burkina Faso, these health data are collected only every 5 years, while in Guyana the last household survey was done in 1999-which makes it implausible to make credits contingent on data that is so untimely. Data is often unreliable. Indicators can vary from year to year just on the basis of survey quality, or changes in definitions. For example, it is widely known that how questions are asked can influence how much income or expenditures are reported in a household survey, and infant mortality data may tend to reflect only information on babies born in government clinics rather than at home. Some PRSP targets are unrealistic or absent. PRSP target outcomes and impacts are often optimistic, and cannot realistically be used as targets for a PRSC (if they were used as conditions or triggers, the program would stall over non-compliance). For a PRSC to use other (lower) targets implies a contradiction with the PRSP. In some cases, PRSPs lack targets that can be used (e.g. Albania). Indicators can distort actions. Emphasizing outcome and impact indicators can also serve as an incentive for the government to distort or misreport the data. When the focus is on outcomes and impacts rather than strategies and actions, there is a tendency to focus on the indicator rather than the achievement of real results. The relationship between the indicator and the result thus begin to change (the so-called Lucas critique). For example, if school enrollment is the outcome target, governments might try to over-crowd school resources rather than build more schools or buy more books. The outcome target is met, but education suffers. To keep governments from acting within the letter of the contract rather than the spirit, the donor would be tempted to specify yet more, and more specific, targets in subsequent PRSC cycles. Outcomes and impacts are uncertain and lagged. Often changes and improvements, particularly in the social sectors, have a long lead time before they show up in outcome and impact indicators. For example, improving the education sector today will make people more productive 15-20 years in the future. In many instances, the long-term effects of actions are highly uncertain and it would be presumptuous to assume that the precise nature of the results chain between actions and outcomes is known with adequate accuracy. Institutional measures cannot easily be quantified. Relying on indicators has a tendency to put emphasis on areaslsectors that have good data or in which results can be more readily measured. Thus, while education and health statistics abound, there are few ways of measuring corruption or good governance. Source: N . Hicks, Monitoring Performance in Povery Reduction Support Credits, A Note on Best Practice, PREM, World Bank, May 2004. B. Supporting the PRSP's Comprehensive Program 63. The breadth of policy areas typically covered by PRSPs testifies to the complex and multidimensional nature of poverty. PRSCs, as one of the Bank's primary means of supporting the policy reform agenda, have been subject to competing expectations: the need to encompass the entirety of PRSP policy reforms on the one hand, and the need to be strategic in terms of directing limited staff time and resources on the other. These tensions lead to trade-offs in terms of sectoral and thematic coverage, and in terms of the modality of support provided. In making these trade-offs, the Bank must take a number of critical factors into consideration, as well as address internal incentives to working multi- and cross-sectorally. Not surprisingly, first generation PRSCs have tended to focus mostly on systems and processes that underpin PRSP implementation, notably public financial management, as well as on well-established sectonvide programs. 1. General Trends 64. The broad scope of the PRSP is mirrored in the number of policy areas addressed by the PRSC. On average, PRSCs cover seven thematic 1 cross-cutting areas and eight sectors, about twice as many sectors and themes typically supported by other policy-based lending (Fig. 10-11). Compared to other DPL operations, PRSCs address more systematically the cross-cutting constraints affecting various sectors. For example, all PRSCs have sought to improve core public sector functions, especially the coherence of government programs, policies, and budgets, as well as the quality of and access to pro-poor service delivery through public expenditure management. The PRSC is an effective instrument to support these efforts since it works directly through the government's planning and budget processes to support reforms that enhance the effectiveness of public sector institutions and systems. This cross-cutting approach has built upon and reinforced the coordination and collaboration among different government ministries and agencies that underpin the PRSP process. 65. Sectoral Coverage. Beyond cross-cutting issues, PRSCs have also addressed in-depth sectoral reforms. As already mentioned, one approach has been for some PRSCs, notably in Uganda and Benin, to address the broader sector policy dialogue by subsuming entire SWAps. In other cases, the PRSC focuses selectively on specific sectoral issues while the broader sectoral dialogue is undertaken through other operations. The education and health sectors are typical examples of the former, since these sectors are more suited to SWAP frameworks, i.e., well- defined sector with budget and program. PRSCs have come to be associated with bringing a more social sector orientation within Bank development policy lending. Approximately 95 percent of PRSCs support reforms in the health sector, while 86 percent address issues in the education and agriculture sectors. This compares to only 53 percent of other DPLs that address health, 42 percent that address education, and 23 percent that address agriculture. There is also a significant difference between PRSCs and other DPLs in terms of their coverage of water and sanitation. Roughly 48 percent of PRSC operations address water and sanitation issues, more than double the number of other DPL operations covering this sector. Figure 10. Sectoral Content Among PRSCs Figure 11. Thematic Content Among PRSCs - --- Law & Justlce I Publ~cAdm~n~strat~on Public Sector Governance Health and Other Soc~alSew~ces Rural Developmenl I Human Development Financial and Private Sector Development Environment, Natural Resource Mgt. Industry and Trade Economic Management I Social Protection and Risk Management Water & Sanllatlon Rule of Law DOthm DPLr Trade and Integration lnformat~onand Commumcat~ons Soclal Development. Gender Urban Development 0% 20'0 40". 60% 80% 100% 120911 0% 2040 40% 60% 80% 100% 120' 5hare of Operallon\ Shareof Oprrrl~ons Source. ALClD Database Source: ALClD Database 66. Distribution of Measures. While PRSCs tend to be much more multisectoral than other DPLs, there has been a tendency to support institutional reforms to strengthen public sector governance (see Figure 12-13). More than 45 percent of all PRSC binding measures entail public sector reforms, of which upward of 64 percent specifically address weaknesses in public financial management. It is noteworthy that the share of measures targeting financial management is very similar to that of the EC's variable tranche." The remaining share of binding measures is more or less evenly distributed among other sectors and themes. PRSCs have a smaller share of binding measures (12 percent) that support financial and private sector reforms, compared to other DPL operations (27 percent). These trends are also broadly similar indicative benchmarks, with some minor differences: among PRSCs, the share of indicative benchmarks increases in environment and natural resource management, and social protection and development. The reverse is true for private sector development. 58EC Budget Support: An Innovative Approach to Conditionality, European Commission, February 2005. Figure 12. Sectoral Distribution of Binding Measures in PRSCs (left) and Other DPL Operations (right) - I Agriculture, - 1 I F ~ ~ h l and n g Urater Sanltdt~orl I Law & Just~ce , &Comm 1 ' Law &Justice Publlc , 4drnin 50% I I 27% Other Soclal Trade I 13% Figure 13. 'Thematic Distribution of Binding Measures in PRSCs (left) and Other DPL Operations (right) Urban Trade and Economlc Develop. Integrat~on Urban ~~~d~and I % Economic Environment 8% Dc\elop Mgmt Env~ronment Integration ~~~t and Natural 03% , lo% and Natural 6%. ,- Resource Soc~alDev & 790 Public Seetor Governance Governance 24% Human Development I - - - - --7% - - - Source World Bank ALCID Note. For operations approved through end-February 2005 67. Broadening Coverage. As the PRSC program progresses, there is a typical tendency to increase the breadth of sectoral and thematic coverage. The Burkina Faso PRSC program grew from four sectors to seven. Vietnam similarly grew from five sectors to eight. The Uganda program began with six sectors and has subsequently folded in two more by PRSC 4. The scope within sectors and thematic areas also tends to broaden. The Benin PRSC's public sector agenda initially addressed public expenditure management and corruption but is slated to broaden in the second operation to encompass internal audit and control, procurement, civil service reform, decentralization, and public communication. The approach taken to widen the scope of the PRSC is largely country-specific. In Vietnam, the PRSC reform agenda moved beyond structural reforms in trade, investment climate, and public expenditure management to other cross-cutting issues of governance, land administration, and access to information. As the PRSC program progresses, the policy dialogue deepens, building upon the achievements of the previous operation. Many other PRSCs begin with cross-cutting issues and move to address sectoral issues in depth. Among the more comprehensive PRSCs, specific criteria were used to determine a sector's "readiness" (see next section), with the breadth of coverage reflecting the Bank's and the government's capacity to manage the transition away from projects. 68. Consolidation and "Maintenance". Over time, given a solid track record of reform implementation, the policy dialogue and support to specific sectors have reverted to what has become known as "maintenance mode," which refers to the steady implementation of well- advanced sector program^.'^ In Uganda and elsewhere, this has implied that while sectoral policy dialogue in education continues through annual reviews that monitor progress of overall reform implementation and outcomes, the sector or thematic area would not be subject to specific binding actions, but rather to broad based indicator of performance. For example, measures supporting education reform comprised 15 percent of total measures in Uganda PRSC 1. That share had dropped to two percent of all measures in PRSC 4 (see Table 6). Table 6. Evolution of Sectoral Content of PRSC Binding and Non-Binding Measures (%) - Burkina Faso Uganda Sector Group PRSC I PRSC2 PRSC3 PRSC4 PRSC 1 PRSC2 PRSC3 PRSC4 Agriculture, Fishing, Forestry ' 29 13 8 7 8 10 9 Education Energy and Mining Information and Communications Transportation Water and Sanitation Finance Health and Other Social Services Industry and Trade Law and Justice I Public Admin. 39 50 47 45 15 64 61 74 Total 100 100 100 100 100 100 100 100 69. Mature Sector Programs in Maintenance Mode. Since some PRSCs have folded in SWAps that had been in existence for some time, they have typically been placed on "maintenance mode" at the outset of the PRSC program. This entails no binding actions to the sector, but rather general evidence of "satisfactory progress in implementation." Such is the case for the "mature" sectors in Uganda (education, health, and water), Madagascar (education, health), Benin (Forestry, Environment). The pace of reforms in a given sector or thematic area ultimately determines how quickly a maintenance mode could take shape. In Burkina Faso, given the difficult transition in the health sector from project to program support, the share of policy measures remained constant throughout the program, reflecting the need to maintain a level of engagement to ensure continued reform implementation. Similarly, given the critical and complex nature of the public sector reform agenda, particularly in public financial and expenditure management, a consistent share of policy measures reflect the need to ensure incremental progress. 