THEWORLD BANK 45408 Decentralization in Client Countries An Evaluation of World Bank Support, 1990­2007 IEG PUBLICATIONS 2006 Annual Report on Operations Evaluation Annual Review of Development Effectiveness 2006: Getting Results Addressing the Challenges of Globalization: An Independent Evaluation of the World Bank's Approach to Global Programs Assessing World Bank Support for Trade, 1987­2004: An IEG Evaluation Books, Buildings, and Learning Outcomes: An Impact Evaluation of World Bank Support to Basic Education in Ghana Brazil: Forging a Strategic Partnership for Results--An OED Evaluation of World Bank Assistance Bridging Troubled Waters: Assessing the World Bank Water Resources Strategy Capacity Building in Africa: An OED Evaluation of World Bank Support China: An Evaluation of World Bank Assistance The CGIAR at 31: An Independent Meta-Evaluation of the Consultative Group on International Agricultural Research Committing to Results: Improving the Effectiveness of HIV/AIDS Assistance--An OED Evaluation of the World Bank's Assistance for HIV/AIDS Control Country Assistance Evaluation Retrospective: OED Self-Evaluation Debt Relief for the Poorest: An Evaluation Update of the HIPC Initiative A Decade of Action in Transport: An Evaluation of World Bank Assistance to the Transport Sector, 1995­2005 The Development Potential of Regional Programs: An Evaluation of World Bank Support of Multicountry Operations Development Results in Middle-Income Countries: An Evaluation of the World Bank's Support Economies in Transition: An OED Evaluation of World Bank Assistance Engaging with Fragile States: An IEG Review of World Bank Support to Low-Income Countries Under Stress The Effectiveness of World Bank Support for Community-Based and ­Driven Development: An OED Evaluation Evaluating a Decade of World Bank Gender Policy: 1990­99 Evaluation of World Bank Assistance to Pacific Member Countries, 1992­2002 Extractive Industries and Sustainable Development: An Evaluation of World Bank Group Experience Financial Sector Assessment Program: IEG Review of the Joint World Bank and IMF Initiative From Schooling Access to Learning Outcomes: An Unfinished Agenda--An Evaluation of World Bank Support to Primary Education Hazards of Nature, Risks to Development: An IEG Evaluation of World Bank Assistance for Natural Disasters How to Build M&E Systems to Support Better Government IEG Review of World Bank Assistance for Financial Sector Reform Improving Investment Climates: An Evaluation of World Bank Group Assistance Improving the Lives of the Poor Through Investment in Cities Improving the World Bank's Development Assistance: What Does Evaluation Show? Maintaining Momentum to 2015? An Impact Evaluation of Interventions to Improve Maternal and Child Health and Nutrition Outcomes in Bangladesh New Renewable Energy: A Review of the World Bank's Assistance Pakistan: An Evaluation of the World Bank's Assistance Pension Reform and the Development of Pension Systems: An Evaluation of World Bank Assistance Poland Country Assistance Review: Partnership in a Transition Economy The Poverty Reduction Strategy Initiative: An Independent Evaluation of the World Bank's Support Through 2003 The Poverty Reduction Strategy Initiative: Findings from 10 Country Case Studies of World Bank and IMF Support Power for Development: A Review of the World Bank Group's Experience with Private Participation in the Electricity Sector Putting Social Development to Work for the Poor: An OED Review of World Bank Activities Small States: Making the Most of Development Assistance--A Synthesis of World Bank Findings Social Funds: Assessing Effectiveness Sourcebook for Evaluating Global and Regional Partnership Programs Water Management in Agriculture: Ten Years of World Bank Assistance, 1994­2004 World Bank Assistance to the Financial Sector: A Synthesis of IEG Evaluations The World Bank in Turkey: 1993­2004--An IEG Country Assistance Evaluation World Bank Lending for Lines of Credit: An IEG Evaluation W O R L D B A N K I N D E P E N D E N T E V A L U A T I O N G R O U P Decentralization in Client Countries An Evaluation of World Bank Support, 1990­2007 2008 The World Bank http://www.worldbank.org/ieg Washington, D.C. ©2008 The International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington, DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org E-mail: feedback@worldbank.org All rights reserved 1 2 3 4 5 11 10 09 08 This volume is a product of the staff of the International Bank for Reconstruction and Development / The World Bank. The findings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgement on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development / The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. Cover: Village meeting in Ilha Ibo, Mozambique. ©Atlantide Phototravel/Corbis; reproduced by permission. ISBN-13: 978-0-8213-7635-5 e-ISBN-13: 978-0-8213-7636-2 DOI: 10.1596/978-0-8213-7635-5 Library of Congress Cataloging-in-Publication data have been applied for. World Bank InfoShop Independent Evaluation Group E-mail: pic@worldbank.org Knowledge Programs and Evaluation Capacity Telephone: 202-458-5454 Development (IEGKE) Facsimile: 202-522-1500 E-mail: eline@worldbank.org Telephone: 202-458-4497 Facsimile: 202-522-3125 Printed on Recycled Paper Contents v Abbreviations vii Acknowledgments ix Foreword xi Glossary xiii Executive Summary xix Management Response xxv Chairperson's Summary: Committee on Development Effectiveness (CODE) 1 1 Background and Objectives 3 Literature Review: Arguments for and against Decentralization and Findings from Empirical Studies 5 Decentralization: Everyone Is Doing It 6 Overview of Bank Support for Decentralization 8 Evaluation of Bank Support to Decentralization 13 2 The Quality of Bank Support 15 The Quality of Bank Analytical Work on Decentralization 18 Coherence of Lending with Country Priorities and Bank Analysis 20 Internal Consistency and Results Orientation 23 Collaboration among Development Partners 27 3 Bank Support for Decentralization Frameworks and Subnational Government Capacity 30 Strengthening Intergovernmental Relations 36 Administrative Capacity and Accountability 40 Overall Findings and Lessons 43 4 Bank Support for Decentralization in Education Services: A Sector Example 46 Improving Intergovernmental Fiscal Frameworks in the Education Sector 47 Enhanced Administrative Capacity of Local Governments 49 Enhanced Accountability of Local Governments/Schools to Citizens 51 Conclusion 53 5 Findings and Recommendations 55 Findings 59 Recommendations 61 Appendixes 63 A: Treatment of Decentralization in Bank Strategies and Policies 71 B: Methodology i i i D E C E N T R A L I Z AT I O N I N C L I E N T C O U N T R I E S These appendixes are available online at http://www.worldbank.org/ieg/decentralization/download.html C: Bank Programs and Projects Reviewed D: Stakeholders Interviewed E: ESW Reviewed for Evaluation 75 Endnotes 81 Bibliography Boxes 4 1.1 What Is Decentralization? 6 1.2 Identifying World Bank Support for Decentralization Poses Challenges 7 1.3 Bank Strategies Reflect Multiple Perspectives on Decentralization 17 2.1 Good Practice ESW in Bolivia 21 2.2 An Incomplete Understanding Reduces Effectiveness of Support 22 2.3 CDD and Local Governments in the Philippines 25 2.4 Development Partner Views on World Bank Decentralization Performance 32 3.1 Good Practice in Russia 33 3.2 Strengthening the Role of Local Governments 36 3.3 Analytic Support to South Africa 38 3.4 Appointment of Chief Accounting Officers in Uganda 47 4.1 Aligning Fiscal Transfers with Local Sectoral Priorities 48 4.2 Gender and Capacity Building at the Local Level Figures xiv ES.1 Evaluation Framework for Assessing Results of Bank Support for Decentralization 9 1.1 Evaluation Framework for Assessing Results of Bank Support for Decentralization 11 1.2 Bank Support for Decentralization to 20 Countries, Fiscal 1990­2007 34 3.1 Distribution of State and Local Revenue Sources 35 3.2 Trends in Fiscal Deficits in Selected Indian States Tables 10 1.1 Typology of Countries Where Bank Support for Decentralization Was Reviewed in Depth 16 2.1 Timeliness of ESW in 20 Sample Countries 30 3.1 Assessing Results of Bank Support for Strengthened Intergovernmental Relations 37 3.2 Assessing Results for Enhanced Administrative Capacity and Accountability 41 3.3 Contribution of Bank Support for Decentralization Frameworks 46 4.1 Assessing Results in the Education Sector 49 4.2 School Autonomy and Accountability in Three Pilot Regions in Russia Supported by the Bank 56 5.1 Support for Decentralization Needs to Be Country Specific 57 5.2 Better Quality of Bank Support Associated with Better Results i v Abbreviations CAE Country Assistance Evaluation CAO Chief Accounting Officer CAS Country Assistance Strategy CDD Community-driven development DAC Development Assistance Committee (OECD) DPL Development Policy Loan ESW Economic and sector work GFI Government Financial Intermediary HNP Health, nutrition, and population IADB Inter-American Development Bank ICR Implementation Completion and Results Report IEG Independent Evaluation Group IMF International Monetary Fund LGU Local government unit MDF Municipal Development Fund M&E Monitoring and evaluation OECD Organisation for Economic Co-operation and Development PFM Public financial management PREM Poverty Reduction and Economic Management (Network) PSAC Programmatic Structural Adjustment Credit UNCDF United Nations Capital Development Fund USAID U.S. Agency for International Development WDR World Development Report v Grain storage huts, Mali. Photo courtesy of Kimberley Fletcher; FletcherGallery@Mac.com. Acknowledgments A team from the Country Evaluations and Regional Dillinger, Matthew D. Glasser, Larry Hannah, Lance Relations unit of the Independent Evaluation Morell, Amitaba Mukherjee, Harry Anthony Group (IEG) team contributed to this report: Gita Patrinos, Geeta Sethi, Mathew Verghis, Deborah Gopal (task manager), Alisa Abrams, Tim L. de Wetzel, and Roland White, and Inder Jit Ruprah, Vaan, Laurie Effron, Carolina Mary Rojas Hayes, principal economist, Office of Evaluation and Victor Orozco, Parameswaran Pramod, Rupa Oversight, of the Inter-American Development Ranganathan, Govinda Rao, Rohit Tharakan, Rajiv Bank. The team acknowledges the assistance of Tharian, Gemi Thomas, and Barbara Yale. Deanna the task team leaders of the evaluated projects and Lijuaco, Arild Schou, Chil Soriano, and Germina economic and sector work. Comments on an Ssemogerere undertook field assessments in the earlier draft of this report from Andrew Reschovsky Philippines and Uganda. Pierre M. De Raet and (professor, University of Wisconsin­Madison) and Manuel Penalver-Quesada participated in country Joerg Faust (head of Department III: Governance, missions. Jonathan Rodden prepared a Statehood & Security, German Development background paper and contributed substantially to Institute) are gratefully acknowledged. the evaluation in several ways. IEG is grateful to the country directors and The team acknowledges the technical guidance country teams, particularly lead staff on decentral- providedbyStevenB.WebbandcommentsbyMark ization, in the countries where field visits were Sundberg, Victoria Elliot (former manager, IEG conducted. The country teams facilitated the visits Corporate Evaluation and Methods), Martha and organized stakeholder consultation meetings. Ainsworth,DeanNeilson,andRoyGilbert.Toneema The IEG team also acknowledges the contribution HaqandSvenjaWeber-Venghauspreparedliterature of several World Bank client country stakeholders reviews on different aspects of decentralization and who provided feedback for the evaluation. poverty reduction. The report was edited by Heather Dittbrenner and William B. Hurlbut. Vikki The evaluation was also informed by and Taaka provided administrative guidance. The team benefited from several background papers on is grateful for the guidance of Lily Chu (interim decentralization, the preparation of which was manager, Country Evaluations and Regional coordinated by Anwar Shah. These papers were Relations) during the approach stage and that of the prepared as part of a larger IEG effort to learn peer reviewers, Jennie Litvack and Ariel Fiszbein. lessons of evaluation in improving public sector performance in developing countries. Six IEG acknowledges the cooperation and input of background papers were also prepared for this Patricia Clark Annez, Kai Kaiser, and Lili Liu, who evaluation on Argentina (Erik Wibbels), Brazil collaborated with IEG in organizing four consulta- (Jonathan Rodden), China (Jing Jin), India tion meetings and contributed to the development (Nirvikar Singh), Mexico (Alberto Diaz-Cayeros), of the evaluation design. IEG also acknowledges and South Africa (Andrew Reschovsky). Their helpful input from Junaid Ahmad, William preparation was coordinated by Chad Leechor. Director-General, Evaluation: Vinod Thomas Director, Independent Evaluation Group­World Bank: Cheryl W. Gray Senior Manager, IEGCR: Ali M. Khadr Task Manager: Gita Gopal v i i Gathering in Bolivian village. Photo courtesy of World Bank Photo Library. Foreword Decentralization shifts responsibility and end of the evaluation period, by which point it accountability for the delivery of public services was rated high in about two-thirds of the focus to subnational (state, provincial, district, or local) countries. It was most successful in strengthen- levels of government, aiming to help improve ing legal frameworks for decentralization and service delivery and local governance. Since the intergovernmental relations, improving public 1990s, the World Bank has devoted an increasing financial management at the local level, and share of its financing to support its client helping central governments establish transpar- countries' decentralization efforts. To assess the ent fiscal transfer systems. It was much less results of such interventions, IEG reviewed Bank successful in helping to enhance own-source support for decentralization in 20 developing revenue at the local levels, clarifying responsibili- countries between fiscal years 1990 and 2007. ties of different levels of government, and This period had associated lending of about $22 strengthening citizen oversight. billion, covering roughly half of all lending that included support for decentralization. IEG's The evaluation confirms that there is no single evaluation did not attempt any systematic assess- approach that can be considered in supporting ment of the costs and benefits of decentraliza- decentralization across widely varying country tion in client countries, noting that in many situations. Solutions need to be tailored to cases, the countries had made a political decision country-specific contexts and driven by a commit- to decentralize well before requesting and receiv- ment to reform at all levels of government. Even ing World Bank support. then, outcomes are sensitive to and positively associated with aspects such as subnational According to the evaluation, Bank support for government capacity and political will. decentralization to the 20 focus countries was highly relevant. Most countries had embarked on That said, a number of directions emerge for a process to shift responsibilities for service seeking greater effectiveness of Bank support. One delivery to subnational governments in areas aspect concerns the need to ensure a cross-cutting such as primary education, basic health care, and understanding of the decentralization process, water supply. World Bank support in these areas which will help to ensure consistent and coherent was responsive to client country demands, support across sector units. Another area of consistent with their development policies, and importance is each government's commitment to aligned with the institution's mandate of improv- reform in this area, together with the need to ing the delivery of basic services to the poor. strengthen capacity, including at local levels. Finally, it pays for countries to develop results- But the support for decentralization was of mixed based frameworks to help assess the outcomes of quality, which nonetheless improved toward the reforms and the support for decentralization. Vinod Thomas Director-General, Evaluation i x Local meeting in Indonesia. Photo by Ray Witlin, courtesy of the World Bank Photo Library. Glossary Administrative decentralization: The process of redistributing authority and responsibility for providing public services from the central or national level of government to a subnational and/or local level "Big Bang" decentralization: A process wherein the central level of government announces decentral- ization; passes laws; and transfers responsibilities, authority, and/or staff to subnational and/or local governments in rapid succession Decentralization: A process of transferring responsibility, authority, and accountability for specific or broad management functions to lower levels within an organization, system, or program (defined narrowly for purposes of this evaluation in chapter 1, box 1.1) Decentralization framework: The full set of institutional arrangements in a given country within which decentralization is undertaken and sustained, specifying among other things how service delivery responsibilities and resource mobilization authority are apportioned among the various levels of government, as well as the associated accountability arrangements Deconcentration: The least ambitious level of decentralization, where responsibilities are transferred to an administrative unit of the central government that is spatially closer to the population where service is to be provided, usually a field or regional office Delegation: An intermediate level of decentralization, where some authority and responsibilities are transferred to a lower level of government, but there is a principal-agent relationship between the central and subnational government in question, with the agent remaining accountable to the principal Devolution: The most ambitious form of decentralization, where the central government devolves responsibility, authority, and accountability to subnational governments with some degree of politi- cal autonomy Fiscal decentralization: The decentralization of government expenditure and revenue-raising author- ity to subnational government structures in line with their allocated functional responsibilities Horizontal decentralization: Delegation of decision-making powers, functions, and resources by a given level of government to other agencies, committees, and/or structures within the same level Intergovernmental relations: Relations and accountability between different domains at a given level of government, as well as between different levels of government (for example, provincial-district) Local government unit: The constitutionally established government structure that operates "closest" (in the sense of level, rather than space) to communities and is responsible for basic service delivery x i D E C E N T R A L I Z AT I O N I N C L I E N T C O U N T R I E S Own-source revenue: Revenue from sources that have been assigned to a given subnational level of government, with the latter determining revenue rates Political decentralization: A process whereby the voice of citizens is integrated into policy decisions at a subnational level and civil society can hold the associated authorities and officials accountable Subnational government: A level of government below the national level; in federal systems, typically the level just below the central level Strategy: The means employed to reach desired objectives, typically comprising such elements as clear policy underpinnings; an action plan based on lessons learned; a monitoring plan; and adequate financial, human, and technical resources for implementation Vertical decentralization: Decentralization that occurs within sectors and departments but is not fully integrated and/or consistent with decentralization in other sectors and departments x i i Executive Summary I n recent years, developing countries have decentralized functions and re- sponsibilities to lower levels of government at an increasing pace. The main reasons for such reforms are often political, but governments also adopt them as a way to improve service delivery and local governance. Typically, after the political decision is made, a country will turn to its development partners-- including the World Bank--for support in implementing the new policies and achieving their development objectives. The Independent Evaluation Group (IEG) tion by subnational governments. This was often assessed the effectiveness of Bank support for a result of lack of political will. decentralization between fiscal 1990 and 2007 in 20 countries, seeking to inform the design and Other things being equal, Bank support brought implementation of future support. Given the better results where there was consensus around difficulties of measuring the results of decentral- the reform within the country prior to Bank ization, the evaluation used intermediate engagement and when the support was outcome indicators--such as strengthened legal combined with incentives for institutional and regulatory frameworks for intergovernmen- reform at the subnational level. Looking forward, tal relations, improved administrative capacity, the results of Bank support for decentralization and increased accountability of subnational can be strengthened with more timely and governments and functionaries to higher levels coordinated analytical work to underpin it, by of government and to citizens--to assess the better coordinating fragmented sector-by-sector results of Bank support in these 20 countries. To interventions, and by accompanying support for examine potential lessons at a sectoral level, the policy reform with technical assistance to evaluation also assessed whether Bank support strengthen local government capacity. for decentralization improved intermediate outcomes for service delivery in the education Introduction sector in 6 of the 20 countries. All 20 countries reviewed in this evaluation have devolved significant responsibilities to lower Bank support contributed to more effective levels of government. Politics may be behind this decentralization--substantially in more than trend, fueled by a desire among citizens for one-third of the 20 cases and modestly in the strengthened democracy and improved gov- others. The most successful aspects of Bank ernance and service delivery. But both the support pertained to the legal frameworks for emergence of strong urban economies and intergovernmental relations, the frameworks for ethnic tensions that threaten national identities intergovernmental fiscal transfers, and sub- also motivate governments to move closer to the national financial management. Bank support people. Governments have not usually asked the was less effective in clarifying the roles and World Bank to help with their decisions about responsibilities of different levels of government whether to decentralize, nor has the Bank and in improving own-source revenue mobiliza- typically advocated decentralization, except in x i i i D E C E N T R A L I Z AT I O N I N C L I E N T C O U N T R I E S particular sectors. Usually--in 12 of the 20 case- evaluation does not assess community-driven study countries--governments have decentral- development, which was assessed in a separate ized for political reasons and only subsequently IEG evaluation (2005b). asked the Bank to help implement the process, make it more rational, and improve service The evaluation framework delivery and accountability. It has yet to be conclusively demonstrated that there is a causal link between decentralization Objectives of the evaluation and improved service delivery, good governance, IEG assessed the effectiveness of Bank support for or macroeconomic stabilization. Moreover, decentralization provided to 20 countries between decentralization is a long-term agenda--in- fiscal 1990 and 2007. These countries were dustrialized countries often took more than a selected to ensure regional representation, and century to reach their current state of decentral- they accounted for 47 percent of all Bank commit- ization, and one or two Bank country strategy ments containing decentralization components periods are simply inadequate to assess that. The during the period. The aim was to examine what decentralization process is also typically worked and what did not to inform the design and disjointed and subject to periods of progress and implementation of future Bank support. reversal. Therefore, rather than focus on the whole decentralization process or the connec- Decentralization has many meanings, but for this tions between decentralization and service evaluation it was defined as the transfer of delivery, IEG focused on a set of intermediate authority and responsibility for governance outcomes that are essential for good service and public service delivery from a higher to a delivery in decentralized settings. lower level of government. The characteristic that distinguishes decentralization from, say, The key desired result is fiscally responsible, simply shifting resources to local governments is responsive, and accountable subnational govern- that decentralization seeks to create relation- ments that are likely, under the oversight of ships of accountability among citizens, service citizens and higher-level government, to improve providers, and subnational governments and service delivery and governance. The desired between the latter and central governments. The result has several components: Figure ES.1: Framework for Assessing the Results of Bank Support for Decentralization Inputs Outputs Intermediate outcomes Final outcomes Lending Fiscal: Rules for Improved legal and Better service revenue and regulatory frame- delivery Economic and expenditure and works for fiscal sector work borrowing relationships and Improved service delivery governance Other nonlending Administrative: support Local control over Improved human resources, administrative budgeting, financial capacity management Better upward and Political: Citizen downward participation accountability Exogenous factors Source: IEG. x i v E X E C UT I V E S U M M A RY · Strengthened legal and regulatory frameworks The quality of the Bank's lending portfolio to for intergovernmental relations these 20 countries was also mixed during the · Improved administrative capacity 1990s. Weak understanding of political economy · Increased accountability of subnational gov- factors and associated risks led to overly ernments and functionaries to higher levels ambitious objectives that often limited develop- of government and to local citizens. ment effectiveness. Bank support for decentral- ization was provided by various sector units, with IEG assessed the extent to which Bank support objectives that were not always consistent at the contributed to progress toward these objectives country level. Bank support focused on from 1990 to 2007. At the sectoral level, the evalua- decentralization frameworks, but did not always tion assessed the extent to which Bank support in provide parallel support to strengthen the the education sector in 6 of the 20 focus countries technical capacity of the subnational govern- helped generate resources for local governments ments to whom responsibilities and resources to deliver services (consistent with formal were transferred. Monitoring of the progress of intergovernmental fiscal frameworks), strength- Bank support for decentralization was weak; the ened institutions and capacity for improved service Bank focused on output- or process-level indica- delivery, and enhanced the accountability of local tors, such as the passage of laws or fiscal governments for service delivery to citizens. transfers, rather than on the performance of local governments and other institutions in delivering Evaluation Findings services. Quality of Bank support In the last five years of the evaluation period, the To understand the quality of Bank support, the quality of Bank support for decentralization evaluation reviewed all Country Assistance improved in 15 of the 20 countries. Bank analyti- Strategies in the 20 countries, more than cal work provided a better understanding of the 40 pieces of relevant economic and sector work, broader implications of decentralization for and 203 lending activities with decentralization service delivery and governance and in turn components. The 203 lending activities in these influenced the design of country strategies. countries had associated financial commitments Country-level assistance was therefore internally of about $22 billion, of which about $7.4 billion more consistent. was specifically for decentralization-related activities. In several countries, support for policy reform was combined with technical assistance to The quality of Bank economic and sector work on strengthen different levels of government, and decentralization was mixed during the 1990s. The the Bank increasingly supported country efforts work was not always timely and in several to assess the results of decentralization in terms countries did not influence the design of Bank of strengthened local government performance. Country Assistance Strategies. Of the 20 country Donor collaboration also improved during this cases studied for this evaluation, decentralization- period, and in several of the 20 countries joint related diagnostic reports existed in 16. Only in 8 diagnostic and analytical work--including at was timely analysis of the implications of decentral- subnational levels--led to joint support for ization policy undertaken within five years of decentralization. issuance of the relevant laws. The evaluation also found little evidence that broader analytical work This said, the organizational structure within the on decentralization had substantial influence on Bank has in general resulted in less-than- Bank operations in the countries studied. Sound optimum support for decentralization at the analysis, when it was done, tended to affect country level. An absence of clear leadership and assistance from the same Bank sector unit that coordination across sectors persists, except in a undertook the analysis, but not usually beyond. handful of cases where country directors and/or x v D E C E N T R A L I Z AT I O N I N C L I E N T C O U N T R I E S vice presidents have broken the sector-silo that were inconsistent with client country approach, thereby enabling more consistent objectives. support to client countries. The evaluation reviewed Bank support for Results of Bank support decentralization in the education sector in The evaluation divided the review of the results greater depth in 6 of the 20 focus countries of Bank support for decentralization into two (there are ongoing evaluations in health, water, parts: support for the development and/or and municipal management). The evaluation strengthening of decentralization frameworks, found that sector-level efforts to decentralize and support for improving service delivery in the education services were not usually sustained or education sector. effective unless they were designed and implemented at the country level within a broad In supporting the development and/or strength- decentralization framework. ening of decentralization frameworks, the Bank generated outcomes that were high or substan- The evaluation did not attempt to aggregate tial in 7 countries, modest in 12, and negligible in ratings of the quality of Bank support or ratings 1. Bank support for decentralization was most of the results of Bank support for decentraliza- successful in helping strengthen the legal tion frameworks into a single rating for each of underpinnings of intergovernmental fiscal the 20 countries. However, a comparison of the relations. The Bank helped establish frameworks ratings for quality and results indicates that when for prudent borrowing and debt management, the quality of Bank support improves, the results generating substantial results in half the also get better. This suggests that closer monitor- countries to which it provided support. Support ing of the quality of Bank support for decentral- for strengthening financial accountability of ization will likely improve the Bank's con- subnational governments to higher levels of tribution to overall results in the country. government also generated substantial results. The Bank was less successful in helping to Recommendations strengthen frameworks for own-source revenue In many of its country programs, the Bank has or in enhancing such revenue; it contributed to made a de facto strategic decision to support substantial and sustained results in only five decentralization and development of sub- countries. The Bank also was not very successful national government capacity. In a few cases-- in helping clarify the responsibilities of the such as where the client country has made various levels of government or in supporting decentralization a cornerstone of its develop- monitoring at the local level. ment strategy and has demonstrated political commitment to decentralizing--Bank support The Bank contributed to better results in has been built on an explicitly cross-cutting countries where the political will to decentralize approach. In most cases, however, Bank support was strong, where there was greater clarity on has taken a sector-specific route, targeting the type of fiscal and administrative decentraliza- decentralization and/or development of tion to be pursued, and where Bank support was subnational government capacity as a logical way aligned with the client's decentralization strategy. of supporting more effective and responsive This was the case notably in two post-conflict service delivery in that sector. In these latter countries, where consensus on the need to cases, the various Bank sector units have not minimize the potential for conflict was always provided consistent or coherent support compelling. In countries where there was less for decentralization. consensus on the approach to implementing fiscal or administrative decentralization, the Looking forward, IEG offers the following results of Bank support were weaker. This was recommendations. They are applicable to every often because the Bank supported approaches client country that has transferred at least some x v i E X E C UT I V E S U M M A RY responsibility for service delivery to subnational itoring and evaluation that focuses on local governments, where the Bank has made a de outcomes (such as enhanced accountability, facto strategic decision to provide support for greater citizen participation, and improved decentralization through either a cross-cutting service delivery) rather than on just the process or a sector-specific approach: of decentralization. · Ensure that Bank support at the country level · Ensure that Bank support--particularly is (among other things): lending--is underpinned by genuine client ­ Founded on a clear analytical framework commitment to decentralized service delivery, based on an integrative understanding of given its importance to the success of Bank in- economic, political, and institutional fac- terventions. Occasionally, a role for the Bank tors at different levels of government and may be justified in the absence of client com- across sectors affected by decentralization mitment (for example, to forestall potentially ­ Accompanied by support (from the Bank or adverse measures), although the evaluation others) to develop and maintain local gov- finds that Bank interventions under such cir- ernment capacity, to the extent feasible. cumstances are not usually effective. · Strengthen institutional arrangements within · Encourage the adoption of a more results- the Bank to ensure that an integrative view based approach to decentralization by helping underpins Bank interventions, particularly develop in-country and Bank capacity for mon- those based on sector-specific entry points. x v i i Village meeting in Indonesia. Photo by Ray Witlin, courtesy of the World Bank Photo Library. Management Response M anagement welcomes the Independent Evaluation Group (IEG) eval- uation of World Bank support for decentralization, covering en- gagement in 20 client countries and operations between 1990 and 2007. Management broadly concurs with the review's recommendations, al- though these require some nuance and clarification; subject to these, man- agement intends to build on the review's findings in its work in supporting countries that ask for assistance on decentralization. Concurrence with the Broad Thrust of with partner countries in supporting their the Analysis and Recommendations decentralization efforts, (ii) support for The evaluation contains important conclusions countries' monitoring of decentralization efforts that management welcomes in the context of its and evaluation of results, and (iii) support for overall assistance to countries on decentraliza- strengthening local governments. tion and improved service delivery. Management recognizes that an evaluation of decentralization Objectives for Bank engagement with was from the outset likely to be challenging, decentralization and attribution given its cross-cutting and often politically driven Management embraces the report's emphasis on nature. Management appreciates the finding that improving service delivery as the ultimate Bank support in this area has improved notably objective of our involvement with decentraliza- over the latter period of the evaluation. We tion. Our work on decentralization per se will be broadly welcome a number of the recommenda- driven by this goal. Management recognizes that, tions that can further support this trend going with regard to decentralization, as with virtually forward. We welcome the overall thrust of the all types of support, alignment of Bank engage- evaluation that engagement on decentralization ment with country commitment is central to should reflect the prevailing client country achieving sustainable development results. context, to support development results. We also However, management notes that decentraliza- appreciate the overall tone of the evaluation, tion has not proven to be either a necessary or a which suggests that in general the Bank has not sufficient condition for good service delivery. Nor acted as an "ideological" proponent of decentral- is there any necessary link between a client's ization, but has sought to support decentraliza- intention to decentralize government and a tion as an intermediate means to improve service primary commitment to improve service delivery and poverty reduction. Management has delivery, which may often be driven by other only a few issues to raise with regard to the (valid) political objectives. Thus, our engage- analysis in the review. ment on helping countries improve service delivery must be guided by our judgment of the Management Observations best steps toward achieving that goal, taking into Management's observations center on three account government commitment on a host of issues: (i) the objectives for Bank engagement issues that affect it, including