2. Public Financial Management 70. Despite the breadth of coverage in PRSCs, most programs have managed to maintain a core emphasis, stemming from the HIPC initiative, on increased poverty-focused expenditures and expanded service delivery. This has, in turn, typically been accompanied by a strengthened 59 Terminology from World Bank Africa Region Strategic Framework ,for IDA'S Assistance to Africa: The Emerging Partnership Model, June 25,2003. drive to improve public financial management. While it is still too early to determine the long- term poverty impact of PRSCs, preliminary results in a number of key areas in several advanced PRSC countries testify to the significant contribution of the PRSC approach to the implementation of reduction irograms. 71. Modernizing PFM Systems. Public policy and budget execution are the two instruments through which governments achieve the objectives of their PRSP. The Bank recognizes the critical role that a transparent and efficient public expenditure framework plays in improving the quality of and expanding access to pro-poor service delivery. More broadly speaking, public financial management (PFM) represents a key government responsibility that contributes to improved overall public administration. Modernizing PFM typically entails improving controls, planning and execution procedures, transparency, and accountability. Given the large donor resource flows to many low-income countries, modernizing PFM is a legitimate interest for donors. particularly for those that provide assistance through budget support, since it is critical to reducing fiduciary risks. For PRSCs, the PFM represents a core component of the reform agenda. This is manifest in the significant analytical and diagnostic work undertaken in PFM, the policy dialogue that centers on PFM issues, and the relatively high concentration of PRSC measures that support the modernization of PFM systems. 72. Ex-ante Underpinnings. All PRSCs have been underpinned by ex-ante fiduciary assessment of the country's PFM and expenditure systems. The CFAA, CPAR, and PER typically focus on aspects such as the comprehensiveness and transparency of the budget, systems that facilitate implementation and monitoring of the budget, fiscal transparency (reliable information on fiscal results and position), and financial accountability for use of public resources. The purpose of the assessments is to identify deficiencies, assess the level of fiduciary risk, and identify additional steps needed to secure acceptable fiduciary arrangements. PFM-related measures with PRSCs have covered a range of issues, including budget preparation, budget execution, accounting and reporting, audit and oversight, and diagnostic studies. Many of the refonns are legal or regulatory, such as enactment of laws on fiscal responsibility, financial administration, procurement, and internal audit. Other measures are more institutional and include the introduction of medium-term expenditure frameworks and information systems, legislative oversight hnctions, and new budget classifications. 73. PFM Measures. Given Table 7. Share of PFM Conditions in PRSCs (%) that the PRSC cycle is short, it has FYOl FY02 FY03 FY04 Avn. been common to support Budgetpreparation 13 40 50 42 4 0 I I institutional reforms such as Budget execution 25 0 14 12 13 Accounting I reporting 25 0 14 38 26 changes in PFM. PRSCs contained Audit and ovenight 25 60 14 8 17 21 policy actions related to Diagnostic studies - 13 0 7 0 4 legislation in PFM, such as Total 100 100 100 100 100 enactment of a Budget Law. Fiscal Solrrce: Financial LManagement Issues in PRSCs, FYOl-04, " Responsibility Act, Financial November 2004 Administration Act, and Procurement Decree. The PRSC program documents discussed the enactment of such legislation but there is not always sufficient attention paid to implementation and how such reforms would improve PFM. The other issues often addressed are the introduction of medium-tenn expenditure frameworks and integrated financial management information systems, and strengthening internal audit, external audit, legislative oversight functions, and new budget classifications. In many cases, PFM conditions have been process- oriented and reflecting difficulties in establishing measurable indicators in this area. However, some of the process-oriented indicators have intermediate outcomes, which are measurable. For example, the Sri Lanka PRSC provided 12 intermediate monitorable and measurable indicators, which should be viewed as good practice examples (Table 7 shows the distribution of PFM conditions in PRSCs). 74. Preliminary Results. Since only a few countries have progressed to subsequent PRSC operations, it is still premature to judge the extent to which PRSCs are contributing to improvements in basic PFM infrastructure. In Uganda, there have been improvements; however, the 2004 CFAA confirms that fiduciary risks are still high because not all of the risks have yet been properly addressed. Although some PFM progress has been made in Vietnam, public resource management there fell significantly short of international best practices. In particular, the government still lacks a rigorous and transparent MTEF. The latest HIPC tracking assessment concluded that Burkina Faso has a solid PFM system (see Box 10). In 2001, it met eight out of 16 benchmarks; in 2004, it met nine out of the 16. At the start of the Tanzania PRSC program, the country had met eight out of the 16 benchmarks, according to the 2001 HIPC assessment, making the country one of the top performers in Africa. The country currently meets 11 benchmarks. Box 10. Public Financial Management Reforms in Burkina Faso Burkina Faso's PRSC program has focused on strengthening domestic budget processes, improving auditing, reforming civil service, and overhauling procurement regulations. ~ h r o u fo& ~ h operations,-the program has achieved several notable results: Budget,fortnulation. Significant progress was made in mastering the preparation of global MTEFs. Their full integration into the budget preparation cycle in 2002-03 led to increasing realism of revenue targets and improved budget arbitration. Moreover, program budgets linking PRSP objectives with budget allocations were established in all key ministries. Budget formulation in key sectors also benefited from PERs, which have been piloted since 2003 by a PER supervision committee. Budget execution. The government fully operationalized a computerized expenditure management system in 1999-2000,which captures all domestically financed expenditure at the different stages of the process and allows close monitoring of budget execution and the level of payment obligations. Moreover, the legal framework was created to delegate credits to regions and provinces and to deconcentrate the issuance of payment orders in order to accelerate delivery of basic services. Budget reporting. The government caught up with its legal obligation to submit budget execution reports to the audit court for the years 1996-2001. To strengthen external controls, an independent Supreme Audit Court was created and made fully operational in 2002. A new organic budget law was adopted in 2003. Moreover, the CFAA was finalized and its recommendations were integrated into the new budget management reform plan adopted in July 2002. Procurement. Roles for the different ministerial departments were clarified by the government. A new procurement code was adopted in May 2003, new procurement manuals and forms were drafted, and a capacity building plan was drawn up. Source: Program Document,for Burkina Faso PRSC 4 (28293-BUR), World Bank, April 2004. 3. Service Delivery in Social Sectors 75. Public service delivery, particularly in the social sectors, has a direct impact on the quality of life of the poor and is a central element of most country PRSPs. Donors have placed considerable importance to this area, manifest in the HIPC initiative's allocation of debt relief to the social sectors and regular Bank tracking of expenditures on health and education in HIPC countries. The PRSCs help to harmonize existing sector-focused projects, leveraging their poverty focus by ensuring appropriate expenditure allocation (level, structure and execution of the budget) and by addressing some of their financing implications for the poor (e.g., tariff setting, cost recovery policies). The PRSC's role in supporting service delivery differs from country to country; some programs address only education and health, while others include agriculture, water and sanitation, nutrition, and other sectors are included. The modality of support also differs; in some countries, PRSCs complement existing investment projects by addressing cross-cutting issues as well as targeted complex sector issues; in others, the PRSC is the Bank's sole instrument for providing sectoral policy dialogue and financing, either through a SWAP framework or through discrete interventions in key areas. 76. SWAP Frameworks. For the Bank, SWAps in service delivery are typically prime candidates to include in budget supportthey have ready-made sector strategies that are central elements of many PRSPs; they entail fairly strong line ministry ownership and capacity; they typically have a medium-term timeframe, monitoring and evaluation framework, and are comprehensive; yet they address detailed sector issues. Donors have also used a variety of financing modalities to support SWAps, traditionally through donor pooled funding;60as such, they have provided a convenient platform for broader harmonization. For these reasons, PRSCs have increasingly folded in SWAps to support public service delivery, primarily in the social sectors, but also in non-social sectors, such as transport and water and sanitation. For recipient countries, the budget support modality has allowed full flexibility on spending-this was why in Ethiopia, for example, the ministries of health and education preferred to fold in the SWAps under the PRSC umbrella. However, many line ministries are concerned that shifting to budget support provides no guarantee of funding, raising the importance of strong public expenditure management and integration of sectoral and central budget processes. This was a concern expressed in Honduras, where SWAP financing will remain separate from the PRSC, but where certain sectoral issues will be brought up in the context of the PRSC policy dialogue. To date, almost all African PRSCs have folded in (or are in the process of folding in) at least one existing SWAP program. In many cases, the PRSC is helping governments to develop SWAps for future financing and other support. 77. Addressing Cross-cutting Issues. The PRSC has been an effective complement to SWAps and other sector-supported projects. In some cases, the PRSC has addressed cross- cutting issues that impact service delivery, such as recruitment and procurement. In Albania, where sector financing is provided through on-going investment projects, the PRSC has focused on cross-cutting issues. It has for example, contributed to further strengthening the links between the MTEF and social sector strategies. In education, it has introduced a public expenditure tracking survey to identify issues related to budget formulation and execution; supported the restructuring of the ministry, including staffing requirements, job descriptions, and costs; supported an audit on management, finances, and human resources of the ministry. In Ghana, the PRSCs have managed to shift the public sector reform program toward service delivery and the need to improve civil service structures and incentives in key sectors, such as education and health. Attention has also focused on payroll management and control because of the need to ensure a hard budget constraint while the institutional reforms are implemented. 