decentralization. x i x D E C E N T R A L I Z AT I O N I N C L I E N T C O U N T R I E S Monitoring and evaluation this area is necessarily unique. Our strategies of decentralized results must accordingly reflect this diversity, and Management is very much in accord with the management must, in partnership with client recommendation to help partner countries countries, develop a vision of intermediate goals monitor results more carefully. Management that may emphasize some sectors more than emphasizes that measurement should be others in different phases. Management shares extended to assessing outcomes in terms of the goal of improving capacity in local govern- improved service delivery. Measurement of both ments but notes its emphasis on "as needed to of these elements would permit a much-needed improve local service delivery," and it will design assessment of the actual transmission mech- its support to encourage this. However, manage- anism between adherence to basic and very ment does not subscribe to the notion that the broadly defined principles of public finance/ only--or best--means of achieving capacity public sector reform and governance (such as goals is through direct support to local govern- enhanced citizen accountability) and actual ments from the World Bank or others, especially improvements in much-needed services. Such because it risks being supply driven. Con- discipline is needed to ensure that intermediate sequently, engagement in decentralization must instruments are ultimately assessed for their pay particular attention to effective and impact on outcomes. Furthermore, the Bank incentive-compatible efforts to strengthen local needs to work closely with client counterparts to governments, for example, through support ensure that monitoring systems are themselves linked to performance (including on intermedi- sustainable and to measure impacts over longer ate process indicators and final outcomes). periods of time, given that decentralization Management also notes that the Bank's recent reforms may be subject to significant lags. Governance and Anti-Corruption Strategy Beyond measurement, the Bank can assist client (World Bank 2007e) encourages governments to countries in designing feasible impact assess- explore demand-side accountability, including ment strategies for decentralization measures. citizen oversight. Management has already committed to report back to the Board on the Support for strengthening local government implementation of this strategy. Management is encouraged that the report notes recent progress in taking an integrative, multidis- Conclusion ciplinary approach to assisting countries that are Overall, management welcomes this evaluation grappling with the decentralization of their from IEG and broadly accepts its recommenda- governments. Management intends to continue tions. Detailed responses to the recommenda- to reinforce and broaden this trend, while tions are outlined in the Management Action recognizing that each country's experience in Record. x x M A N AG E M E N T R E S P O N S E Management Action Record IEG Recommendation Management Response Ensure that Bank support--particularly lending--is underpinned Ongoing/Agreed in Principle. Management agrees in princi- by genuine client commitment to decentralized service delivery, ple with this recommendation; however, it notes that the degree given its importance to the success of Bank interventions. Oc- of country consensus around appropriate decentralization modal- casionally, a role for the Bank may be justified in the absence of ities may differ by sectors and levels of government. Conse- client commitment (for example, to forestall potentially adverse quently, the Bank can often play a critical role helping countries measures), although the evaluation finds that Bank interven- identify and implement entry points intended to promote qual- tions under such circumstances are not usually effective. ity in decentralized service delivery. However, a number of cases suggest that political drivers for decentralization are often an im- portant country dynamic, and Bank country strategy design and implementation must be responsive to these realities. Management agrees to continue to emphasize cross-cutting client commitment for decentralized service delivery as a critical factor in defining entry points for successful engagement, and to calibrate assistance to support sustainable service delivery and poverty re- duction as good practice in relevant Country Assistance Strategy (CAS) design and implementation. To implement the Governance and Anti-Corruption (GAC) Strategy, the Bank's Regional vice pres- idential units have identified 26 countries that are initiating country- specific GAC processes. Where country conditions involve significant issues of decentralized service delivery (for example, the Democratic Republic of Congo, Indonesia, and the Philippines), country teams are systematically deepening the analysis around political-economy circumstances in shaping development effec- tiveness. At the conclusion of this learning process, Bank man- agement is committed to reporting to the Board whether and how it intends to systematize and scale up its GAC work, including an- alytic and advisory activities. Reporting on the above agreed ac- tions will be done in the context of overall GAC reporting. Encourage the adoption of a more results-based approach to de- Ongoing/Agreed. The Bank will continue to place particular em- centralization by helping to develop in-country and Bank capac- phasis in its dialogue with client countries on supporting the ex- ity for monitoring and evaluation (M&E) that focuses on local istence of a credible data-collection and reporting system for outcomes (such as enhanced accountability, greater citizen par- relevant services, one that is consistent with the structure of de- ticipation, and improved service delivery) rather than on just the centralization both in terms of its level of aggregation and man- process of decentralization. agement responsibilities and that can be used to make midcourse corrections as needed. At the same time, the design of M&E sys- tems needs to highlight the presumed results chains between in- termediate outcomes and ultimate outcomes (service delivery). For operations supporting a broad-based engagement in decen- tralized outcomes, countries will be encouraged to adopt a sys- tem that reflects the M&E system in place at the national level. x x i D E C E N T R A L I Z AT I O N I N C L I E N T C O U N T R I E S Management Action Record IEG Recommendation Management Response At the national level, the Public Expenditure and Financial Ac- countability (PEFA) 8 indicators (concerning intergovernmental fis- cal relations) and the PEFA 23 indicators (concerning front-line school and health facility financing) provide for systematic as- sessments of intermediate outcomes. The PEFA Secretariat also recently issued guidance on subnational government applications of this public financial management diagnostic tool, useful to coun- tries to consider for M&E at this level. As part of the GAC Im- plementation Plan, management has committed to strengthen the application of Actionable Governance Indicators, including those in the area of decentralization. Management will continue its efforts to provide country and in- dividual operations teams with guidance on helping countries strengthen M&E around decentralization. Ensure that Bank support at the country level is (among other Ongoing/Agreed in Part. Given the cross-cutting nature of things): decentralization, management agrees that an integrated ap- · Founded on a clear analytical framework based on an inte- proach is important as part of general and specialized economic grative understanding of economic, political, and institu- and sector work, including attention to the political-economy of tional factors at different levels of government and across decentralization (for example, Public Expenditure Reviews and sectors affected by decentralization sectoral diagnostics, including around GAC issues). · Accompanied by support (from the Bank or others) to develop and maintain local government capacity to the extent feasible. Management agrees with the first part of this recommendation and will implement it and monitor progress in the context of the GAC work noted in its response to the first recommendation. Management agrees that support for subnational capacity can be a vital ingredient to strengthening service delivery outcomes. However, management notes that strengthening government capacity must be linked to ultimate service delivery outcomes and based on appropriate engagement models and Bank compara- tive advantage, especially in the presence of a large number of diverse subnational jurisdictions. Local capacity building cannot be limited to, for example, the supply of training, but depends on appropriate incentives. Experience shows that it must be de- mand driven to be effective. Although strengthening local capacity often represents an important element for effective decentral- ization, management does not commit to always supplying local capacity building as an element in the activities it supports. x x i i M A N AG E M E N T R E S P O N S E Management Action Record IEG Recommendation Management Response Strengthen institutional arrangements within the Bank to ensure Ongoing/Agreed. Management underscores the role of coun- that an integrative view underpins Bank interventions, particu- try management units and Regional vice presidential units (with larly those based on sector-specific entry points. support from multisectoral communities of practice in the Regions) in ensuring that consistent approaches are implemented as part of the CASs and operational review processes. Management will ensure that the issue is raised early in relevant CAS discussions. At a Bank-wide level, the Decentralization and Sub-National Regional Economics Thematic Group will continue to serve as a platform to promote integrative approaches to strengthen results- based decentralization engagements, working together with other related cross-cutting thematic groups, notably the Urban Economics, Finance, and Management Thematic Group. Within the Bank-wide initiative to revitalize and support communities of practice, management sees this as a cross-network area to pri- oritize. This work will facilitate coordination across net- works/sector teams at the Regional level. To support coherent approaches, the thematic group structure will offer senior facil- itation and advisory services on a demand basis to country teams engaged in upstream CAS or project design issues. Specific steps to strengthen Bank engagement on issues of decentralization and local governance are under preparation under the guidance of senior operational management. x x i i i Meeting on women's role in society, Dushanbe, Tajikistan. Photo by Gennadiy Ratushenko, courtesy of the World Bank Photo Library. Chairperson's Summary: Committee on Development Effectiveness (CODE) O n May 21, 2008, the Committee on Development Effectiveness (CODE) discussed the "Evaluation of Bank Support for Decentralization in Client Countries" and the draft management response. Background decentralized service delivery; (ii) encourage a Key strategy documents include Strengthening more results-based approach by helping to World Bank Group Engagement on Governance develop in-country and Bank capacity for and Anticorruption (World Bank 2007l) and monitoring and evaluation that focuses on local Reforming Public Institutions and Strengthening outcomes; (iii) ensure that Bank support is based Governance: A World Bank Strategy (World Bank on a clear analytical framework, accompanied by 2000c). The updates on the implementation of the strengthening of local government capacity; and 2000 Bank strategy were prepared in 2002 and (iv) strengthen institutional arrangements within 2005 as part of the "Sector Strategy Implementa- the Bank to ensure that an integrated view tion Update: FY05" (World Bank 2005g), which was underpins Bank interventions, particularly those discussed by CODE. Related Independent Evalua- based on sector-specific entry points. tion Group (IEG) evaluations include the recent Public Sector Reform: What Works and Why? (IEG Draft management response 2008), considered by CODE in March 2008. Although concurring with the broad thrust of the analysis and recommendations, management Main Findings and Recommendations commented on three issues: (i) the objectives of IEG assessed the effectiveness of Bank support the Bank's engagement with partner countries in for decentralization between fiscal 1990 and 2007 supporting their decentralization efforts, which is in 20 countries, seeking to inform the design and centered on helping improve service delivery; (ii) implementation of future support. The evalua- support for countries' monitoring and evaluation tion also assessed the results of Bank support for systems; and (iii) support for strengthening local decentralization of service delivery in the governments, where it notes the importance of education sector. IEG found that the most incentives and demand-driven approaches. successful aspects of Bank support pertained to the frameworks for intergovernmental relations DGE statement including fiscal transfers and subnational In his statement, the Director-General, Evalua- financial management. Bank support was less tion (DGE) noted a few challenges specific to this effective in clarifying the roles and responsibili- evaluation: the consequent need to formulate an ties of different levels of government and in operative definition, an absence of a specific improving their own-source revenue mobiliza- articulated sector/thematic strategy, and the tion. Looking ahead, IEG makes the following long-term nature of the process. Accordingly, the recommendations: (i) ensure that Bank support evaluation needed to be based on intermediate is grounded in genuine client commitment to outcome indicators. The DGE highlighted three x x v D E C E N T R A L I Z AT I O N I N C L I E N T C O U N T R I E S key lessons from the evaluation: (i) Bank support Notion of decentralization and scope for decentralization is highly relevant, (ii) Bank of evaluation support to decentralization should be tailored to Speakers expressed divergent views about the varying country contexts, and (iii) results of Bank notion of decentralization and the scope of the support have been better when such support IEG evaluation. Some members felt that limiting was framed around a country-led strategy. the notion of decentralization to service delivery was misleading. IEG responded that the subject Overall Conclusions and Next Steps for evaluation had been narrowed intention- The Committee welcomed the opportunity to ally because the service delivery element was discuss the IEG evaluation. CODE found the the critical one for poverty reduction and a management response to be constructive, accept- major element underlying intergovernmental ing, and appreciative of IEG's findings and relations. A member noted that the decentraliza- recommendations. A few speakers wondered tion process always involved costs (for example, whether in such instances of substantive entente fiscal, administrative, and costs of coordination there was any value added in holding a CODE failures and income distribution). To this end, discussion. Other members disagreed, because of members suggested measuring the efficiency of the strategic salience of the topic and the need for decentralization by costs and benefits as well as an independent validation of confirmation of what by its contribution to effective public service the Bank was or was not doing. Some speakers delivery. In this regard, speakers expressed a expressed their disappointment with (i) the overly preference to include in the evaluation an narrow definition employed, focusing on service analysis of costs and benefits of decentralization. delivery; (ii) neglect of the whole cost side of A member suggested addressing the cost side in decentralization and, more broadly, of its sustain- the Public Expenditure Reviews. IEG responded ability, noting in particular aspects such as fiscal that chapter 3 of the report partially covered the costs, coordination failures, administrative cost side by describing the fiscal costs and the duplication, and distributional consequences. potential macrostabilization issues. At the same Relatedly, a few other speakers wondered whether time, IEG noted that (i) the evaluation intended the Bank's efforts (and hence the evaluation) to assess the outcome of Bank support for were not sidestepping the key policy debates and decentralization, not to weigh costs and benefits suggested it could be more ambitious in helping of decentralization; (ii) costs and benefits had a client countries think through the options. There number of dimensions that were beyond the was broad endorsement of the recommendation Bank's mandate and were political or quasipo- on client commitment and political will, although litical in nature; and (iii) it was difficult to a cautionary note was expressed regarding the capture all dimensions of costs and benefits. complexity of the issue and attendant nuances in Management agreed to adopt a more systemic reading the messages. There was overall consen- approach to reviewing Public Expenditure sus among the speakers that decentralization is a Reviews, especially when they include analysis means and not the objective per se; it is a country- of the fiscal costs of decentralization. driven process and the Bank should play a supportive, demand-driven role in it. In this Other speakers broadly agreed with the thrust of regard, it was noted that understanding the politi- IEG's recommendations and appreciated the cal economy of the decentralization context is constructive draft management response. They important. The findings and recommendations found this evaluation to be very strong and noted (accepted and indeed elaborated on by manage- that decentralization is a new approach to ment) regarding institutional fragmentation and development. Members agreed that decentral- the need for increased coherence and integration ization was just a tool that could serve different across sectors resonated with many speakers. purposes, including but not limited to improving public service delivery. In this context, a member The following main issues were raised. preferred the term "intergovernmental relation- x x v i C H A I R P E R S O N ' S S U M M A RY: C O M M I T T E E O N D E V E L O P M E N T E F E C T I V E N E S S ( C O D E ) ships" to the term "decentralization." A speaker Development capacity at the local level noted that promoting spatial planning and A member was concerned about IEG's regional development is an important objective recommendation to develop the capacity of local of decentralization and appreciated some indica- governments. He noted that Bank clients are tions from IEG on the relevance of this objective central governments and did not believe that the for the Bank's work. Bank's engagement at the local level would have a significant impact. In this regard, some Focus of Bank's support for decentralization speakers underlined that such engagement Although welcoming IEG's finding about should be demand driven. Both IEG and improved quality of Bank support for decentral- management agreed with the need for Bank ization, some members suggested that the Bank involvement at the local level to be demand be more ambitious in encouraging countries' driven. A member sought clarification from decentralization efforts where it is appropriate management about whether the development of and contributing to policy debates, particularly in local capacity was an issue of incentives or the small and fragile states. They recommended training. Management noted that both aspects enhancing Bank support by sharing best practice. are important; however, experience shows that IEG broadly agreed with this recommendation training needs to be demand driven. but stressed the importance of proceeding with caution, given that decentralization in many Coherence in the Bank's advice countries (including several small states) Speakers took note of the IEG finding about the reflected deeply held political beliefs about its existing fragmentation within the Bank in value. Some speakers encouraged the Bank to providing support to decentralization. They play an advisory role and focus more on analytical emphasized the importance of consistent Bank work and technical assistance. A speaker sought advice to client countries and requested more information about the appropriate staff skill information on the specific actions aimed at mix to effectively support decentralization efforts ensuring internal collaboration and avoiding in client countries. fragmentation be included in the Management Action Record relating to the report. Manage- Country context ment explained that the Bank did not find it Members agreed that country specifics should desirable to establish one central unit to be taken into account. A speaker emphasized the coordinate all work. At the same time, manage- importance of understanding which factors drive ment implements specific actions to improve the decentralization agenda. The need for coherence in three areas: (i) at the country studying a country's political economy was level, through strengthening upstream diagnos- reiterated. A member suggested discussing and tic, embedding decentralization agenda in analyzing the final outcomes of decentralization Country Assistance Strategies/Country Partner- in countries. Some speakers felt that decentral- ship Strategies, and the existing Anchor Groups ization is the right option for the African in SD, HD, and PREM; (ii) at the central level, countries. One member was not certain about through cross-cutting thematic groups such as the findings of IEG that Bank support is weak those on decentralization and urban econom- when it is inconsistent with the government view ics, finance, and management; and (iii) at the and that it is successful in the countries with Bank-wide level, through the Governance existing political will. He noted that there is no Council. A member asked about the additional clear concept of political will as far as decentral- costs of ensuring coherence. Management ization is concerned. In this context, another clarified that this is not a major resource issue. member remarked that the level of decentraliza- tion often depends on the central government's Donor harmonization decisions and, hence, decentralization has a very Some members stressed the need for donor collab- political nature. oration and harmonization for effective partner- x x v i i D E C E N T R A L I Z AT I O N I N C L I E N T C O U N T R I E S ships with client countries. Management clarified tives (that is, applying both top-down and that the Bank is actively involved in the bottom-up approaches for analysis and evalua- harmonization process through the Development tion). He said that in many countries, the Partners Group on Local Governance and decentralization process comes as a reaction of Decentralization that is informally aligned with the central government to bottom-up requests the Organisation for Economic Co-operation and and development and asked IEG to take this into Development-Development Assistance Committee. account for future evaluations. The importance of applying the notion of subsidiarity to Globalization and localization decentralization was also emphasized in this A member suggested looking at decentralization context. from both globalization and localization perspec- Jiayi Zou, Chairperson x x v i i i Chapter 1 Evaluation Essentials · Countries have increasingly adopted decentralized forms of governance and have turned to the Bank for sup- port in their efforts. · Potential gains from decentralization can include enhanced government accountability and more effective and efficient service delivery. · Risks associated with decentraliza- tion include macroeconomic insta- bility, corruption, and elite capture. · In assessing intermediate outcomes in 20 countries from fiscal 1990 to fis- cal 2007, IEG included the quality of in- tergovernmental fiscal frameworks, subnational institutions and adminis- trative capacity, and accountability of subnational governments. Multipurpose Hall in Sri Lankan village. Photo by Dominic Sansoni, courtesy of the World Bank Photo Library. Background and Objectives C ountries have been turning to the World Bank for support with de- centralization for more than two decades. Beginning in the 1980s, and with a marked increase in pace and scale beginning around the mid- 1990s, the Bank has provided some form of lending and nonlending support to 89 client countries. This evaluation, which covers the period fiscal the emphasis more recently has been on 1990 to 2007, assesses the quality and results of economic arguments. Development agencies, that support in 20 of those 89 countries where among others, have tended to focus on the support was most substantial in terms of World benefits of decentralization for service delivery Bank financing (the support to these 20 based on the principle of subsidiarity1 countries represents about 47 percent of total (Shah 1998; Litvack and Seddon 1999; Recent arguments for Bank financial support for decentralization Manor 2003; World Bank 1997b, 2004j; decentralization focus on during the evaluation period). A key aim of the Ahmad and others 2005; and Shah efficient service delivery. evaluation is to draw lessons for future Bank 2006b). The focus has been primarily support for decentralization. on fiscal decentralization, with the goal of improv- ing allocative efficiency by bringing citizens closer Decentralization is the transfer of administrative to decision making and giving them the chance to and financial authority and responsibility for prioritize the use of public resources. Such fiscal governance and public service delivery from a decentralization can also, under certain higher level of government to a lower level. The conditions, strengthen accountability, because precise dimension (or ambition), appellation, level citizen participation, monitoring, and control of of responsibility, and set of government authorities local governments are relatively easier at the local involved has varied widely by country. In terms of level (OECD-DAC 2004). general patterns, however, the development litera- ture identifies three dimensions and three modes The 2004 World Development Report: Making of decentralization (box 1.1). This evaluation seeks Services Work for Poor People (World Bank to cover all three dimensions and modes. 2004j) notes that decentralization is one institu- tional mechanism that is often proposed to Literature Review: Arguments for and improve service delivery. It adds that experience against Decentralization and Findings with the use of this mechanism has varied and from Empirical Studies has produced mixed results. The World Develop- Although political arguments in favor of decentral- ment Report (WDR) notes that in Bolivia the ization have been advanced for several centuries, creation of rural local governments was associ- 3 DECENTRALIZATION IN CLIENT COUNTRIES Box 1.1: What Is Decentralization? Dimensions Modes Administrative decentralization--How responsibilities and au- Devolution--The deepest form of decentralization, in which a thorities for policies and decisions are shared between levels government devolves responsibility, authority, and accountability of government and how these are turned into allocative to lower levels with some degree of political autonomy outcomes Delegation--Some authority and responsibilities transferred, Fiscal decentralization--The assignment of expenditures, rev- but with a principal-agent relationship between the central and enues (transfers and/or revenue-raising authority), and bor- lower levels of government, with the agent remaining account- rowing among different levels of governments able to the principal Political decentralization--How the voice of citizens is inte- Deconcentration--The shallowest form of decentralization, in grated into policy decisions and how civil society can hold au- which responsibilities are transferred to an administrative unit of thorities and officials accountable at different levels of the central government, usually a field, regional, or municipal office government Source: IEG staff. ated with dramatic shifts in public allocations Gershberg (2000) found good outcomes associ- away from infrastructure and into social ated with increased local autonomy, although sectors--and a sharp fall in the geographic they also found significant impediments to realiz- concentration of public investments as they ing the potential from decentralization in some become more evenly dispersed across regions countries. Studies have also found that increasing (Faguet 2001). In some other countries, in parents' participation in community-managed contrast, decentralization has led to increased schools led to significantly lower rates of student regional inequalities and the "capture" of public and teacher absenteeism in El Salvador (Jimenez resources by local elites. and Sawada 1999); that decentralization improved students' test scores in Argentina Case studies at a sectoral level also do not lead to (Galiani and Schargrodsky 2001; Eskeland and any firm conclusions about the benefits of Filmer 2002); and that decentralized manage- decentralization. Some case studies in the health ment of schools led to improved achievement sector have found that decentralization has had scores in Nicaragua (King and Ozler 1998). negative effects on the quality of service delivery. Lakshminarayanan (2003) found that devolution The literature also highlights risks associated in the Philippines disrupted the integrity of the with decentralization when local capacity, referral chain related to the delivery of services accountability, and institutions are weak. The such as emergency obstetric care. In Zambia and 1997 WDR (World Bank 1997b) warns of the risk again in the Philippines, Aitken (1998) found that of macroeconomic instability caused by changes in salary levels, inadequate funding of weakened fiscal discipline under decentraliza- local health programs, and the politicization of tion. Poor local implementation capacity can lead local appointments increased uncertainty and led to increased inefficiency and waste in service to deterioration in staff morale and quality of care. delivery (Bird, Ebel, and Wallich 1995; Prud'homme 1995; and Tanzi 1996). In addition, Research has found In the education sector, in contrast, capture by political and other local elites can weaker connections case studies have generally found readily emerge as power is transferred to the between decentralization positive results from decentralization, local level, where entrenched inequities may and service delivery in particularly where local communities help elites orient service delivery toward health but stronger became involved. In a survey of themselves and away from the poor (Jette 2005; connections in education. international experience, Winkler and Manor 2003; OECD-DAC 2004; and von Braun 4 BACKGROUND AND OBJECTIVES and Grote 2000). Corruption may also increase whether to decentralize, because this Risks include with decentralization if personal or family ties political decision will often have been macroeconomic begin to override legal or regulatory considera- made and is outside the sphere of instability, corruption, tions (von Braun and Grote 2000). Government influence of technical specialists and and elite capture. ownership can falter, leading to a reversal of the policy advisers. Instead, it is how best process, if decentralization results in stronger to implement decentralization so that service political opposition to the ruling party (Brosio delivery, particularly to the poor and disadvan- 2002; Crook and Sverrisson 2001; von Braun and taged, does not deteriorate. Grote 2000). Decentralization: Although decentralization has a mixed record Everyone Is Doing It with regard to service delivery, there is general Most World Bank client countries have decentral- recognition that certain elements are necessary ized to at least one level of elected subnational (but not sufficient) for positive impact. These government. Decentralization started as early as include adequate financial resources, accounta- the 1960s in Nepal and Tanzania. bility for the use of resources, and government These early efforts focused on social Certain elements are commitment and ownership (Crook and Sverris- service delivery by local governments necessary (but not always son 2001; Manor 2003; OECD-DAC 2004; and with participatory planning. By the sufficient) for a positive Ahmad and others 2005). If decentralization 1980s many other countries were impact. takes place without these three conditions, pursuing various forms of decen- diffused accountability and poor service delivery tralization--in Latin America (Bolivia, Brazil, are likely to result. Some therefore argue that Chile, Colombia, Guatemala, Haiti, Jamaica, certain threshold levels of local capacity need to Nicaragua, and Mexico); Africa (Cameroon and be in place to ensure that decentralized funds Uganda); South Asia (Maldives, Nepal, and Sri and services can be managed properly. Lanka); and East Asia (Vietnam). In the 1990s, many Eastern European countries began In short, the literature underscores that decentralizing, and by the late 1990s most decentralization is not a panacea for poor service countries served by the Bank were undergoing delivery. As the 2004 WDR notes: "If decentraliza- some form of devolution, delegation, or tion just replaces the functions of the central deconcentration of administrative, political, or ministry with a slightly lower tier of government fiscal authority and responsibility. (a province or state), but everything else about the environment remains the same--compact, Countries have embraced decentralization for a management, and client power--there is little range of reasons. In Eastern Europe and the reason to expect positive change." Thus, whether former Soviet Union, decentralization was part of and how to decentralize or not is a question that the political and economic transformation that can be answered only against the background of came with the end of dominance by a country-specific contexts and institutions. highly centralized Soviet Union. In Most of the Bank's client South Africa, Sri Lanka, and Indonesia, countries have This said, many countries decide to decentralize decentralization was also part of a decentralized. for political rather than economic reasons. strategy to dampen ethnic or regional Decentralization of power is seen as an instru- discontent and conflict. In Chile, Uganda, and ment for absorbing regional and ethnic conflicts Côte d'Ivoire, decentralization was explicitly (von Braun and Grote 2000; Brosio 2002). pursued to improve delivery of basic services Decentralization is also seen as encouraging civic (Shah 2004). participation because it allows communities to be consulted over matters that affect their daily Each country has pursued its own form and pace lives (Katsiaouni 2003). The main issue before of decentralization. Some countries have moved development practitioners in such cases is not gradually, often stopping or reversing the process 5 DECENTRALIZATION IN CLIENT COUNTRIES over the years or moving in "fits and starts." usually not based on a constitution, and the Experience in Albania, Peru, the Russian Federa- central government maintains the right to tion, and Uganda fits this pattern. As indicated change the responsibilities and powers accorded earlier, only a few countries have opted for a "Big to lower levels of government. Bang" approach. Among these is Ethiopia, which rapidly devolved substantial resources from Nevertheless, the distinction between federal central agencies to regions in 1993 (and only a and unitary states is blurred in some countries. few years later decentralized from regional to Uganda, for example, is a unitary country; local levels), and the Philippines, which started in nevertheless, local governments and elected 1991 to devolve both political and expenditure local councils are established through constitu- responsibilities simultaneously to both regional tional provisions and laws. and local governments and to transfer govern- ment staff from the center to these subnational Overview of Bank Support for governments. Decentralization Decentralization has differed between federal Lending volumes and trends and unitary states. Although the latter can have Bank support for decentralization has more different tiers of government among which often than not been embedded in multisector or responsibilities and powers are shared, in federal sector loans, credits, and grants; few operations countries, such as Argentina, Brazil, Ethiopia, have been devoted solely to decentralization. As and Russia, the division of responsibilities and a result, establishing the total amounts and powers is typically enshrined in the trends of Bank support for decentralization is Some countries have Constitution and cannot be unilater- very difficult (box 1.2).2 decentralized gradually, ally revoked or altered by the central often stopping or government. In unitary states, these Using the Bank's coding system to identify reversing the process. responsibilities and powers are projects in which decentralization is featured as Box 1.2: Identifying World Bank Support for Decentralization Poses Challenges The task team leader or a task team member assigns thematic assignment mean that the reliability of the system identifying Bank codes to every Bank activity that directly serves the Bank's ex- support for decentralization is not assured. Additionally, the sys- ternal clients. Every lending activity can be assigned up to 5 tem cannot report on how much of the total commitment may be themes (selected from a list of 65), one of which is attributed to support for decentralization, but can only point to decentralization. loans or credits where decentralization or municipal finance is one The code for decentralization falls under the auspices of the element (as coded by the task team). Public Sector Board and is defined broadly. For instance, it can ac- The evaluation reviewed the objectives of a large number of commodate any activity relating to "delivery of public services." lending activities in the 20 countries where support for decen- The intent, of course, is that any such activity should involve de- tralization was assessed in depth. The goal was to understand centralization in some form or fashion. The evaluation found cases whether they included support for decentralization as defined by where this would be difficult to argue. In addition, there are two the evaluation. Once these lending activities were narrowed down other codes that are the responsibility of the Urban Sector Board to 203 projects and programs, a more in-depth review of objectives and that cover decentralization-related activities: municipal gov- and components allowed estimates to be made of the commitment ernance and institution building and municipal finance. for decentralization in each activity. About one-third of the total was The multiplicity of codes and the inevitable subjectivity in their found to be specifically targeted to decentralization-related support. Source: IEG desk review. 