78. Addressing in-depth Sectoral Issues. In many cases, PRSCs complement sector investment projects by addressing detailed sector issues, particularly those that have not been adequately dealt with through projects. In Ghana, the PRSC complements an education project that increases the availability of facilities, equipment, textbooks and teachers. The role of the '' Fiducia~Arrangements for Sector-Wide Approaches (SWAPS), OPCS, World Bank, April 2003; Andrew Cassals, A Guide to Sector- WideApproaches.for Health Development, WHO, 1997; CIDA Primer on Program- Based Approaches, CIDA, February 2004. PRSC has been to focus on selected policies that require broad consensus within government to be tackled. These issues include the implementation of an incentive scheme to provide scholarships to girls, the removal of government fees for primary schools, and the introduction of capitation grants to primary schools in deprived communities. In Rwanda, the objectives of a water resources project are to integrate the management of water resources, improve water access to poor people, and improve government capacity in the sector. The PRSC supports these objectives by focusing on output-based mechanisms and increased private sector management of water supply and sanitation. 4. Selectivity and Sequencing of Coverage 79. The multisectoral approach of the PRSC has raised some important questions that reflect tensions and competing expectations. PRSCs are meant to anchor the Bank's policy dialogue across a broad range of areas; thus there has been pressure for the PRSC to mirror the breadth of policy areas contained in the PRSP. At the same time, a PRSC's coverage cannot be too comprehensive without becoming overloaded. 80. Managing Trade-offs. These competing expectations have manifested themselves in different ways. Concerns have arisen within the Bank, for example, that not enough attention has been given to the rural development and growth agendas with the same depth and breadth as other sectors, notably health and ed~cation.~'Similarly, guidelines are being drafted to further advance water and sanitation coverage in PRSCS.~~At the same time, discussion has surfaced as to the feasibility of having multiple PRSCs in a country, each addressing the policy dialogue of different sectors and themes (e.g., PFM, social sectors, etc.). In several countries, such as Albania, Pakistan, Honduras, and Tanzania, the PRSC is running concurrently with other policy- based operations. In determining the scope and coverage of the PRSC, teams must assess the inter-sectoral and intra-sectoral trade-offs in the PRSC, the specific type of support provided to the sector (policy dialogue, financing, maintenance) and a host of other considerations that inform priorities for overall growth and poverty reduction. In this regard, country specific solutions will have to be defined as part of the CAS process without expecting uniform treatment across all sector and thematic areas of the PRSP (Box 11 provides a good case study on the selection of sectors and themes in the Benin PRSC). 81. Considerations Determining Content. In ultimately deciding which sectors and themes to support, how to support them, and when to sequence them into the PRSC, the Bank takes a number of factors into consideration: Government Preferences. Providing the government the space to express its opinion on what, when, and how the PRSC supports particular areas of the PRSP is critical in reinforcing country ownership. This has frequently proven to be a delicate balancing act for the Bank; providing advice to the government on the major impediments to poverty reduction must be balanced with the need to foster a genuinely country-driven process. As already noted, in many African countries, the consolidated approach that folds in many sectoral programs and investments under the PRSC responds to governments' call for less fragmentation and greater ownership of development assistance. In many non- '' A Review ofRural Development Aspects of PRSPs and PRSCs, 2000-2004, ARD, World Bank, 2005. 62 Using Budget Support and PRSCs to Scale up Rural WaterSupply (Draft), Sanitation and Hygiene, Energy and Water Department, World Bank, forthcoming. African countries, the preference has been to maintain sectoral financing through SWAps in order to guarantee a proper level of funding and technical support. More broadly speaking, the PRSC's heavy emphasis on improving and expanding pro-poor service delivery reflects the priorities of most PRSPs. Box 11. Selectivity and Sequencing of Sectoral Programs in the Benin PRSC In Benin, the selection of sectors in the PRSC was determined by three major and weighted considerations: (1) the Bank's comparative advantage and the potential for coordination and synergy with other donors; (2) the consistency and continuity with the CAS under which the Bank supports the Government's strategy to move toward programmatic lending and sector projects to be folded in programmatic support; and (3) the readiness of the sector, that is whether conditions for ensuring effective participation of the sector in programmatic financing are considered acceptable, judged notably by the quality of the sector strategy, the reform agenda for increasing expenditure efficiency, and the capacity to prepare and implement program budgets. The core sectors supported by the first PRSC, namely, education, health, and rural water and sanitation, were identified as priority sectors in the PRSP and have developed over the past three years a minimum level of capacity and ability to participate in programmatic support lending. The water and sanitation strategy was prepared by the Government with the support of sector donors ten years ago and continues to provide the sector with a sound framework for increased public expenditures while addressing remaining policy issues. Similarly, a health sectoral strategy and policy has been in place since 1995 and has served as the basis for the design of multi-year expenditure programs supported by donors through budget support and investment projects. An education sector policy and strategy was adopted by the Government in 1998 and, helped by the PRSP's operational plan for the sector, has provided a framework for performance-based programs in basic education. The PRSC will also address policy issues that are critical to help the country accelerate the pace towards the MDGs (universal completion of primary education). Source: Program Doczlmentfor Benin PRSC 1 (26272-BE) Sectoral Readiness. Among the African PRSC programs that are consolidating the lending portfolio under the PRSC, the question of sectoral readiness is of paramount concern. Budget support implies a loss of control over program implementation and earmarked financing. The experience with the health sector in Burkina Faso illustrates the need to ensure a calculated transition to budget support. In Uganda, the Bank has established strict criteria for sectors entering the PRSC framework. Criteria include the quality of the sector strategy, an existing sectoral MTEF, a strong monitoring and evaluation framework, strong coordination with the central ministries, and overall administrative capacity of the line ministry. In Rwanda, the approach was to include the energy and water sectors in the first PRSC with the intent to prepare these two sectors for fuller engagement in subsequent operations. In order to make informed decisions on the reforms to support, PRSCs must also be sufficiently underpinned by a set of analytic work. In Uganda, although there was no previous sectonvide approach in the water and sanitation sector, the Bank had undertaken extensive analytic work on the needed policy and institutional reforms that had created the basis for a strong program supported by the PRSC. Indeed, one explanation for the apparent weak coverage of growth and rural development among PRSCs is the lack of coherent strategies in these two areas in the PRSPs (see Box 12). Bank's Comparative Advantage. The content of the PRSC must be in accordance with the Bank's Articles of Agreement which prohibit the Bank's direct involvement in political concerns. This has been brought up in the context of efforts by several bilateral donors to attach political governance conditions and human rights concerns to joint budget support programs. The PRSC must also reflect the Bank's comparative advantage in the country, and in almost all cases, the areas targeted by the PRSC are those where the Box 12. Coverage of Rural Development and the Growth Agenda in PRSCs The PRSC has been strongly identified with carrying out the PRSP's emphasis on social sector reforms and expenditure. Many view this emphasis as precluding adequate focus on other critical areas vital to poverty alleviation, such as rural development and private sector-led growth. While more than 85 percent of PRSCs support reforms in these two areas, a relatively small share of binding measures target them specifically: 5 percent for rural development and 15 percent for private sector development (compared to 10 percent and 32 percent in other DPL operations, respectively). This compares to about 20 percent of PRSC binding measures supporting social and human development reforms and 45 percent addressing public sector governance issues. There are several factors that explain this trend. The first generation of PRSCs have focused appropriately on the overall framework for program lending, namely public financial management, which accounts for about 35 percent of total binding measures. Also, sectors that have had long-standing sectonvide approaches with developed MTEFs, a good track record, monitoring and evaluation frameworks, analytic underpinnings, and relatively strong ministerial capacity, have naturally been the first sectors picked up by PRSCs. On this note, a review of JSAs in PRSC countries reveals that nearly 67 percent of PRSPs need to provide greater elaboration on underlying strategies, including rural and private sector development. A study conducted by USAID on the PRSP's coverage of growth confirms this finding.63 More broadly speaking, one may concede that the cross- sectoral nature of rural and private sector development make it difficult to integrate these two policy areas in a budget support framework, unlike well-defined sectors such as health and education. In many countries, rural and private sector development is characterized by institutional fragmentation with multiple ministries involved. What is incontestable is the critical importance of rural and private sector development in promoting economic growth and poverty reduction. In this regard, future PRSPs are indeed expected to more fully address growth and to elaborate on rural development issues. The PRSC programs in Pakistan, Sri Lanka, and Vietnam have placed the growth agenda at the center of the reform program. Several PRSCs, notably Madagascar, Rwanda, Mozambique, and Ethiopia are helping to prepare sectorwide programs in agriculture or to undertake comprehensive rural development strategies. Bank has had significant experience in. There are several exceptions, however. In Cape Verde, the Bank had very little previous experience in health sector investments but the PRSC was seen as an opportunity for the Bank to engage in policy dialogue on the health system, particularly insurance mechanisms and other financial framework issues. In Senegal, the areas of justice and labor had been relatively neglected by donors but are priority areas for the PRSC, given their significant associated externalities for the country's growth and equity agenda. Comparative advantage also extends to the PRSC itself and the value added it offers as a high profile instrument. In Ghana, for example, local stakeholders broadly view the PRSC as the centerpiece in policy dialogue, supporting key issues in growth, service delivery, and governance. The selection of issues and their elevation to the center of government would increase the chances of progress on decisions, with maximum buy-in from important segments of the society. Complementarity and Harmonization with other Donors. The Bank looks to avoid duplication with other donors and as much as possible, to harmonize budget support programs. In Albania, the decision to keep the PRSC relatively small reflected the CAS emphasis on infrastructure financing, a necessity in light of the vacuum left by the EU's shift away from this critical area. In Tanzania, donor harmonization resulted in the PRSC focusing narrowly on cross-cutting issues, an arrangement that had been in place by the existing budget support donor group. With the government drafting the next PRSP, the Bank and other donors are discussing the possibility of extending the coverage of the joint budget support programs to include social and other sectors. 