6 BACKGROUND AND OBJECTIVES a theme or activity, the Independent Evaluation potential to increase the efficiency of service Group (IEG) found that total commitments delivery, but also with risks (box 1.3). under these projects (spread over 89 client countries) over the period fiscal 1990­2006 Bank approaches to decentralization amounted to $31.6 billion, 8 percent of total In the early 1980s, Bank support for decentral- World Bank commitments during that period. ization was mainly in the urban sector,3 starting For the most part, that lending was not uniquely in Latin America and spreading to other or even primarily in support of decentralization, Regions.4 In the late 1980s and early 1990s, the so it cannot be taken as a meaningful estimate of Bank focused on strengthening community Bank financing for decentralization. This evalua- participation in development planning and on tion estimates that about one-third of this sum increasing control over resources at the (about $10.6 billion) can be considered to have community level. In the mid-1990s, specifically targeted decentralization as defined after several failed adjustment No single network, sector for the purposes of this evaluation (see box 1.3). programs that led to deterioration in board, or thematic group Note that this includes both still-open and service delivery, the Bank began oversees decentralization, closed commitments (the distinction is an supporting delegation of the delivery and there is no overall important one). of social services, in particular in Bank strategy in support education, health, and water, to local of it. Bank strategy levels of government. Decentralization is a cross-cutting theme, so no one network, sector board, or thematic group Consequently, Bank support for fiscal decentral- within the Bank has clear leadership, and no ization and intergovernmental fiscal frameworks unifying decentralization strategy or guidelines broadened. In the late 1990s, the emphasis have been articulated. Instead, since the late shifted toward governance at intermediate 1990s, several sector boards have issued relevant government levels, focusing on fiscal reform, strategies. These strategies tend to treat including strengthening of financial manage- decentralization as a cross-cutting issue with the ment, procurement, and related capacity. More Box 1.3: Bank Strategies Reflect Multiple Perspectives on Decentralization The Bank's approach to decentralization underwent a shift in the In contrast, the 2007 health and 2005 education strategies rec- mid and late 1990s. It moved from neutral encouragement of ognize the trend toward decentralization in Bank client countries stakeholder participation in decision making concerning re- but are more cautious: they note the challenges and suggest a case- source allocation to more proactive support for bringing gov- by-case approach to relying on or supporting local governments. ernments "closer" to the people. For health, this reverses the approach of the 1997 strategy, which The Bank's current approach is a more cautious one that em- made decentralization one of three implementation pillars. phasizes the need for strengthened institutions and risk mitiga- Finally, the 2003 Water Resources Sector Strategy, the 2003 tion. The 2000 Public Sector and Governance Strategy portrays Forestry Strategy, and the 2001 Environmental Strategy point out the decentralization as one of eight key elements of public sector re- risks of decentralization. The first suggests that optimum water form. At the same time it recognizes the risks of elite capture and resource management may need to be at the national or even in- lack of capacity at the local level. The 2000 Urban and Local Gov- ternational level, and the last two note that decentralization pres- ernment Strategy and the 2003 Rural Development Strategy also ad- ents special challenges to natural resource management and vocate a strong role for the Bank in strengthening local governments. environmental regulation and management (see appendix A for The 2005 Social Development Strategy makes the case for linking the relevant passages on decentralization in each of the sector community-driven development and decentralization. strategies). Source: World Bank data and IEG staff review. 7 DECENTRALIZATION IN CLIENT COUNTRIES Support started in the recently, with increasing devolution of The evaluation focuses on intermediate out- urban sector and spread responsibility for service delivery by comes rather than attempting to assess final to projects involving several client countries, the Bank has outcomes for two reasons. One is paucity of data community participation, further enlarged and deepened its on final outcomes in Bank reports. The other is service delivery, and support to include strengthening of the challenge of attribution: even if data on capacity building. local government institutions and service delivery and governance at the local capacity. government level were reliably available, it would be difficult to assess whether such improve- Throughout the review period, the Bank ments were due to decentralization, let alone to advocated and supported fiscal delegation and support for decentralization provided by the administrative deconcentration, although it Bank, because of the many factors that influence encouraged some degree of central control and such outcomes. monitoring. Support for political decentraliza- tion covered a range of activities, including Figure 1.1 provides a framework to help trace empowering local citizens to design and how Bank input might help bring about final implement development activities, enhancing outcomes such as better service delivery and access to information on the use of public improved governance. As shown in the figure, resources, and building the capacity of citizens three intermediate outcomes (results) are to monitor service delivery. relevant: strengthened intergovernmental fiscal frameworks, enhanced administrative capacity, Evaluation of Bank Support to and increased accountability of subnational Decentralization governments. As noted at the beginning of this chapter, the term "decentralization" means different things to differ- In addition, IEG divided the review of results ent people. To bring in a sectoral perspective, the along two lines: Bank support to strengthen evaluation also reviewed operational activities for decentralization frameworks and Bank support decentralization in the urban, education, health, to improve service delivery. Bank instruments and water sectors in countries in which Bank and design of the support, as well as the stated support was assessed in depth. The evaluation objectives of the lending, were often quite differ- reviewed Bank support for decentralization in ent between these two lines. Lending in support education in particular depth (see chapter 4), of frameworks was frequently provided through benefiting from case studies undertaken for a development policy loans (DPLs) or economic recent IEG evaluation of Bank support for and sector work (ESW) and was aimed at public education (IEG 2006a). This evaluation covers two sector reform more broadly; lending for service key aspects of the Bank's support: its quality and delivery was more typically provided through its results. Appendix B details the evaluation investment and technical assistance projects and methodology. limited to sector-specific reform. For these reasons, this report discusses results separately, The evaluation addresses several questions regard- focusing first on the intermediate outcomes of ing the quality of support. It reviews whether the Bank support for decentralization frameworks support was based on a clear understanding of (chapter 3) and then on the intermediate country context, whether it outcomes of Bank support for service delivery in The evaluation focuses on appropriately reflected the specific the education sector (chapter 4). Bank support for circumstances associated with strengthening decentralization in those countries, The discussion of intermediate results focuses decentralization whether it was internally coherent only on those aspects that were supported by the frameworks and and results based, and whether it Bank and does not assess the full results of improving service was provided in collaboration with decentralization. It is possible, for example, that delivery. other development agencies. decentralization has been working well overall in 8 BACKGROUND AND OBJECTIVES Figure 1.1: Evaluation Framework for Assessing Results of Bank Support for Decentralization Inputs Outputs Intermediate outcomes Final outcomes Lending Fiscal: Rules for Improved legal and Better service revenue and regulatory frame- delivery Economic and expenditure and works for fiscal sector work borrowing relationships and Improved service delivery governance Other nonlending Administrative: support Local control over Improved human resources, administrative budgeting, financial capacity management Better upward and Political: Citizen downward participation accountability Exogenous factors Source: IEG. a country, even if the specific intermediate contained some sort of World Bank support for outcomes associated with Bank support have decentralization. been less satisfactory, or vice versa. As with all IEG evaluations, the findings on the results of Figure 1.2 shows the annual commitments of Bank support therefore cannot be generalized to World Bank financial support related to a country's overall effort. decentralization in the 20 countries, with a cumulative total of about $10.2 billion. The To capture Bank support in different country majority of these projects are now closed, but contexts, IEG selected 20 countries from the set figure 1.2 also includes still-active commitments. of 89 that were coded as receiving Bank support To assess the quality and results of Bank support, for decentralization, municipal governance and the evaluation considered only 203 lending institutional strengthening, or municipal finance. activities that closed before June 30, 2007 (about Given that decentralization is a process that takes 30 percent of which were DPLs, the remainder significant time to achieve maturity, the evalua- being investment projects). The commitments tion reviewed Bank support for decentralization associated with the parts of these activities that to these 20 countries between fiscal 1990 and related specifically to decentralization total $7.4 2007. The set of countries reviewed for the billion. In addition to the closed activities, a few evaluation includes countries representing all open activities were reviewed to see if there were Regions, with both small and large populations, any discernable shifts in the quality of Bank and exhibiting eligibility for both International support for decentralization. Bank for Reconstruction and Development and International Development Association funding, Chapter 2 contains the evaluation's findings on the federal and unitary systems, Big Bang and gradual quality of Bank support for decentral- approaches, as well as past or present conflict- ization to the 20 countries. To assess The evaluation assessed affected status (see table 1.1 for a typology of the quality, all relevant documents, includ- Bank support to 20 20 countries). ing Country Assistance Strategies countries that received (CASs), ESW, and lending documents about half of all Bank The commitments of financial support to these (appraisal through implementation funding with 20 countries represented 47 percent of the completion for the 203 closed projects) decentralization- Bank's total commitment to the 89 countries that were examined. A three-point scale related activities. 9 DECENTRALIZATION IN CLIENT COUNTRIES Table 1.1:Typology of Countries Where Bank Support for Decentralization Was Reviewed in Depth Recent effort Type of Subnational government Unitary or at strengthening administrative dependent on transfers Type of Country federal decentralization decentralization for >75% of revenue decentralization Albania U 2000 Devolution Y Gradual Bolivia U 1994 Deconcentration Y Gradual Brazil F 1988 Devolution N Gradual Burkina Faso U 1998 Deconcentration Y Gradual Ethiopia F 1993 Devolution N Big Bang India F 1993 Deconcentration Y Gradual Indonesia U 2000 Devolution Y Big Bang Madagascar U 2003 Deconcentration Y Gradual Morocco U 1992 Deconcentration N Gradual Nepal U 1999 Deconcentration Y Gradual Nicaragua U 1988 Deconcentration Y Gradual Pakistan F 2001 Devolution Y Big Bang Peru U 2003 Deconcentration Y Gradual Philippines U 1991 Devolution N Big Bang Russia F 1999 Devolution N Gradual Rwanda U 2000 Deconcentration Y Gradual Sierra Leone U 2004 Deconcentration Y Gradual Tanzania U 1997 Deconcentration Y Gradual Uganda U 1993 Devolution Y Gradual Yemen, Rep. of U 2000 Deconcentration Y Gradual Source: IEG desk review. Note: All local governments established through elections. F = federal; U = unitary. (high, medium, and low) was used to rate the Given the small size of the sample, the assess- quality of Bank support. ment is aimed only at providing some insights into the issues, strengths, and weaknesses that For decentralization frameworks (chapter 3), have characterized at least some Bank work on results were assessed in the 20 countries listed in decentralization related to service delivery. table 1.1. A four-point scale (high, substantial, modest, and negligible) was used to rate the To assess the results of Bank support at the results of Bank support, and a before-and-after country level in addition to reviews of Implemen- methodology was used to assess progress at the tation and Completion and Results Reports country level. (ICRs) and other Bank self-assessments, IEG undertook field missions to 8 of the 20 The evaluation assessed For service delivery (chapter 4), IEG countries--Burkina Faso, India, Madagascar, whether Bank support assessed Bank support for education Peru, the Philippines, Russia, Tanzania, and helped improve services in a subset of six countries. Uganda--to gather additional evidence regard- decentralization These were selected based on the ing the decentralization efforts. Additionally, frameworks in 20 amount of support provided by the IEG's Country Assistance Evaluation (CAE) countries and service Bank and the availability of evaluative mission to Indonesia and other recent CAEs on delivery in education in 6. information from other IEG reports. Albania and the Republic of Yemen provided in- 10 BACKGROUND AND OBJECTIVES Figure 1.2: Bank Support for Decentralization to 20 Countries, Fiscal 1990­2007 (includes still-open activities between 2000 and 2007) 1,200 s)n 1,000 (millio n 800 alizatiortn 600 decer fotn 400 200 commitme kn Ba 0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 Source: World Bank database. Note: This includes total commitment of $10.22 billion specifically targeted at decentralization activities (that is, all projects with a decentralization, municipal governance, or municipal finance theme). depth information on decentralization issues. survey on donor collaboration to the participants. Several IEG Project Performance Assessments The results of that survey are integrated into this Reports also deepened the understanding of report. With financial support from the Norwegian results. Agency for International Development (better known as Norad), two participatory field assess- IEG participated in a workshop on decentraliza- ments in the Philippines and Uganda were tion organized by the Organisation for Economic undertaken to bring the voices of client country Co-operation and Development­Development citizens into the evaluation in a systematic way. Assistance Committee (OECD-DAC) Informal Other stakeholders consulted during missions are Working Group on Decentralization and Local listed in online appendix D (see http://world Governance in September 2007 and issued a joint bank.org/ieg/decentralization/download.html). 11 Chapter 2 Evaluation Essentials · The quality of Bank support for decentralization has improved sig- nificantly in the past five years. · In countries that have devolved responsibilities for some service delivery, Bank support for local gov- ernments is more recent and in most cases began only after devolution was well under way. · Analytical work is mostly sectoral and is typically not used outside the relevant sector. · Treatment of political economy issues in ESW is typically weak and--even where it exists--rarely influences recommendations. · Donor collaboration on decentral- ization is increasing, although several challenges remain; harmonization of financial management and procure- ment processes must be catalyzed to reduce the costs to local govern- ments of doing business with the Bank. Senegalese village center. Photo by Curt Carnemark, courtesy of the World Bank Photo Library. The Quality of Bank Support T his chapter details the results of IEG's assessment of the quality of Bank support, both analytical/advisory and financial, in the 20 countries over the evaluation period. The assessment reviewed Bank support to un- derstand whether it was based on a clear understanding of country context, whether the support appropriately reflected the circumstances in these coun- tries and the Bank's own diagnoses of decentralization issues, whether the sup- port was internally coherent and results based, and whether the support was provided in collaboration with other development partners. The assessment is based on a desk review of all The Quality of Bank Analytical Work Bank country strategies, relevant ESW on on Decentralization decentralization, and all World Bank financing IEG found decentralization-related diagnostic (projects and fast-disbursing loans) in support of reports for 16 of the 20 sample countries (see client countries' public sector as well as of the online appendix E).1 Of the remaining four urban, health, and education sectors. In the eight countries, analytical work is ongoing in three. In countries where field missions were conducted, a few countries, including Uganda and Sierra the findings of the desk review were validated Leone, even though there was no analytical work through interviews with relevant client country prior to Bank financing, the operational work stakeholders as well as bilateral and multilateral was undertaken in collaboration with other development agency representatives. development partners, based on successful United Nations Capital Development Fund The quality of Bank support for decentralization (UNCDF) pilot projects.2 during the entire evaluation period was of medium quality, but there was significant In about half of the sample countries, the Bank improvement in the last five years of the evalua- undertook analytical work within five years of a tion period, when the quality of Bank work was decentralization law being passed. In the other high in about two-thirds of the 20 countries. Still, half, the Bank neither anticipated nor examined better coordinated and more comprehensive the implications of the decentralization law for approaches are needed in about one-third of the key areas of the Bank's work within the five-year sample countries