63 The Treatment of the Private Sector in AJi-icanPRSPs and APRs, USAID, November 2004. "Aid Modality". There are obvious interventions that are more appropriately supported by well-targeted project financing rather than PRSCs. These include large multi-year infrastructure type projects that require up-front commitments, multi-year contracts, and international procurement. Similarly, capacity building entails specific activities linked to a defined budget that is more suited to project financing. This also applies to other types of interventions, such as micro-finance and CDD approaches, which address issues at the local or grassroots levels. 5. Implications of Multisectoral Approach 82. The multisectoral approach of the PRSC reflects the Bank's internal incentive framework. On the one hand, many sector staff have recognized the limited institutional development impact of sector projects when they rely heavily on PIUs that are often weakly integrated into government structures. They also acknowledge the PRSC's advantage of engaging in policy dialogue directly with decision makers within the government, particularly on cross-cutting issues that impact sectoral service delivery. Issues such as procurement procedures, budget management, and legislation must be addressed at the level of the central government. This could only be done through an approach such as the PRSC that involves cross-ministerial dialogue, coordination, and collaboration. The multisectoral approach also allows teams to identify cross-sectoral linkages to poverty. Expanding access to education in rural areas may entail improving transportation services and rural roads. 83. Policy Dialogue. While it is still relatively early to arrive at definitive conclusions, preliminary results have shown that the nature of the sectoral policy dialogue has been enhanced. As one sector staff noted, "the discussions no longer revolve around the procurement of what types of ceiling material used in school construction but rather around substantive policy choices in education reform." Particularly for cross-cutting constraints that affect service delivery, PRSCs often provide a more effective platform than sector projects. The Albania Country Assistance Evaluation (CAE), for example, notes that the PRSC's cross-sectoral dialogue created opportunities and incentives for sector staff to rely on the PRSC to advance their agenda rather than rely upon their own sector programs.64 84. Staff Issues. On the other hand, sector staff have expressed a number of concerns over the potential impact of consolidating projects under the PRSC banner. For many, the traditional project framework has offered a clear added value. The Bank uses its expertise to help the government identify, prepare, finance, and supervise projects. With the PRSC, there is less hands-on involvement in implementation, raising questions over local technical capacity. For example, would the government, through budget support, hire a foreign consultant to fill in technical gaps? Given capacity constraints, how could quality be assured? Are budget processes robust enough to ensure that sectors are properly funded? Within the Bank PRSC team, would there be enough resources in terms of staff, budget, and timing to address the host of issues in a given sector, e.g., health sector issues include health financing, pharmaceuticals, recruitment, training, construction, etc. Another concern relates to the flexibility of the PRSC and the possibility of full disbursement even when sectoral performance is judged inadequate. In this regard, what is the role of sectoral staff when it comes to assessing sectoral performance beyond specific binding actions? Many of these concerns have manifested themselves in attempts to load PRSC policy matrices with policy actions across a broad range of sector issues. Sector staff 64 rllbar7ia Cbuntty Assistance Evaluation, OED, World Bank, forthcoming. have also called for a share of the Bank budget to adequately prepare and supervise the PRSC. A parallel concern has been the potential undercutting of funding to sectoral ministries. 85. Co-Task Management. These difficulties are underscored by the fact that eight of the 33 PRSCs have been co-task managed by multiple Networks. The Albania PRSC introduced this approach by setting up a co-management arrangement between the Human Development (HD) and the Poverty Reduction and Economic Management (PREM) Networks. Other PRSC programs that have been co-task managed include: Benin (PREM, HD), Cape Verde (HD, PREM), Ethiopia (PREM, HD), Honduras (HD, PREM), Mozambique (PREM, CMU). Co-task management requires setting aside the traditional single manager model to focus on a more client-centered approach to preparation of the PRSC. While at times costly, time consuming, and challenging, co-task management has paid off in a much more coordinated PRSC task team that has been able to link public sector and social service reform agendas effectively. 86. Multisector Teams. These concerns reflect a broader sentiment among sector staff about the choice of PRSC task management. PRSCs tend to be managed by PREM staff, even though a large part of policy aspects is dominated by social sectors. And there have been occasional concerns when sector staff felt excluded from the core decisions on the PRSC. For new PRSC teams, what lessons have been learned from working cross-sectorally? In Nicaragua, the PRSC team found that a concerted effort was needed to ensure inclusiveness, open communication, and dialogue to determine linkages. At the same time, clear rules and criteria had to be established to keep the reform program at a reasonable size and cohesion. This involved mundane issues such as making sure policy measures were phrased appropriately or having a common definition of what constituted priority meas~res.'~However, the Bank's internal incentive structure continues to be a challenge for the implementation of multisector PRSCs. IV. HARMONIZATION PREDICTABILITY AND 87. The PRSP represents an opportunity to enhance aid effectiveness. A key element of the new development architecture is the concept of mutual accountability: governments undertake a wide range of policy and institutional reforms critical to improving development outcomes and achieving national objectives on poverty reduction, and in return, donors support government leadership of the development process by reducing aid fragmentation and associated transaction costs, and improving aid predictability. In heavily aid-dependent countries with multiple budget support programs, the emerging architecture envisions harmonized donor budget support programs and processes that are aligned to the government's annual PRSP process and budget and planning cycles, and that provide timely and predictable resource flows. The harmonization and alignment agenda is ambitious and challenging. Full effectiveness is still hampered by different approaches and requirements on the part of donors and governments, the weak links between the PRSP and domestic processes, and the fact that practice is still emerging and subject to change. While experience suggests the Bank is generally moving in the right direction, expectations on the speed and magnitude of the shift to the new aid architecture must be properly managed. 65 PREMKnowledge and Exchange Brief (No. 27491), World Bank, 2004. A. Donor Coordination and Harmonization 88. Over the past few years, the political momentum for harmonization and alignment has grown. At the March 2005 High-Level Forum on Aid Effectiveness in Paris, donors strengthened the commitments to coordination and harmonization made in Rome and M~nterrey.~~Donor coordination and harmonization are not ends in themselves, but the means to enhance aid effectiveness by moving away from the fragmentation, duplication, and high transaction costs that to a large extent still characterize foreign assistance. The Bank has been a leader in this agenda, both at the corporate and country levels, and the PRSC has provided an effective vehicle to coordinate and harmonize with other budget support donors. 1. Coordination and Harmonization Modalities 89. Operational experience during the past five years has produced various models of how the PRSC facilitates Bank coordination and harmonization with other donors: for example, traditional coordination and sharing of analytical work, donor cofinancing, and harmonization of budget support programs. Each approach reflects different country circumstances, the role of the government, and the nature, timing, and role of the PRSC in the country. In some countries, the PRSC provides the framework for harmonization; in others, coordination and harmonization evolve through successive operations. In any given country, the Bank engages in various types of coordination and harmonization activities with both budget support donors and donors focused on project financing. In Benin, for example, the Bank provides budget support jointly with the EC and other donors and works closely with non-budget-support donors, such as USAID, to elaborate on the private sector development strategy, and UNDP, to support the PRSP M&E system. 90. Traditional Coordination. In certain PRSC countries where there are few donors, or where budget support plays a relatively minor role in donor assistance, the Bank and other donors undertake coordination in traditional ways: ensuring adequate exchange of information and identifying areas of comparative advantage to avoid overlap and reduce the likelihood of critical gaps. In Albania, whose relations with the European Union are leading to closer alignment with EC legislation, the PRSC has been kept small, in relation to both overall donor flows and the Bank's lending portfolio. Donor coordination on the PRSC has included joint work with DFID on the PRSP policy monitoring and evaluation program, and with UNDP on the MDGs. In Pakistan, the Bank undertook several joint analytic products with the Asian Development Bank (ADB) and DFID. The Government considers the PRSP and PRSC to be good instruments for increased coordination on policy dialogue and program support. DFID has expressed interest in disbursing budget support against PRSC actions and the US has produced a document on "Shared Objectives" that was prepared in consultation with the Bank and includes many of the PRSC's areas of 66 www.aidharmonisation.com. 67 Program Document.for Pakistan PRSC I (28268-PK), World Bank, August 2004. In Honduras, the design of the PRSC reflected close collaboration with the Inter- American Development Bank (IDB), particularly on the policy benchmarks and the analytic work on the country's fiduciary framework. DFID and the Swedish International Development Agency (SIDA) are cofinancing the IDA technical assistance project that complements the PRSC. The German Financial Cooperation Agency (KfW) and SIDA have also indicated strong interest in parallel financing future PRSC operations. 91. Donor Financing. Development partners view the PRSC as an effective vehicle for supporting the policy dialogue on the objectives and programs of country PRSPs. In Vietnam, development partners parallel financed the PRSC, disbursing commitments against policy measures in the PRSC policy matrix. The governments of Denmark, the Netherlands, UK, and Sweden provided close to $75 million in parallel financing for the first two PRSCs and it is expected that total parallel financing for PRSC 3 could exceed the $100 million committed by IDA, and would include resources from, in addition to those noted above, ADB, Canada, EC, and Japan. Many of these development partners have also played important roles in the policy dialogue and in the preparation and design of the PRSC and, together with others, have provided TA grants to support the PRSC. They have divided the task of engagement in tripartite meetings with the Bank and the government amongst themselves, according to their comparative strengths." KfW has provided parallel financing to PRSCs in Armenia, Nicaragua, and several African countries, disbursing its funds against the Bank's assessment of the fulfillment of prior actions and conditions. 92. Donor Harmonization. In many PRSC programs, including all those in Africa, harmonization around budget support is a key element in the PRSP process and a key objective of the Bank's engagement. Broadly speaking, harmonization entails a concerted effort by budget support donors and the government to reduce transaction costs of aid delivery in terms of time and effort, preparation, and reporting and monitoring, by supporting one broad comprehensive reform program based on the PRSP and by engaging in a consistent policy dialogue. Key characteristics typically include formal, signed memoranda of understanding (MOUs) that lay out the rules for the provision of budget support, including the joint assessment of country performance and disbursement against a common PAF; multiyear donor commitments; joint donor missions; and streamlined reporting requirement^.'^ Such frameworks also aim to facilitate early commitments of aid, the integration of these commitments into the budget formulation process, the front-loading of disbursements within the fiscal year to enhance the predictability of aid flows, and setting out the obligations of countries and their partners in the PRSP process. It is still too early to judge the effectiveness of such arrangements. However, the start-up costs of establishing joint monitoring frameworks have been considerable, and the process often time-consuming.70 Ethiopia and Mozambique, for example, the process of agreeing on a common set of indicators and monitoring modalities took well over a year. - See Table 7 of the Program Documentfor VietnamPRSC 3 (28916-VN), for a detailed description of the areas of involvement of each cofinancier. 69 Good Practice Note on Budget Support Groups and Joint Financing Arrangements, OPCS, World Bank, June 2005. 70 One donor field staff noted the need to "harmonize harmonization," given the number of surveys and studies undertaken to track progress on country-level harmonization. 7' PRSP -Progress in Implementation, PREM, World Bank, September 2004. 2. Approaches to Harmonization 93. The Paris Declaration on Aid Effectiveness pointed to the need to implement common donor arrangements for financing, disbursements, and monitoring and evaluation, and to reduce duplicative missions and diagnostic reviews.72Indeed, as more donors provide aid in the form of budget support, key principles and good practices on harmonization have emerged (see Table 8 for a list of budget support donors in PRSC countries). And while harmonization approaches differ from country to country, donors generally strive to stay within the spirit of these emerging principles and practices. 94. Country Leadership. As much as possible, the government should drive budget support harmonization, spelling out the "rules of the game" in keeping with the PRSP's country-driven approach. Many country teams have suggested that multidonor budget support initiatives are helping governments take leadership of overall development assistance coordination. This requires the establishment of clear mechanisms for coordination, and usually the assignment of a lead agency. In Ghana, Pakistan, Rwanda, Tanzania, Uganda, and Vietnam, one central ministry is responsible for strategy formulation, budget preparation, and external assistance coordination. Honduras is also consolidating development assistance coordination mechanisms. However, genuine country leadership in coordinating donor assistance is often hampered by capacity and institutional constraints. In Nicaragua, coordination between the Bank and the IDB has been a challenge, partly because the IDB's counterpart is the Ministry of Foreign Affairs, while the Bank's is the Ministry of Finance. In Cape Verde, the Ministry of Foreign Affairs takes responsibility for government relations with bilateral donors, while the Ministry of Planning and Finance oversees relations with multilateral donors. Nevertheless, the harmonization process in these countries has moved forward through a partnership framework with the government. In countries where the governments have taken a much greater role in donor harmonization, they have actively taken part in developing their respective partnership framework agreements that stipulate the separate accountabilities of government and budget support donors. 72 Paris Declaration on Aid Effectiveness: Ownership, Harmonization, Alignment, Results and Mutual Accountability, High-Level Forum, March 2005; access at www.aidhamonization.org . Table 8. Harmonized Budget Support in PRSC Countries using a commonPAF; data based oripr&am documents of the most r e c e n G ~ sapproved ~ s as of April 2005. 95. Performance Assessment Framework (PAF). Good practices suggest that the PAF should be country-driven and used by all budget support donors.73Most PAFs reflect an ongoing policy dialogue and derive to varying degrees from negotiated processes between donors and governments. Ethiopia stands out as a good example of where the government led an iterative process of identifying a subset of indicators from its own PRSP policy matrix to establish a common framework for all budget support programs, including the PRSC. In Nicaragua, Uganda, and Vietnam, the PRSC policy matrix serves as the common framework for several budget support donors who disburse their commitments when the Board approves the PRSC.74 In other countries, the PAF was developed before PRSC approval. In Tanzania, when the Bank joined the existing framework, changes to the PAF were made to accommodate the PRSC, such as broadening the coverage to include rural development and environment. In Burkina Faso, there were originally difficulties in integrating the PRSC into the existing MoU of bilateral donors that foresaw only disbursement against outcome indicators (see Box 13), and in Ghana, the first PRSC proceeded without a PAF in place. These initial difficulties were eventually 73 OECD-DAC, Harmonising Donor Practices for Effective Aid Delively, Volume 2: Budget Support, Secior-wide Approaches and Capacity Development in Public Financial Management, 2005; available at www.oecd.org/dac/harmonisin~ractices. 74 In Uganda, these donors include DFID, Ireland Aid, KfW, and the Netherlands. Several other donors, including the EU, DANIDA, Norway, and SIDA, remain closely associated with the process but disburse against a separate set of indicators. overcome and in most countries with multiple budget support programs, donors have moved toward a common PAF. Box 13. Donor Harmonization in Burkina Faso IIn Burkina Faso, the Bank and the AfDB could not formally sign a memorandum of understanding (MOU) with the I multidonor budget support group (Soutien budgetaire corqoint azr cadre strategique de lzrtte contre la pauvrett;, or SBC-CSLP), because the narrow definition of disbursement conditions conflicted with the Bank's PRSC guidelines and the AfDB's programmatic lending operations. However, the Bank worked as a very close associate to the SBC- CSLP, participating in almost all meetings of the group and coordinating the preparation of the PRSCs with SBC- CSLP members and other donors present in Burkina Faso. In particular, the PRSC team regularly exchanged views on the PRSP and program implementation and shared its written conclusions with the donor community; and the Government invited donors to participate in Bank missions. The coordination ensured consistency of supported measures and avoided duplication of Government reporting efforts. The update of the PRSP in 2003 allowed the Government to further improve donor coordination by broadening the evaluation and monitoring framework to achieve buy-in from all donors. In particular, the updated PRSP is accompanied by a Priority Action Plan (PAP), a set of measures and results indicators providing the unique framework to assess progress made in implementing policies and achieving results. This led to a revised MOU in January 2005, centered on the Government's PAP that accompanies the revised PRSP. All donors providing budget support draw their disbursement criteria from the PAP measures and indicators as well as from the IMF's assessment of macroeconomic policies. 96. Unanimity vs. Pluralism. A certain level of consensus among budget support donors on key issues is needed to ensure proper harmonization. Agreement on a common PAF, for example, typically defines a joint budget support group. However, donor groups differ from one another on the flexibility individual donors have to determine specific benchmarks. For example, while the Bank bases disbursement decisions on achievement of specific actions, the EC bases its decision on achievement of outcome targets (see Box X). In Ghana, Uganda, and Vietnam, donors reach the same judgment on performance, based on a common set of indicators in the PAF. In Tanzania, individual donors make separate assessments of how the common assessment of progress translates into financing decisions. This practice conforms to an emerging view among budget support donors that approaches allowing staggered or multitranche approaches by different donors may limit the "herd behavior" that could lead to an all-or-nothing response of disruptive fluctuations in aid flows if performance is deemed less than satisfactory. Such an approach allows donors to disburse a set commitment amount against different triggers, so that satisfactory progress in one policy area could be rewarded independently of progress on other areas. In this view, allowing different donors to respond in different ways to performance spreads risks for the recipient country. On the other hand, it also dilutes any signaling fbnction and prevents the expected reductions in transaction costs of a coordinated donor response. 97. Donor Views on Bank Harmonization. Discussions with development partners often reveal reservations about the Bank's role within budget support groups and its willingness to engage in harmonization efforts. On the one hand, most development partners respect the Bank's technical capacity, understanding of complex developmental issues, and ability to take a lead in many areas of the policy dialogue. On the other hand, the Bank is often viewed as the oversized member of the budget support group, with unmatched influence within the group and disproportionate access to government. Although the PRSC has increased opportunities for joint work within the donor community (in Tanzania, donors point to the Bank's proactive role in involving donors in the PER process; and in Ethiopia, to the Bank's insistence with government that other donors be included in the policy dialogue), donors still express a concern that the Bank can and does proceed on its own when it wants to. The size of PRSC teams can often be overwhelming during missions, with important meetings, discussions, and decision-making compressed in a narrow timeframe. The scope of these missions may complicate logistics and communications of ensuring that all partners are well informed of changing schedules and agendas, and that they reach consensus on key issues before negotiating with the government. Despite the Bank's strong decentralization efforts over the past years, in several countries donors encourage an even stronger field presence of senior sectoral staff who can engage in continuous dialogue with partners and with government and, hence, a reduced need for large missions. 