where the quality challenge period (see table 2.1). In Pakistan, timely and endures. extensive ESW, some of which was undertaken 15 DECENTRALIZATION IN CLIENT COUNTRIES Table 2.1:Timeliness of ESW in 20 Sample Countries ESW undertaken Percent of cases Within two years after decentralization law was passed 20 2­5 years after decentralization law was passed 35 5­10 years after decentralization law was passed 20 No completed ESW found that dealt substantially with decentralization 25 Source: World Bank data. jointly with other major development partners, decentralization, but the sector perspective meant influenced financial support from the Bank for that major cross-cutting issues of local govern- decentralization in the country. In Brazil, an in- ment accountability and capacity were not depth study of the challenges of decentralization addressed (for instance, in Brazil, Burkina Faso, to the municipal level fed into the design of Bank Ethiopia, and the Philippines). In India and support for municipalities. Morocco, a rural-urban segmentation of the analysis was predominant, resulting in the During the 1990s, During the 1990s, the concentration of (sometimes virtually simultaneous) preparation analytical work was analytical work was on sector-specific of separate reports focusing on rural and urban mostly sectoral in nature, aspects of decentralization, depending local governments, respectively. and the type of work on the Bank unit that undertook the done often depended on analysis. Such analyses influenced Until recently, the treatment of political economy which unit did it. subsequent Bank interventions in the issues in most ESW was weak. Most pieces did sector in question. Beyond this, the not discuss the roles of important interest analyses had very limited influence in that they groups and different public officials. Insofar as did not lead to a holistic understanding of political economy factors were discussed, they decentralization and its implications or of cross- tended to be portrayed as post hoc explanations sectoral synergies. As a result, they were unable for past government policy failures. Most of the to catalyze the formulation of coherent country- earlier ESW did not anticipate the responses of level decentralization support strategies, even important interest groups (such as teachers and under circumstances where the client countries public sector unions) or of public sector officials had devolved service delivery to local govern- to proposed institutional reforms. ments across multiple sectors. Although the Bolivia Institutional and Gov- In Peru and Russia, the analyses initially concen- ernance Review (World Bank 2006b) and the trated on the fiscal aspects of decentralization and Madagascar study on decentralization (World paid less attention to the close linkages with Bank 2004c) stand out for their treatment of many aspects of broader public sector reform. In political economy issues, the Tanzania study on some countries, such as Nepal and Sierra Leone, fiscal decentralization (World Bank 2006k) and the Bank studied the implications of decentraliza- the study on Pakistan's reform of provincial tion through the lens of Public Expenditure finances in the context of devolution (World Reviews, which tended in some cases to highlight Bank 2000b) offer only limited treatment of such expenditure issues without linking them to issues. Additionally, in fewer than half of the 16 revenue-side issues, notably the resources countries for which decentralization-related needed to implement devolved spending diagnostic reports could be found, the ESW responsibilities. examined the fiscal costs of decentralization for the various levels of government. As a result, Sector analyses in health and education frequently recommendations were often prefaced with had at least some treatment of the implications of general headings and discussions of "interna- 16 THE QUALITY OF BANK SUPPORT Box 2.1: Good Practice ESW in Bolivia The Bolivia Institutional and Governance Review Towards an In- enue mobilization and otherwise strengthening incentives that clusive Decentralization (World Bank 2006b) is a good-practice local officials face in contexts marked by dramatic inter-regional ESW that provides guidance for Bank interventions in support and urban-rural income inequities. It buttresses the discussion of specific sectors in the country as well as for the design of a with detailed data and analysis relating to such inequities and CAS. Its strong features include the following: highlights trade-offs when discussing reform possibilities. 3. It recommends the most attractive ways of avoiding, or at least 1. It takes a comprehensive view of the decentralization process minimizing, risks associated with an ineffectively designed in Bolivia. It pays careful attention to potential problems of and politically motivated decentralization program. It rec- fiscal discipline--identifying them as most crucial--but also ommends a sequence of reforms and emphasizes trade-offs relates these to service delivery in key sectors, to broader gov- associated with each possible choice. ernance issues, and to accountability, all within a common 4. The review concludes that decentralization to regional gov- framework focusing on institutions and incentives. ernments in Bolivia is likely to be successful in improving ei- 2. It establishes trade-offs across goals. In contrast with the lim- ther service delivery or local governance under existing itations of a sector-specific report, the comprehensiveness of conditions. It sets out why the Institutional and Governance this review forces the authors to consider whether multiple re- Review arrives at this assessment and relates it to the recent form goals might be mutually incompatible. For example, the re- history of municipal decentralization and political economy view discusses the trade-offs between efficiency and equity, in Bolivia, as well as outcomes in other countries that faced especially in relation to the difficulties of enhancing local rev- similar circumstances. Source: World Bank documents. tional best practice." They were disconnected Although up to a point all ESW Discussion of political from the circumstances of the countries. discusses monitoring systems, this economy in most ESW has was done meaningfully in only seven been weak. Only in four countries (accounting for some 10 countries, where the ESW assessed percent of reports reviewed) did ESW meaning- the quality of existing systems, areas of fully review and assess Bank support for decentral- weaknesses, and the nature of data that needed ization and provide recommendations on suitable to be collected. The Madagascar ESW is good entry points, phasing of decentralization-related practice in this area and recommended that a reforms, and necessary preconditions for Bank monitoring system be established as a first-phase support. In the remaining countries, an opportu- activity before deepening decentralization. nity was missed for ESW to deepen its influence on the design of Bank support. More recently, decentralization has received more comprehensive treatment in 40 percent of the The Madagascar ESW (World Bank 2004c) is countries (Madagascar 2004; Pakistan 2004; Bolivia notable in that it identified conditions under which 2006; Russia 2005; East Asia and the Pacific 2006, the decentralization of certain activities might be covering Indonesia and the Philippines; Burkina desirable and proceeded to compare these Faso 2007; and Ethiopia 2007). The associated conditions systematically with the actual situation. reports seek to understand the important shift in This led to a set of phased recommendations, governance that decentralization has some of which involved centralization. The brought about and the implications for Only in four countries Tanzania report on fiscal decentralization (World the development agenda, including did ESW provide Bank 2006k) is good practice in terms of the clarity Bank strategies. Although the reports guidance to Bank staff on with which it provides recommendations for the address the issues from different angles preconditions or entry design of Bank support for decentralization. (for example, fiscal reform in Russia points for Bank support. 17 DECENTRALIZATION IN CLIENT COUNTRIES versus service delivery in Ethiopia), the analysis of local capture and improving local governments' decentralization is more comprehensive than in responsiveness to their citizens. There is little the past, with a focus on both the demand and evidence that this set of publications had supply side aspects of decentralization and their substantial influence within the Bank, although complex interaction. this evaluation found that other development partners were aware of them and considered the The majority of these reports were prepared by Bank as having significant technical knowledge multisectoral teams in collaboration with other and expertise in this area. development partners and with the participation of country stakeholders. This has enhanced the Coherence of Lending with Country quality of the analysis and ownership by the Priorities and Bank Analysis stakeholders. In Russia, for example, stakehold- The evaluation assessed the extent to which each ers interviewed for this evaluation noted the Bank lending operation was prepared jointly participatory manner in which the analysis of with the relevant client country, reflected decentralization was undertaken and were country-specific circumstances, and was rooted particularly appreciative of the technical in the Bank's own diagnosis of the decentraliza- knowledge that Bank teams brought to bear in tion process in the country. helping them design their local systems. (Examples in box 2.2 on page 21 illustrate the Participation of country stakeholders importance of sound analysis of country context Country stakeholder participation in the for the design of Bank support.) preparation of Bank lending was high or substan- tial in 14 of the 20 countries. Stakeholder partici- The influence of broader The Bank has also prepared broader pation was highest in the Africa Region and Bank analyses of analyses of decentralization issues. The lowest in the Middle East and North Africa decentralization is public face of this work was the Region. Stakeholder participation was also high unclear. decentralization page of the Poverty in Russia, where government and other Reduction and Economic Management stakeholders reported that they had led the (PREM) Public Sector Web site. In addition, project design, with the Bank bringing its techni- "Rethinking Decentralization in Developing cal knowledge, including a comparative perspec- Countries" (World Bank 1998b), issued by the tive on good practices across countries, to bear. PREM anchor, provided guidance on moving away from a normative approach to decentralization. Reflection of country-specific circumstances The "Decentralization Briefing Notes" (Litvack Weaknesses in the understanding and treatment and Seddon 1999), published by the World Bank of political economy risks in ESW and in project Institute in collaboration with PREM, contained 21 design have given rise to overly ambitious brief, well-focused papers on aspects of the ration- objectives in Bank lending. Bank lending ale for decentralization, project design, service documents typically acknowledge and address delivery, and potential impacts of decentralization. these risks. However, they also tend to assume Fiscal Decentralization and the Challenges of that the Bank can mitigate virtually any relevant Hard Budget Constraints (Rodden and Eskeland risk. Even serious risks (high-probability and/or 2003), a set of studies to guide Bank staff and high-loss events) are considered to be contained other practitioners, provided guidance on how to by various mitigation measures. avoid excessive debt accumulation as subnational units gained greater fiscal independence and In Peru, a better dissemination strategy was worked through the implications for how to expected to address risks associated with the decentralize (or not). lack of consensus on decentralization. In Pakistan, the threat of rapid disengagement by A final work, "Decentralization and Governance" the Bank was expected to mitigate the risk of (World Bank 2001a), guided staff on avoiding dwindling political will to carry out reforms. The 18 THE QUALITY OF BANK SUPPORT relative downplaying of risks in this manner Philippines--either the influence of Weak understanding of makes "stretch" objectives appear more credible ESW on Bank support is not evident political economy issues and realistic. or the ESW is too recent to assess led to overly ambitious whether it has had meaningful objectives. In contrast, overly ambitious objectives entail influence. lower achievements--hence reduced effective- ness--vis-à-vis the former. Additionally, in many Analytical work prepared by one unit generally cases, sector-level activities did not display had little influence on operations prepared by awareness of ongoing general fiscal reforms. The other units. For example, an analysis of fiscal Tanzania Human Resource Development Project decentralization and intergovernmental finances (approved in 1997) focused on facility-based in Albania (World Bank 1994a) pointed to the decentralization, transparent fiscal transfers on a lack of clarity in allocating responsibilities among per capita basis, and community participation; the different levels of government in the social this marginalized the role of subnational govern- sectors. It cautioned that Bank support for ments that a 1997 law on decentralization decentralization must first address this lack of provided for.3 clarity and investigate the benefits of decentraliz- ing noninfrastructure services in the social Influence of own diagnosis in design sectors. However, the 1998 Albania Health of Bank support System Recovery and Development Project In 12 of the 16 countries for which ESW was supported the decentralization of carried out, at least one report appears to have health services to the Tirana Regional ESW frequently pointed to influenced the design of one or more Bank Health Authority, thereby in effect measures subsequently lending operations. Most of these operations adding a layer of management costs supported by DPLs. were DPLs with specific reform measures without ensuring the autonomy ("conditions") derived from decentralization- necessary to realize the benefits. This aspect of related studies (Albania, Brazil, Burkina Faso, and the project has complicated decentralization in Pakistan). The influence of ESW on the design of the health sector and now calls for corrective investment projects was not as clear. action.4 The Tanzania decentralization study (2001e) Differing sectoral or thematic perspectives can provided the Bank with a basis for supporting give rise to inconsistencies (and ultimately to the local government program during a phase reduced development effectiveness) in a Bank when there was not yet a consensus among country program. In the Philippines, PREM development partners about whether and how studies (World Bank 1999b, 2004g) identified a to support decentralization. The municipal need to operationalize the 1996 framework for sector decentralization study in Brazil (1992a) Local Government Unit (LGU) financing, includ- contributed significantly to paving the way for ing transforming the Municipal municipal development loans in many states. Development Fund (MDF) into a full- Operations have tended fledged financial intermediary and not to consider relevant Two reports on Russia (one on the fiscal costs of requiring government financial analysis done by other reform and the other on regional-local reform) intermediaries (GFIs) to help LGUs units. contributed to the government's fiscal decentral- reduce their reliance on government ization reform plan for the medium term at the transfers.5 The PREM studies note that access to regional and municipal levels (World Bank private sources of capital remains largely 2005h, 2005i); they also supported the design of untapped because of the virtual monopoly of the Bank's Fiscal Federalism and Regional Fiscal GFIs and the MDF.6 Reform Loan. Regarding the remaining four countries for which ESW was carried out-- In contrast, a 2006 project by a regional urban Ethiopia, Morocco, Nicaragua, and the unit continues to support lending by GFIs to 19 DECENTRALIZATION IN CLIENT COUNTRIES local governments because, according to the related support to operational units is split project documents, private sector funding for among PREM, Social Development, Urban local government infrastructure is still limited, Development, and other sectoral units, the and increasing it can only be a long-term support is often fragmented in a field where a objective. A proposed urban sector project aims highly collaborative and coordinated approach is to strengthen the MDF (still nestled in essential for effectiveness. As a result, support to Inconsistencies in the Department of Finance) to client countries has not capitalized on the signifi- objectives across support the lower tier of LGUs.7 cant depth and breadth of expertise on the individual sector subject that exists within the Bank.8 interventions reduced In Morocco, a PREM report (World development effectiveness Bank 1992b) noted that the key issue One illustration of this is Nicaragua, where a at a country level. with municipal governments was not sector-level perspective encouraged the lack of funding, but lack of capacity. adoption of a law that introduced a transparent Yet the 1993 first MDF project provided funding system of formula-based transfers to municipali- for infrastructure loans to municipalities after ties. These transfers effectively contributed to dropping a significant capacity-building com- the equalization process and addressed an ponent because of (central) government existing unfunded mandate on municipalities' reluctance to use loan proceeds for the latter part. However, they also raised new issues of purpose. The second loan, in 1998, which contin- macroeconomic-level fiscal imbalance, particu- ued lending to municipalities, was also approved larly because municipalities did not have without any provision for capacity building commensurate expenditure responsibilities. The because of resistance from the government to transfers contributed mainly to increased borrow for capacity building. employment and salaries at the municipal level rather than investment in local priorities.9 Although each of the individual sector or thematic perspectives within a given Bank The Bank has since begun helping revise this country program may be underpinned by a result through DPLs, where corrective measures sound rationale, improved development feature among the reforms supported by these effectiveness requires that tensions between operations.10 As a Bank Policy Note (World Bank differing perspectives be resolved, ideally 2004e) rightly concludes, "Without an appropri- through a country-owned strategy or at least ate institutional framework and monitoring through a unifying framework for Bank system, hasty and aggressive decentralization can assistance. lead to waste of resources, a worsening of the provision of public services, and increased Internal Consistency and Results macroeconomic instability."11 Orientation Differences in ways of pursuing macroeconomic Internal consistency stability and equity have given rise to some Much of the Bank's support for decentralization inconsistencies across objectives in at least 6 of the