3. Risks and Challenges in Harmonization 98. While harmonization can enhance aid effectiveness, some notable risks and challenges are inherent in current harmonization approaches. Volatility. The issue of volatility increases substantially as more donor resources are provided through budget support. In some countries, levels of budget support receipts have varied greatly between years. Malawi offers a worrying example of unsatisfactory performance resulting in a total disruption of budget support flows. As already noted, the risk of volatility is increased if all donors use the same disbursement indicators and respond in the same way to performance and underperformance. Size of PAF / Number of Indicators. While the PAF is a useful guidepost and management tool for the government's reform program, care must be taken to limit the size of the PAF matrix. Donors must judge how significant indicators are to the overall success of the program. Moreover, while in most cases donors have flexibility to choose disbursement indicators within the PAF, this may lead to a net increase in conditionality for the government. Weight of Donor Influence. Budget support groups range in size from three donors in Cape Verde to 17 in Mozambique. Harmonization has raised various issues in relation to how groups should be managed and how decision-making should be undertaken. Should donors that provide relatively small volumes of budget support (in relation to their overall aid program and to total budget support flows) have an equal say in decision-making with donors that provide relatively larger shares of aid through budget support? What should be the influence of donors with limited technical expertise in any given policy area, or of non-budget-support donors whose programs may be affected by the decisions of budget support groups? In Tanzania, it has been suggested that entry fees could be set at $50 million in annual commitments, or 70 percent of the donor's total as~istance.~~Such an approach can reduce transaction costs; but the approach also risks deterring some donors, particularly small ones and those that are not "likeminded" from continuing their venture into budget support. Donor Collusion. To avoid increased donor dominance through coordination, governments may prefer maintaining bilateral relations alongside those they have with the joint donor group. In this regard, the harmonization process in some cases has brought about contradictory reactions from governments. On the one hand, partner governments welcome that harmonized donors speak with "one voice"; on the other, they are weary of donors taking a common position on difficult or sensitive issues, as in Pakistan. In Ethiopia, the government reluctantly agreed to accept bilateral donors as 75PRSP Annual Progress Reports and Joint StaffAssessrnents - A Review of Progress, ODI, September 2004 part of the budget support policy dialogue after the World Bank insisted on a harmonized approach to budget support. Some governments have requested "clear rules of the game." Without such rules, although governments prefer budget support as a modality for financing PRSPs and lowering transaction costs, they may be willing to forego those benefits if they perceived harmonization as donor collusion. B. Improving Predictability 99. Improving the predictability of resource flows is an explicit objective of many PRSC's and budget support programs in general. Predictability is of particular concern in aid-dependent countries, where budget support represents a large share of public expenditure. The concern is both of a short and medium term nature since: (a) more donors are channeling aid through budget support, and the reliability of commitments and the timing of disbursements are critical for budget execution; and (b) the PRSP process has underlined the need for governments to improve their medium-term budgeting and planning and reduce resource volatility so they can carry out the country's development program. Increasingly, the donor community is also framing the challenge of resource predictability in the long-term context of achieving MDG targets, in line with the Monterrey Consensus. This section seeks to assess the extent to which PRSCs are improving resource predictability. It looks at how the programmatic design of the PRSC helps to reconcile the need for predictability with strong country performance, and provides suggestions on how to enhance predictability through improved transparency and better alignment with domestic planning and budgeting processes and timetables, including the PRSP Annual Progress Reports (APRs). 1. Programmatic Alignment with the PRSP 100. The programmatic approach of the PRSC reflects an evolution in the design of policy- based lending. Adjustment lending operations in the 1980s and 1990s were typically multitranche operations with a short-term horizon based on ex ante conditionality that often proved insufficiently flexible to reflect changing circumstances. Traditional conditionality was also viewed as being insufficiently tailored to specific country circumstances and overlooking important considerations of political economy, government capacity, and commitment. As a result, governments often were not able or willing to implement policy conditionality, leading to a "tap-on, tap-off' pattern of disbursement and high volatility in resource flows.76Alternatively, to disburse funds, donors frequently waived unmet conditions and proceeded with disbursement, undermining reform implernentati~n.~~The PRSC was designed to reflect country priorities and to be flexible, with a step-by-step medium-term approach based on completed (as opposed to promised) actions. These design features help reconcile the tension between the need to stabilize public expenditures through resource predictability and the need to demonstrate that the policy environment allows for an effective use of aid resources. 101. Medium-Term Perspective. The programmatic design of PRSCs is based on a medium- term program of support. Ideally, the CAS overlaps with the PRSP timeframe and sets out the three or four annual PRSC operations within that timeframe. The CAS assigns an indicative 76 While there is also considerable unpredictability in investment or project lending, it does not directly affect government planning and budgeting since disbursements typically reflects physical implementation and not recurrent expenditure. Volatile project disbursement, therefore, does not necessarily imply a funding gap that needs to be filled. See Mick Foster, Improving the Medium to Long-Term Predictability o f Aid, January 2005. 77 Adjerstment Lending Retrospective, OPCS, World Bank, June 3001. commitment amount for each operation in the program, providing an envelope of budget support for the PRSP program and a critical input to the government's medium-term expenditure planning. There are no formal mechanisms to provide a rolling commitment scenario (i.e., a PRSC approved and disbursed in FY05 indicates the budget support envelope for FY06-08, and the next PRSC in FY06 updates the envelope for FY07-09), since the CAS does not indicate commitments beyond the timeframe it covers. In the context of the government's and the IMF's annual medium-term macroeconomic and fiscal projections exercise, the Bank's informal indications about multiyear financing commitments are naturally subject to the country performance allocation and available IDA resources. Over time, it may be useful to consider formalizing commitment projections on a rolling basis that are in line with the programmatic design of the PRSC and the need to enhance governments' medium-term budget planning. 102. Triggers and Flexibility. Another key element of the programmatic design is the use of triggers to assess country performance and to serve as the basis for disbursement decisions. While each PRSC is based on prior actions, triggers allow the Bank to determine whether sufficient progress has been made to move to the next operation. Triggers help shape the preparation of the subsequent PRSC, and they typically form the basis of its prior actions. However, triggers entail considerable flexibility: the prior actions of the subsequent PRSC need not be identical to the triggers; they can be adapted to changing circumstances or modified for better clarity, specificity, or measurability (see Box 14). The Bank's judgment could consider to what extent new evidence or circumstances render a key measure or indicator less relevant than originally anticipated. The flexibility inherent in the programmatic design of the PRSC comes not from defining vague or easily met triggers, or from allowing continued financing despite poor perfoimance, but rather from identifying key reform measures that the government is committed to, and then measuring progress against them, using reasoned judgments to allow for disciplined adaptation. Box 14. Different Approaches to Triggers I I Triggers have been designed and used in various ways. Typically, once triggers are fulfilled and the country moves on to the subsequent operation, prior actions are drawn from the triggers themselves. Prior actions need not be identical to the triggers. For example, the Burkina Faso PRSC 2 maintained the triggers set out in PRSC 1 as prior actions, except for two actions that were rephrased to better clarify government policy. In some PRSC countries not all the triggers were defined with geat precision, including broad definitions such as "satisfactory progress in implementation of procurement reforms," "provide sufficient budgetary resources for the provision of subsidized health services," or "implement an adequate institutional capacity building plan". While this type of progress indicator leaves the ~ a nwith flexibility to determine how to make a "bottom line" assessment, it k also increases the degree of subjectivity in assessing performance and could potentially lead to misunderstandings. Eventually, the Bank and the government must agree on specific performance indicators- that is, precisely worded prior actions-as they did in Uganda for nearly 80 percent of the triggers in PRSC 2,67 percent of PRSC 3 triggers, and 57 percent of PRSC 4 triggers. In Albania, 40 percent of the triggers were modified to better specify prior actions. The approach taken in Vietnam, however, was quite different. Although triggers were met, the Bank and the Vietnamese government agreed to an almost new set of prior actions in PRSC 2 and PRSC 3 that had very little resemblance to the original triggers. 103. Graduated Response. The PRSC emphasis on predictable budget support clearly does not imply an entitlement to aid flows irrespective of country performance. The PRSC's programmatic approach allows for a graduated response in case of uneven country performance. Proceeding with Full Amount. If overall progress is judged to be adequate even if certain triggers were not fully met as expected-perhaps because delays in some areas were compensated by advances in others-the prior actions of the program can be adapted and the next operation can proceed with a full amount. To date, most triggers have been fully met; only five percent of triggers were not met at the time the program document was completed. In those cases where actual implementation deviated from expectations, the Bank respond in a graduated manner-it determined that substantial progress in other areas demonstrated continued government commitment and justified moving toward the next operation. Many of the unmet triggers were simply delayed; they did not reflect a lack of government commitment. Postponing Operations. If triggers were not met because of backtracking or failure to implement, the Bank gave serious consideration to delaying the operation. In Tanzania, the Government preferred to postpone the subsequent operation by about two months to fully implement all prior actions that, for technical reasons, had taken somewhat longer than originally anticipated. The decision was made in lieu of the Bank's offer to continue the program as scheduled but with a reduction in the commitment amount. Delays have also marked the PRSC programs in Nepal and Sri Lanka, two countries with an uneven track In both cases, delays were caused by inadequate performance (see Box 15 regarding the Nepal PRSC program). Reducing Commitment Amounts. Another option has been to reduce the amount of disbursement. This graduated modulation of PRSC commitments avoids the inflexible "tap on, tap off" pattern of disbursement in traditional policy-based lending, with the entire tranche commitment held up or cancelled when conditionality was not met. To date, the Bank has not had to reduce the commitment amount for a subsequent PRSC operation. For Ethiopia PRSC 1, the original commitment was somewhat lower than originally intended prior to Board approval given the lack of progress in one key policy area. Box 15. Nepal PRSC Program in "Wait and See Mode" 7 The World Bank's Board approved the first Nepal PRSC in November 2003. The decision was based on an impressive array of reforms camed out amidst political uncertainty and conflict. A core group of reformers at a senior technocratic level were driving the reform agenda on the basis of a comprehensive and participatory PRSP process. Regular dialogue on the key actions required for PRSC I1 revealed that despite an overall slowing of the pace of reforms, certain bold measures such as fill adjustment of petroleum prices, improved trade facilitation, and accelerated devolution of schools and health clinics had taken place. A January 2005 mission identified three outstanding actions that remained in order for the Bank to proceed with PRSC 11. On February 1, 2005, the King dismissed the coalition government, appointed himself as head of the new Cabinet, and suspended a number of basic constitutional rights. The new Government announced a 21-point economic program, largely consistent with the PRSP; but it is too early to tell how effectively reforms will be implemented, and the initial small steps have been somewhat mixed. The Government's ability to implement reforms effectively and improve public service delivery will also depend on whether the security situation on the ground improves. In light of these uncertainties, and the fact that the three prior actions are still outstanding, the Bank has decided not to go ahead with a PRSC I1 in FY05, but to continue with dialogue until the repercussions of this political change on reform implementation become clearer. The Bank will coordinate and consult closely with other development partners as part of this process. 104. Interrupting Programs. PRSC programs have been suspended in a few cases where inadequate progress in improving the policy environment did not warrant moving to the next operation. Guyana is the only country where the PRSC program has been disrupted. A deadlocked parliament, political and ethnic tensions, and institutional weaknesses have made it '' The Sri Lanka PRSC program may be modified to reflect the revised PRSP currently under draft by the new government. difficult for the government to implement triggers for PRSC 2 in a timely fashion. The ICR for the first operation suggests that the original schedule of one PRSC per year was too optimi~tic.~~ Eventually, the country team decided to discontinue the PRSC as a series of annual policy loans and to prepare a new set of policy-based operations that are more widely spaced, taking into account lessons learned and country circumstances. 2. TheRecord on Predictability 105. PRSCs have contributed to improved predictability of resource flows in PRSC countries, as reflected in: (a) the extent to which the Bank has delivered on its medium-term commitments outlined at the outset of the PRSC program; (b) the regularity and volume of PRSC disbursements; and (c) the regularity of PRSC approvals.s0 The seven countries that have moved to subsequent PRSC operations have succeeded in demonstrating improved predictability on all counts. Table 9 compares commitments to actual disbursements for countries that have moved past their second PRSC. The data show strong correlation between commitments made at the start of the program with disbursement upon Board approval. Table 9. Commitments and Actual Disbursements (%) Original Commitment (OC) Disbursement (D) Countrv - ..~ PRSC I PRSC2 PRSC3 PRSC4 PRSC I PRSC2 PRSC3 PRSC4 D/OC - - ~ ~~ Albania 20 10 10 20 18 10 120 Burkina Faso 45 35 40 60 45 35 50 60 106 Ethiopia 150 150 120 130 83 Ghana 125 125 125 125 100 Tanzania 150 150 132 150 94 Uganda 150 150 150 150 150 150 150 150 100 ~ietnam 250 100 100 250 100 100 100 Source: SAP Business Warehouse Note: Original commitinents are based on figures provided in the program document for PRSC 1or the CAS. 106. Increased Volumes. The record also shows that flows from policy-based operations were disbursed with much greater regularity and predictability in PRSC countries compared to the years prior to the PRSC program. This is particularly evident for Albania, Burkina Faso, and Vietnam, which utilized policy-based lending sporadically in the past. Another trend in disbursement shows an increase in the volume of DPL resources, even in countries where the PRSC is a small share of the country's total portfolio. While the sample is still relatively small and the time period relatively short, the trends so far fulfill the expectation that the PRSC's programmatic, annual cycle helps to reduce volatility and improve the regularity of annual aid flows for countries demonstrating consistent progress in reform implementation. 107. Aggregate Disbursements. Aggregate disbursements (DPL plus investment) have increased sharply in most PRSC countries (see Figure 14). The finding parallels the OED assessment that PRSPs in general have provided a basis for higher, more stable flows." This particular trend, however, is dependent on the structure of the country portfolio. In Vietnam, 79 The ICR for Guyana PRSC 1 nevertheless rated the operation "satisfactory." One other indicator of predictability is the synchronization of disbursement with the government's budget cycle, in the sense that resource flows should be timed as much as possible to when the government needs them most, typically at the beginning of the fiscal year. K 1The Po~tertyReduction Struteg?.Initiative: An Independent Evaluation, World Bank, OED, 2004, pg. 38. while PRSCs provide a more predictable flow of quick-disbursing resources, the higher aggregate disbursement levels are dnven by new investment projects. Figure 14. Increased Predictability and Volume of Bank Disbursements to PRSC Countries Burkina Faso Uganda I OAdjustmnt 1 Investmnt Investment 1 FY95 FY96 FY97 FY98 FY99 FY00 FYOl FY02 FY03 FY04 FY95 FY96 FY97 FY98 FY99 FYOO FYOl FY02 FY03 FY04 I Vietnam Tanzania I I I I , I OAd~llstrnPnt I I@Adjustmnt HInvestment 1 332 Investmnt I FY9S FY96 FYY7 FY98 FY99 N O 0 FYOl FY02 FY03 FY04 FY95 FY96 FY97 FY98 FYY9 FYOO FYOl FYO2 FY03 FYD4 Source: SAP Business Warehouse Note: Dotted line indicates the beginning of PRSC disbursements 108. Regularity in Timing. PRSCs have been broadly successful in maintaining an annual cycle of operations (see Table 10). This design has proven to be beneficial in several respects. As long as performance is judged sufficiently satisfactory, the government could expect improved regularity and predictability of resource flows. Among the seven PRSC programs that have moved past the first operation since FYO1, the average lag time between Board approvals was 379 days, or approximately one year." In contrast, during the same period, the average lag time between tranche disbursements of multitranche policy-based lending was 441 days, or 14 months. More than 26 percent of these multitranche operations also required waivers or tranche restructuring. While actual lag time varies and is influenced by a number of factors, the annual cycle provides greater focus and emphasis on regularity, and some constraint on the amount of time both donors and government have to make decisions and to take actions. 82 The data do not include the time lag between Vietnam PRSC 1 and PRSC 2, since the fonner was a two tranche operation, with the second tranche disbursed 200 days before the Board approval of the next operation. Moreover, actual lag time varies, from 10 months between Burkina Faso's PRSC 3 and PRSC 4 to nearly 16 months between Albania's second and third PRSCs. Source: SAP Business Warehouse Note: asterisks indicate use of two-tranche operations; the first "x" for each country marks the approval of the first PRSC operation. 3. Design Options to Further Improve Predictability 109. Although the Bank's PRSC disbursements have been quite predictable, the Bank can exercise a significant degree of discretion in modulating disbursements of subsequent PRSCs as a response to under-performance. Similarly, an element of discretion is also introduced in some PRSC programs when triggers are vaguely defined, (e.g., "satisfactory progress with implementation"). While there may be an advantage to the flexibility it provides Bank teams when addressing concerns over underperformance, undue discretion also carries the risk of uncertainty. Transparent decision rules and clearly specified triggers can help set out clear donor and recipient expectations and clarify their mutual accountability. Implementation experience by the Bank and other budget support donors has produced a variety of innovations that addresses the concerns over excessive discretion. 110. Bank Experience with Formal Graduated Approaches. Some PRSC countries have incorporated a more formalized response to underperformance. In Benin, for example, each of the three core sectors amounts to one-third of the total commitment, so that if implementation were off track in one sector, commitments would be reduced accordingly; however, the definition of "satisfactory implementation" is not clearly articulated. The Vietnam program uses a more specified approach: attainment of all triggers is considered a strong indication that Vietnam is ready to move to high-case lending, as established in the CAS; progress on most triggers and no backtracking on any of them leads in principle to the preparation of the next PRSC operation in base-case lending; and backtracking could lead to delaying the next operation, which in practice is equivalent to moving to the CAS low case. In terms of explicit tranching, all but three PRSC operations have been single-tranche annual operations: the first PRSC operations in Vietnam and Nicaragua were two-tranche operations, a design choice that reflected pragmatic considerations such as IDA allocations and parliamentary approval, not the need for greater transparency or predictability. The Uganda PRSC 2 had a floating tranche whose conditions were particularly sensitive but were satisfied by effectiveness; thus it was disbursed along with the main tranche. 111. Other Design Options to Improve Predictability. Given the PRSC's country selectivity, an argument has been made to move toward longer-term predictable commitments through an aid allocation model similar to IDA'S.'^ This approach proposes that commitment levels be determined by an assessment of public expenditure needs and broad country performance ratings, such as the CPIA. There would be less need for detailed conditionality at the operational level or for any specific actions or indicators that may be binding-as long as there is sufficient ownership and the country broadly pursues its overall development program. In fact, some donors have argued in favor of moving away from explicit conditionality on specific policy areas and confining conditionality to fiduciary concerns and overall adequate budget management. For Bank-supported programs, it is not clear that such a clear distinction between policy and fiduciary conditionality can and should be made: typically PRSCs address fiduciary issues as part of the development program. Nonetheless, for well-performing countries that have benefited from more mature multiyear PRSC programs, the concept of a broader assessment of overall performance based on fewer and perhaps less detailed progress indicators at the operational level and a broader assessment of overall country performance would be a natural evolution of emerging good practice lessons. 112. Fzked-Variable Tranching. The European Commission and several bilateral donors have incorporated a "graduated ?.pproach" with a fixed and variable tranches. Under this approach, the fixed tranche is designed to provide relatively predictable financing, or a "base flow" of budget support. It is broadly linked to input, output, or outcome indicators that are fundamental to program success and have a very high likelihood of being met, such as sound macroeconomic management (for which judgment is based on the IMF's assessment). The variable tranche is explicitly linked to performance. It may be linked mechanically to performance against specific indicators, whether formulated as specific measures, outputs, or in the case of the EC, on outcomes; and it may be disbursed partially. The Bank's Regional Strategy for Africa encourages experimentation with approaches such as variable tran~hing.