appears to have seized opportunities as they 20 countries. For example, recentralizing the tax arose across sectors. However, the multiplicity of base can advance both equity and macroeconomic units that support decentralization, stability and reduce fiscal imbalances; however, Inconsistencies arose the differences in approach across the recentralization also reduces the fiscal autonomy between centralization of various units, and the diversity of of subnational units, thereby reducing their ability tax authority and objectives underlying Bank support to respond to local needs and priorities. decentralization of for decentralization have entailed expenditures, as well as uncoordinated support at the expense In Tanzania, PREM staff, together with the in human resources of development effectiveness. Because International Monetary Fund (IMF), supported treatment. responsibility for decentralization- the centralization of local revenue-raising author- 20 THE QUALITY OF BANK SUPPORT ity, and the urban unit supported the decentral- wage bill. At the same time, subnational govern- ization of expenditures. In Russia, revenue ments were under pressure to hire more staff to administration has been recentralized with Bank meet their newly assigned responsibilities. Not support, ostensibly to ensure macroeconomic surprisingly, neither wage containment nor stability and equitable allocation of revenue at efficient human resource capacity build-up were the regional level. However, this has negatively achieved under such circumstances. affected the revenue autonomy of several regional governments. In four countries, tax Differing approaches to decentralization have administration projects were implemented after led to tensions in several countries where Bank the passage of a decentralization law, but the support was not coordinated across the units projects dealt primarily with centralized revenue delivering the support. In the majority of the administration structures. country cases reviewed, a community-driven development (CDD) approach, as typically If well designed and well executed, administra- featured in Bank support (essentially a form of tive decentralization should involve the rational- deconcentration that transfers implementation ization and reallocation of human resources and responsibility to community groups), generated result in the enhanced efficiency of government some tensions. This happened in some services. The evaluation found that in a few situations where responsibilities for service countries, including Bolivia, Pakistan, and delivery and community participation had been Tanzania, macroeconomic adjustment programs devolved or delegated by law, often to newly required the central government to contain the created local governments. Box 2.2: An Incomplete Understanding Reduces Effectiveness of Support The Ethiopia Health and Education Sector projects (each $100 setting decentralization policies, along with two other national million, approved in 1999) were designed after a social sector re- coordinating bodies for decentralization, complicating the institu- view that did not fully consider decentralization issues and be- tional framework for decentralization (see World Bank 2004b, pp. fore a 2000 study of regionalization that undertook a more in-depth 16 and 20). review of the issues. Deficient project design and weak institu- In the Republic of Yemen, without prior ESW, the 2002 country tional capacity partly explain the weak outcome ratings. strategy planned to provide support for ambitious decentralization In Nicaragua, without prior analysis of the institutional frame- objectives under the assumption that a broad consensus had been work for decentralization, Bank support encouraged the govern- achieved around the decentralization agenda through the 2000 Local ment to require a transfer of resources to municipalities in the (not Government Law. Some analytical work initiated in 2002 was de- entirely correct) belief that the latter were subjected to unfunded livered in 2005 but not published because of disagreements with mandates. Now the challenge for Bank support to the Nicaraguan the government. A Learning and Innovation Loan to support de- authorities is to help bring about a transfer of expenditure re- centralization was appraised, but not carried through to approval, sponsibilities to match a predetermined level of resource transfers. again because of a disagreement with the country authorities on In Bolivia, without comprehensive prior ESW, both the first the approach. The subsequent CAS candidly noted that government and second Programmatic Structural Adjustment Credits (PSACs) commitment to decentralization was questionable. Meanwhile, (which incorporated 39 reform measures, 27 of them relating di- the United Nations Capital Development Fund (UNCDF) and other rectly to decentralization) failed to achieve their objectives, owing development partners in the Republic of Yemen have ongoing in part to ongoing conflict and in part to an incomplete under- decentralization-related activities in the country and believe that standing of the highly complex institutional setting. The Ministry the Bank-advocated social fund approach is undermining the ef- of Finance, the counterpart for the PSAC, is today an additional body fectiveness of local governments. Source: Desk review of Bank documents. 21 DECENTRALIZATION IN CLIENT COUNTRIES CDD can strengthen Although CDD can strengthen the planned and designed social programs and voices of local capacity of local governments to often created parallel community-level institu- communities but if not participate in development activities, tions to oversee and monitor services that were carefully designed can enhance community voice in local distinct from the village local government or undermine the role of development, and reduce risks of panchayat in charge of local development and local governments. political capture, it also relies in many community participation.13 cases on centralized allocation of resources12 and/or parallel community bodies for In the Philippines, after the ambitious decentral- implementation (see box 2.3). This can ization program of 1991, health services were undermine the local government's role in placed entirely under the authority of local development activities. That said, increasingly, governments, but the Bank supported centrally the approach to resource allocation and coordi- planned and designed health sector interven- nation with local governments in Bank- tions through project implementation units. This supported activities may be changing. is not a problem per se; in some industrialized countries, deconcentrated mechanisms have led Similar tensions have arisen where Bank to effective service delivery. Furthermore, in support was not fully aligned with the client many situations (for example, a concerted effort country's decentralization strategy and policies. to meet the Millennium Development Goals), In India, although Bank support for fiscal such a centralized approach may be necessary if discipline and reform at the state level helped local governments do not have the capacity or strengthen macroeconomic stability, its support the political will to allocate funding. However, in the social and rural sectors was not equally such a deconcentrated approach to implementa- consistent with the country's policies for tion can reduce the roles of local governments to deepening decentralization. During the evalua- whom delivery of service has been devolved or tion period, the Bank supported centrally delegated. Recent efforts have attempted to mitigate such Box 2.3: CDD and Local Governments tensions. In the Philippines, the Kalahi CIDSS in the Philippines project (2003) has supported more coordinated ways of working with local institutions. In this Interviews with local stakeholders of the Bank-supported Water Dis- project, allocation of resources is undertaken in tricts Development Project demonstrate the tensions between ap- a competitive manner by village-level govern- proaches that primarily involve community groups as implementers of ment units in participating municipalities. In local development activities and those that involve local governments. Tanzania, the second social action fund The Barangay (village government) officials commented that it was attempted to provide expanded, more explicit difficult to accept the "Bank condition" that the water and sanitation roles for district and village governments. These services should be managed by a community association and not by the efforts notwithstanding, development partners Barangay, because they felt that they have the mandate to provide these continue to perceive the fund as an opaque services under the Decentralization Code. The Barangay captain in instrument, largely because of the centralization Mabunao noted that the city and the Barangay had agreed in princi- of allocation decisions with the final oversight ple that after the project closed, the facility should be turned over to and responsibility resting in a project implemen- the Barangay. At the same time, the community association officers said tation unit in the president's office. Partners that the project increased citizen participation, accountability, and argue that this weakens the role of local govern- transparency. They felt that although the community had previously de- ments in a process of devolution. The Bank's pended only on the Barangay and municipal governments to provide 2005 Social Development Strategy (see appendix water and sanitation services, the citizens were now more empowered A), which calls for greater linkages between as co-owners of the facility, with a voice in management. decentralization and CDD, is likely to encourage Source: IEG field assessment of the Water Districts Development Project in the Philippines. greater coordination with local governments under CDD initiatives.14 22 THE QUALITY OF BANK SUPPORT Results orientation of Bank support Ensuring that these monitoring Inconsistencies can arise During the evaluation period, Bank operations mechanisms are properly maintained when Bank support is not have focused monitoring on fiscal decentraliza- and systematically reported on during aligned with country tion. They have tracked indicators such as project implementation will be the strategies. transfers to subnational governments or key to enhancing knowledge on the increases in own-source revenue. However, IEG links between decentralization and improved found that in 10 of the 20 country cases, the most service delivery. recent country strategies had broadened their focus to include assessments of the performance The recently introduced Public Financial Manage- of local governments and institutions. ment (PFM) Performance Measurement Framework15 contains two indicators (trans- At an operational level, about 43 percent of Bank parency of intergovernmental fiscal relations and decentralization-related lending activities in 11 public availability of information on resource countries during the evaluation period include at allocations received by service delivery units) that least one indicator to measure the results of can help monitor aspects of decentralized service decentralization (with some activities in Sierra delivery at a country level. These indicators will Leone, Russia, and Uganda representing good underpin an assessment of progress. practices). In the remaining nine countries, such monitoring indicators were not found. In the 11 This framework has also been adapted Supporting client countries, the indicators have been tracked in by development partners in preparing countries to monitor the more than three-fourths of ICRs, although local government fiduciary assess- progress and impact of several reports discuss achievements in a qualita- ments such as in Tanzania (World decentralization at the tive manner. Intermediate outcome indicators Bank 2006l). This is a good practice project level is important. are rare, and even when available they are vague, report led by the government of such as "number of well-functioning municipal Tanzania. It provides a candid assessment of councils." Neither appraisal documents nor ICRs fiduciary issues at the local level as well as quanti- define "well functioning." tative information on local government perform- ance. The Bank has also helped develop a A review of still-active projects shows that the checklist for assessing the distributional impact improvement in recent CASs is reflected at the of decentralization as part of an effort to improve operational level. Newer operations include poverty and social impact analysis of decentral- fiscal and administrative indicators such as the ization (Kaiser 2006). Strengthening and replicat- share of unconditional transfers and own-source ing such efforts can lead to improved monitoring revenue in local government budgets, the and evaluation (M&E) of local government number of municipalities with development performance. plans prepared in accordance with guidelines, and local staffing levels. Accountability indicators Collaboration among Development are also included, to track trends in voice and Partners participation (number of local authorities who Collaboration with development partners is hold regular community meetings with pub- increasing at the country level. A review of recent lished minutes, public access to relevant country strategies indicates that 55 percent information) and to track local government included high or substantial collabora- accountability (comparison of unit costs of tion among development partners. Active projects show infrastructure built by local governments with The strategies included a joint strategy improved attention to those of line agencies, transparent financial for decentralization as well as joint monitoring fiscal, management systems at the local levels). Most diagnostic and analytical work and administrative, voice, operations also set aside funds for household lending activities within a context of and accountability and beneficiary surveys to assess citizen percep- general alignment and harmonization dimensions of tions of service delivery. at the country level. At the operational decentralization. 23 DECENTRALIZATION IN CLIENT COUNTRIES Collaboration with level, about two-thirds of all activities government. This has helped harmonization at development partners has have been prepared with high or the subnational levels also. In Brazil, an included joint strategy substantial donor collaboration. agreement has been reached among the Inter- development, analytical American Development Bank (IADB), the World work, and lending. Mechanisms for collaboration on Bank, and the government on the use of country decentralization support are increas- systems for financial reporting. Implementation ing and include both formal and informal of this agreement still lies ahead. working groups. In the Philippines, the Bank has been instrumental in initiating the Philippines Although Bank instruments such as Sector-wide Development Forum, a government mechanism Approaches have sought to harmonize financial for facilitating substantive dialogue on the reporting and procurement procedures at country's development agenda. It has also various sector levels, these efforts do not helped facilitate a working group devoted to necessarily lead to improved harmonization at decentralization and local governments, which the local levels. In Tanzania, each local govern- has reportedly given local governments a voice ment continues to have multiple bank accounts through their associations. for each sector project or program, and sometimes separate ones within each sector to Collaboration with In several countries, including accommodate the reporting obligations of the development partners in Tanzania and Uganda, significant various development partners. The shift to decentralization has diagnostic work is now undertaken greater use of country systems could help increased in the majority jointly with the development of the reduce such transaction costs for Bank client of the evaluation Public Expenditure and Financial countries. countries. Accountability (PEFA) PFM checklist and the OECD-DAC tool for assess- Overall, the evaluation finds that the quality of ment of national procurement systems. These Bank support for decentralization was of tools have been adapted to cover local levels and medium (neither high nor low) quality during are applied jointly by the government's key the evaluation period but has improved signifi- development partners.16 Joint assistance strate- cantly since the early 2000s; recent quality has gies in several of the 20 cases are helping been high in two-thirds of the 20 countries. establish a strategic framework for assistance, leading to joint missions and assessments, and Recent Bank analytical work is leading to a better supporting clients with the development of understanding of the broad implications of common measurement indicators. However, it is decentralization for service delivery and clear from donor responses to a questionnaire governance and is in turn influencing the design and other reviews by donor groups that develop- of country strategies as a whole. At an ment partners need to deepen collaboration on operational level, support for greater clarity in support for decentralization (box 2.4).17 decentralization policies is typically combined with technical assistance to strengthen the This evaluation also finds that harmonization of capacity and accountability of different govern- development partners' financial management ment levels (and the linkages between them). requirements and procurement practices at the Finally, the Bank is getting better at ensuring local level has generally been slow in the 20 substantial client participation and ownership countries. Nevertheless, there are among the different levels of government and Collaboration among noteworthy examples of progress. In among citizens in the design of assistance and development partners is the Philippines, major development support to monitor the results of decentraliza- improving, but partners harmonized their procure- tion. Collaboration with other development harmonization is ment guidelines and financial report- partners is improving but is still not consistently progressing slowly. ing requirements with the central effective. 24 THE QUALITY OF BANK SUPPORT Box 2.4: Development Partner Views on World Bank Decentralization Performance Seven development partners responded to IEG's questionnaire willing to work with other bilateral partners and that it sometimes on coordination (see below). Most appreciated the substantial funded conflicting programs and developed parallel systems. technical competence and experience of Bank staff and their They also remarked on the tendency of the Bank and other presence in countries. They also appreciated the influence of multilateral agencies, particularly in Asia, to segregate Regions the Bank's financial resources on national policies and its abil- or sectors. They also opined that Bank staff are "politically ity to get the country government to focus on critical issues. naïve" and do not have a historical perspective essential in However, respondents noted that the Bank was often un- this area. Question High­substantial (%) Modest­negligible (%) Extent of your agency's collaboration with Bank in supporting decentralization 29 71 Extent of your agency's collaboration with other partners 71 29 Extent to which Bank approach is consistent with that of your agency 100 Extent of satisfaction with Bank approach to decentralization 33 66 Source: Responses from seven partner agencies to IEG questionnaire (2007). The lack of a unifying strategy or framework systems that span different sectors at Overall, the quality of within which to address decentralization is the local level. There is a need for Bank support was of evident in Bank support for local governments. high-level leadership and oversight medium quality for the Support for decentralization through individ- of the process of designing Bank evaluation period, but it ual sectors in general is not sufficient to support to ensure greater coherence has improved in the last strengthen institutions and accountability and consistency. 