~~ However, some caution is warranted in exploring this approach in the World Bank context. In the first place, variable tranching would re-introduce the traditional ex-ante conditionality of multiple tranching that must be specified and met without flexibility. Because the programmatic approach is based on a series of annual operations, variable tranching in PRSCs would imply tranches that are explicitly tied to deadlines and precise measures rather than leaving the freedom to modulate, postpone, or withhold lending based on a new performance assessments. Under the Bank's operational policy, even a fixed tranche would still be conditional on an adequate macroeconomic policy framework and overall satisfactory implementation. Moreover, the difference in the size of the fixed and variable tranches would need to be significant to create incentives for governments to perform; otherwise, a large fixed tranche with minimal conditions might be viewed as an entitlement. - X3 M. Foster, Improving the Medium to Long-Term Predictability ofAid, January 2005. 84 Strategic Framework for IDA'S Assistance to Africa: The Emerging Partnership Model, World Bank, Africa Region, June 25,2003. 4. Timing of Disbursements and Aligning with Country Processes 113. Beyond design options, another way of enhancing the PRSC's predictability is to strengthen the alignment of the PRSC cycle (preparation, approval, and disbursement) to the country's planning, review, and budget calendar and processes. While PRSCs have generally succeeded in providing stable annual disbursements, the aim of improving alignment with country processes is to ultimately disburse at, or close to, the beginning of the fiscal year, thus providing the government the maximum level of confidence and certainty it needs to execute the annual budget and commence program implementati~n.~~ Achieving such alignment requires that donors time their assessments of country performance in a way that allows them to confirm commitment amounts when the government is setting the resource ceilings for budget formulation. This entails a proper sequencing of a number of processes: sectoral and cross- sectoral program reviews, macroeconomic assessments, budget preparation, donor confirmation of commitments, donor disbursement, and budget reporting (see Table 11 for an illustration). Delays in any one process could disrupt sequencing. 114. PRSP Annual Progress Report (APR). An emerging consensus is that the PRSP APR should play a central role in this process. The OED evaluation suggests that APRs have played a constructive role in reporting on progress and in adjusting strategies. This role could be strengthened, however, by more closely linking them with the budget and other domestic processes and ensuring that their timing is country-driven.R6 Essentially, the framework envisions a steady-state scenario that gives the APR a critical role in government planning, budgeting, and review functions, as well as providing donors much, if not all, the reporting requirements on reform implementation. To fulfill this vision, the APR must be both forward- and backward-looking: it needs to review the past year's progress across a range of sectoral and cross-cutting areas, and also to inform the policy choices moving forward and articulate the government's priorities and budget needs for the next year. Such a comprehensive review would necessarily be linked to sectoral program reviews, PERs, macroeconomic assessments, and budget projections for the current and up-coming years. In this regard, the APR could contain the input, output, and outcome indicators on which donors and governments have agreed for assessing country performance. The forward-looking perspective of the APR would entail the laying out of future policy choices and articulating the government's priorities and budget needs for the next year. The framework would also benefit greatly if results and recommendations of key analytic and diagnostic work were also incorporated. Such a comprehensive review would provide donors with the basis for making firm commitments of support for well-costed programs that are drawn from the PRSP." 85 There are, however, varying degrees to which the timing of disbursements is a priority for some countries. The governments of Uganda, Tanzania, and Ethiopia have noted that the firmness of the commitment is more important than the actual timing. 86 The Povero?Reduction Strategy Initiative, An Independent Evaluation of the World Bank's Support Through 2003, Operations Evaluation Department, World Bank, 2004. 87 Increasing the Predictability ofFlows of Donor Budget Support, IMF, January 2004. Table 11. Example of Synchronized Cycle for Budget, PRSP Review, and Budget Support Month Government Donors t-9 Reporting on budget implementation and Review of performance under existing donor outcomes programs and in implementing the PRSP in the Draft I'RSP annual review of previous previous year fiscal year Discussion of key performance indicators for Discussion of progress and policy PRSP implementation in next year implications with domestic stakeholders Indicative budget support commitments in the and with donors context of preparations for national budget p-y t-7 Preliminary update of macroframework Donor formulation of support programs for next and MTEF fiscal year, based on government policy Preparation of government policies for priorities next fiscal year t-6 Budget information to sector ministries t-4 Finalization of government policy Donor confirmation of budget support amounts priorities and preparation of budget PRGF mid-term review completed It-2 Submission of budget to Parliament PRGF discussions of program for next fiscal year based on finalized budget and updated MTEF t 1 Approval of budget Approval of PRGF annual program 'A I Approval of PRSC Sourc.r: Increasing the Predictabilih of Flows of Donor Budget Support, IMF, February 2004 115. Challenges. Obviously, in order for this alignment to be realized, the government must first ensure strong links between the annual PRSP process and the budget and planning cycles. These links are still not well developed and many governments have expressed uncertainty over the value and the intended purpose of the APR, often viewing it as an added burden.8xA recent SPA survey revealed that among 15 countries currently receiving budget support, only Rwanda had complete alignment between the APR and the budget cycle and six countries had partial alignment. The report also noted that the APR in most countries was still a pilot exercise or a limited progress report with shortcoming^.^^ It is important to note that, just like the PRSP itself, the APR should be seen as an instrument of domestic accountability. It should primarily serve domestic purposes and hence, each country will deternine for itself how it will cany out its annual reporting, and how it wishes to use the APR. Guidance has been perhaps been purposely vague specifically to give countries flexibility to determine the content of the reviews.90 However, as a result of these weaknesses, many donors also remain uncertain about the operational usefulness of the information contained in the APR.9' 116. Role of Donors. While the role of the APR is an area that is clearly the responsibility of the government, donors may provide assistance to facilitate domestic PRSP alignment. As is the case in other areas, the Bank's engagement through the PRSC could help remedy some of the weaknesses in PRSPs. This is particularly so in countries where the Bank and other budget support donors have coordinated their programs behind the policy priorities of the PRSP and where a substantial part of dialogue with the government has focused on improved links between 88 Strategic Partnership with Africa, Budget Srrpport Workshop, Kigali, Ruunda, June 2004. 89 Survey of the Alignment of Budget Support und Balance of Pu>jmentsSupport with National PRS Processes, SPA, February 2005. 90 A Concept ,Vote on Aligning Donor Budget Support ~ Y t hthe PRSP Process, IMF and World Bank, February 2003. 91 Strategic Partnership with Africa, Budget Support Workshop, Kigali, Rwandu, June 2004. policy, budget, and results monitoring and reporting. Good examples include Ghana, Ethiopia, Mozambiqlle, Tan~ania,and Ilganda. In Ethiopia, improved alignment with domestic processes was instrumental in the government's effort to design one monitoring and evaluation system for the PRSP, the PRSC, and other budget support programs. In Mozambique, budget support donors have encouraged the use of an existing annual progress report to Parliament as the main instrument for assessing PRSP implementation and as the basis for an annual joint donor-government assessment on budget support. The move places considerable importance on improving the quality and coverage of the report to Parliament. In Albania, one of the main components of the PRSC is to improve PRSP monitoring and evaluation through a variety of actions, such as establishing a baseline for regional distribution of health facilities and sufficiently staffing and funding the Institute of Statistics. Efforts to strengthen PRSP monitoring has gone the farthest in Tanzania, where the PRSC focuses mainly on institutionalizing the monitoring and evaluation of PRSP targets. 117. Disbursement Delays. Aligning domestic processes and calendars has implications on when to assess whether prior actions have been met. Although the aim of disbursing at the beginning of the fiscal year is important for many countries, PRSC disbursements often come later as a result of insufficient alignment with the budget cycle. 118. Improving TimelyDisbursements. In order to align PRSCs more closely with the budget cycle, PRSC teams have proposed to modify PRSC timetables to align with the budget cycle. In Burkina Faso, the Bank accelerated the preparation and approval of PRSC 4 to follow the previous operation after only eight months. The aim is to eventually "catch up" with the government's fiscal calendar, with PRSCs approved and disbursed in April-May. In Mozambique, the PRSC currently disburses half way through the year and the plan is to approve three PRSCs within six months of each other in order to align with the joint review and to disburse at the beginning of the calendar year. However, for several PRSC countries, such as Ethiopia, Tanzania, and Uganda, the timing of resource flows is less important than the early confirmation of disbursement. In Ethiopia, the government indicated that as long as they have reasonable certainty regarding the donor's commitment of budget support, actual disbursements can take place up to half-way through the fiscal year. 119. The stocktaking paper concludes that the PRSC has emerged as a robust and effective lending instrument that has been successfully introduced in almost all Regions of the Bank and broadly delivered on its promises. It finds that the PRSC is an appropriately flexible approach that is most suited to countries with siror~gcommitment and sufficient capacity to undertake a medium-term reform program. Although results vary from country to country, the PRSC has reinforced country ownership by (a) building upon the objectives, priorities, and measures set out in the government's PRSP with strong analytic foundations and sustained policy dialogue; (b) providing more predictable medium-term budget resources based on reform implementation and accompanied by increased attention to core public sector functions; (c) aligning with the annual PRSP process to better streamline donor performance reviews, assess implementation challenges, and determine future direction; and (d) reducing transaction costs by providing a framework for donor harmonization around budget support. Certain features of the PRSC have emerged as strong points, in particular its combination of flexibility and predictability in responding to uncertainty and uneven performance, and its reliance on analytic and diagnostic work. As PRSC programs become more mature and are extended to more countries, emerging issues include the challenge of closer donor harmonization and alignment with country processes and building a strong results framework to link implementation with development outcomes. 4 ~ p m h o nt ~ and Country Services r ~ q 1898HSf&,W Washington, DC 20433