15 years. 25 Chapter 3 Evaluation Essentials · Decentralization frameworks im- proved in several respects following Bank support--substantially in 7 countries, modestly in 12, and neg- ligibly in 1. · The most successful aspects of Bank support for decentralization frameworks pertained to the legal frameworks for intergovernmental relations, the frameworks for inter- governmental fiscal transfers, and subnational financial management. · Bank support was less effective in clarifying the roles and responsibil- ities of different levels of govern- ment and in improving own-source revenue mobilization by subnational governments, often because of lack of political commitment. · Bank support brought better results when there was consensus around the reform within the country prior to Bank engagement and when the support was combined with incen- tives for local institutional reform. Outside of the Parliament Building in Cape Town, South Africa. Photo by Trevor Samson, courtesy of the World Bank Photo Library. Bank Support for Decentralization Frameworks and Subnational Government Capacity A sound decentralization framework at the country level--defined through- out this evaluation as the full set of institutional arrangements within which decentralization is undertaken and sustained, specifying, among other things, how service delivery responsibilities and resource mobilization authority are apportioned among the various levels of government, as well as the associated accountability arrangements--must cater to a wide range of goals, from macroeconomic stability to equity. The framework needs variously (depending on the such as UNCDF, the Danish International Develop- country context) to ensure clarity in assignments, ment Agency, and the U.S. Agency for International absence of unfunded mandates, reasonable stabil- Development (USAID) supported decentraliza- ity in revenue assignment, effective budgeting and tion, often starting before the Bank did. financial management at both the central and subnational levels, prudence in subnational In all the cases with high or substantial achieve- borrowing and debt management, and adequate ment of targeted results, Bank support was accountability of subnational governments. combined with strong government commitment to reform. These caveats notwithstanding, Even where Bank support is significant, attribution interviews with government stakeholders and of results to that support is a very challenging development partners during field visits suggest exercise. First, results are influenced by many that the Bank's superior knowledge of variables, including local institutional capacity, decentralization--based on the diversity of its political factors, and aid dependence. These experience in different country contexts, its variables are themselves difficult to measure. ability to influence governments to seek policy Second, the Bank is only one of many develop- reform, and its ability to provide attractive levels ment partners that provide support; in several of of funding--were all important for achieving the 20 countries, other development partners progress. 29 DECENTRALIZATION IN CLIENT COUNTRIES Recognizing these challenges, this evaluation Strengthening Intergovernmental limited itself to assessing intermediate outcomes Relations in areas where the Bank has provided support in Bank support to strengthen intergovernmental these 20 countries.1 The intermediate outcomes frameworks led to high or substantial progress in were grouped in two broad categories: strength- the achievement of intermediate outcomes in 8 ened intergovernmental relations and improved of the 20 countries (table 3.1) and modest administrative capacity and accountability, each progress in another 9. The basis for this finding is with three or four subcategories (see table 3.1). explained below. Given the varied starting points and contexts in each country, the evaluation assessed progress in Strengthened legal framework these dimensions, using a before-and-after The development literature highlights the analysis. importance of a strong legal foundation to Table 3.1: Assessing Results of Bank Support for Strengthened Intergovernmental Relations Receipt of Enhanced framework Strengthened more timely and for own-source Improved framework regulatory predictable fiscal revenue/discretionary for prudent borrowing Quality of framework for transfers by resources of local and debt management results decentralization subnational governments governments at subnational level High · Key laws or regulations · Improved institutional · Improved framework · Rules for prudent borrowing related to decentraliza- framework in place for in place to enhance and debt management tion or intergovernmental intergovernmental own-source revenue strengthened and being effec- fiscal framework (such as fiscal transfers · Increasing availability tively implemented at the sub- fiscal responsibility and · Subnational governments of discretionary national level subnational borrowing) receiving more timely resources of sub- · Improved access to prudent passed or issued and predictable funds national governments borrowing · Being implemented Substantial · Critical laws supported · As above, but subnational · Improved framework to · Framework for prudent bor- by Bank passed but governments not yet enhance own-source rowing established, improving implementation not receiving transfers in a revenue or own-source debt management, although begun fully predictable or timely revenue of local govern- little or no improved access to manner ments increased as a prudent borrowing result of Bank support, likely to be replicated and sustained in the future Modest · A few subsidiary laws · No institutional framework · Some initial outputs in · Clear rules for debt manage- passed, but critical areas in place, but donor funds terms of framework or ment, but debt management still unregulated and being sent through piloting of mechanisms, not improved and no change in subject to political or transparent formula but no substantial access to prudent borrowing executive discretion improvement Negligible · Some policy dialogue · No difference in institu- · Reduction in own-source · No rules for debt management but no output tional framework, and no revenue, and no frame- formulated, and no increased increase in timeliness or work in place despite access to prudent borrowing predictability in receipt several years of policy of funds dialogue and/or support Source: IEG desk review. 30 BANK SUPPORT FOR DECENTRALIZATION FRAMEWORKS AND SUBNATIONAL GOVERNMENT CAPACITY ensure permanence and predictability of the condition for legal amendments that Assuring legal rights to institutional arrangements under which sub- should have clarified responsibilities information is a first national governments function (see, for instance, across levels of government. In step to enhancing Ahmad et al. 2005; Bird and Vaillancourt 1994; Nicaragua, although laws have been transparency. Boadway 2006; Shah 1998). passed with Bank support, key aspects need to be amended. The legal framework for decentralization improved to substantial or high degrees in 12 of Even when the government passed legislation to the countries studied. The piloting of implemen- clarify the rules for decentralized governance, tation mechanisms through Bank-supported the evaluation found that the effectiveness of projects led to the significant revision of such laws was enhanced if the Bank simultane- decentralization laws in Burkina Faso and the ously supported implementation of these laws at formulation of legislation in Rwanda.2 Legal local levels by providing fiscal and other changes, supported by the Bank through DPLs, incentives for institutional and behavioral helped improve intergovernmental relations by changes in the various sectors and levels of deepening or fine-tuning fiscal decentralization government. frameworks (Ethiopia, Russia, and Uganda) or by establishing laws for fiscal responsibility and Timeliness and predictability of transfers discipline at the subnational levels (Brazil, The Bank supported improving intergovernmen- Andhra Pradesh and Karnataka [India], Pakistan, tal fiscal transfer frameworks in 19 of the 20 and Peru). In Ethiopia, Bank support contributed countries. Bank support contributed to substan- to clarifying the legal framework for functional tial results in seven countries and modest results assignment in woredas and municipalities.3 in another five. The Bank has also supported laws that have Bank support helped governments Support for legislation is improved access to public information at develop formula-based transfers in not enough by itself; the subnational levels, thereby increasing the Ethiopia, Indonesia,6 Russia, Rwanda, Bank must help to accountability of governments to their citizens. Pakistan, Uganda, and Sierra Leone. For provide incentives to Several states in India passed right-to- example, in Russia (see box 3.1), reform enforce the laws. information legislation with Bank support. Initial has furthered the evolution of a clear analysis suggests that commercial and govern- and transparent system of transfers to regions; the ment stakeholders are using these right-to- government uses formula-based equalization information systems more often than ordinary grants, earmarked financial assistance for regional citizens are, but with increased awareness and finance reform, and capital grants that are distrib- capacity this measure can be a powerful instru- uted competitively and require cofinancing.7 ment in helping shed light on local government decision making.4 In Uganda, although a formula has been developed and applied, most development In other countries, the new laws had only modest funds are still earmarked by the central govern- effects because of reduced government commit- ment. In response to increasing concerns about ment to implementation or lack of consensus on the limited autonomy of local governments, with the way forward. For example, in Tanzania, signif- Bank support the Ugandan government is icant policy dialogue has only partly resolved implementing a pilot that gives 15 districts the conflicts between various legal enactments. In flexibility to reallocate recurrent conditional Albania, the Law of Prefects was passed with grants within and between sectors.8 Bank support, but several other critical laws essential for establishing the decentralized In Pakistan, with Bank support the federal and framework were not passed during the evalua- provincial governments established a fiscal tion period.5 In Bolivia, the Bank waived a loan framework to finance the district governments 31 DECENTRALIZATION IN CLIENT COUNTRIES In other countries, serious problems remain in Box 3.1: Good Practice in Russia the design of transfers. For example, mining resources in Peru are distributed based on the Decentralization reform efforts supported by the Bank during the late origin of the resource, which results in inequitable 1990s (including with three sector adjustment loans) were not suc- distribution. Also, municipalities in Nicaragua cessful, given the political volatility at the time. But a second set of Bank receive transfers but not clear or commensurate operations had better results. assignment of expenditure responsibilities. The Bank first helped the government prepare the Fiscal Federalism Strategy of 2001 and supported the enactment of two laws to underpin the In five countries, outcomes of Bank support have reform: the Law on General Principles of the Organization of Government yet to materialize, in part because of the lack of a in Subjects of the Russian Federation and the Law on Local Self- clear framework. Although the Bank began Government. The Bank also provided technical assistance and financed supporting urban and rural decentralization in a system of performance-based grants that encouraged fiscal reforms by Burkina Faso in the 1990s, that support has not regional governments. led to the establishment of a clear framework for This appears to have resulted in greater clarity in the assignment intergovernmental fiscal transfers given the system, instituted a formula-based transfer mechanism that allocated more country's predominant focus on deconcentration. funds to poorer regions, and provided incentives to subnational govern- As of June 2007, transfers to local authorities ments to reform their fiscal systems. These systems were later extended continue to be discretionary and based on a to government funds at the regional level, triggering significant reform of simple measure of population, which gives dispro- regional fiscal management as well as of the regional-to-local fiscal portionate advantage to the biggest and richest transfer system. cities (World Bank 2007a). The results are similar By 2003, about one-third of the Russian regions were using the formula- in Madagascar (World Bank 2004c), where the based transfer system and receiving funds in a timely and predictable man- policy direction may become clearer-- ner. The status of federalism in Russia, however, remains unclear because decentralization legislation was passed in 2004-- of some recent political decisions, such as the presidency taking on the but a system of transfers has not yet been authority to appoint regional governors. instituted. In the Philippines, support was Source: Field visit to Russia (April 2007). provided mainly through ESW (ADB and World Bank 2004; World Bank 1994b)9 and policy With Bank support, one- and facilitate the regular flow of funds. dialogue but has not yet led to discernible results. third of the sample This has made more resources countries have available at the subnational level. Transferring large amounts of funding to munici- established formula- palities and local governments without a clear based transfers. In 7 of the 19 countries, allocations of strategy and without development of local government funds continue to be either capacity can reduce accountability. The evalua- discretionary or opaque. For example, in Albania tion found that it is important to support local the central government introduced a formula, governments with technical assistance to developed with Bank support, that sets the amount enhance capacity and accountability institutions. of state resources to be allocated unconditionally to In some countries, Bank support has helped the local governments. The bulk of the develop- create incentives for fiscal and administrative ment funds, however, continue to be distributed as reform at the local level, and this constitutes conditional transfers that are arbitrary and negoti- good practice (see box 3.2). ated. In other countries, progress is limited to a few sectors or to donor funds. In Tanzania, formulae for Additionally, such transfers should avoid the risk transfers are sector specific; more than 80 percent of softening the budget constraints that have of the funds are earmarked, and the been built up in several countries. Equity consid- Critical elements of effectivetransfersofthebudgetedfunds erations are also important in determining the transfer systems are yet to to district governments are not always nature of transfers. In Bolivia and Russia, assess- be addressed and remain timely or predictable (World Bank ments have found that shifting more and more discretionary or opaque. 2006k). resources to departmental and municipal 32 BANK SUPPORT FOR DECENTRALIZATION FRAMEWORKS AND SUBNATIONAL GOVERNMENT CAPACITY governments without equity considerations can revenue. In the states of Bahia and Lack of client consensus lead to disparities in service delivery. Parana (Brazil) and in Nicaragua, Bank on direction has some- support helped enhance municipal times made change more Own-source revenues own-source revenues through com- difficult. The literature on decentralization, including puterization of records, rationaliza- Bank ESW, suggests that when local governments tion of tax rates, and increased efficiency of tax raise their own revenue, they will be more collection. In Brazil, the current transfer system autonomous from the central government and to municipalities, however, has been identified more likely to be responsive to their citizens for as reducing states' incentives to collect own- quality and efficiency in service delivery (Ter- source revenue.12 Minassian 1997, p. 49; Taliercio 2005).10 Bank support has aimed to increase own-source In Russia, Bank assistance supported It is important to provide revenues at the local level in the 20 countries that the centralization of revenue adminis- incentives for fiscal, were the focus of this evaluation, through policy tration. This has helped to improve financial, and admin- dialogue, DPL conditions, or project support.11 intergovernmental fiscal frameworks istrative reform at the Desired results have, however, been difficult to through a more equitable allocation local level before achieve. (See figure 3.1 for data on distribution of resources among the regions. Since transferring funds. of resources in a set of six countries.) 2005, both tax assignment and transfer formulae have been refined in the The evaluation found no single best approach for Budget Code. This made the system more stable addressing this issue. Bank support has varied in and transparent, although regions can no longer the five countries where it has contributed to tailor the intergovernmental system to suit local improving the framework for own-source conditions and have reduced incentives for Box 3.2: Strengthening the Role of Local Governments In about half of the country cases, including Karnataka (India), stitutional strengthening. For example, the program created an op- Russia, Rwanda, Sierra Leone, Tanzania, and Uganda, the Bank portunity for the first-ever audit of lower local governments (below provided support directly to local governments to facilitate the the district governments), which required the preparation of final ac- implementation of decentralization laws and to initiate institu- countsandidentifiedsomefinancialmanagementproblems.Thefield tional reform through demand-based capacity-building pro- assessment found, however, that the regularity and scope of local grams. Typically, such support provides development block audits has decreased because of a lack of funds after the project grants from the central government (through the budgetary closed. system) to local governments for physical investments for man- Access to information has been increased. Central govern- dated services that are their responsibility. Access to funds is ments receiving support are required to publish the amount of contingent on a minimum set of institutional, financial, and op- funds transferred to local governments in newspapers, and local erational criteria. Local governments that do not meet these cri- governments are expected to post the amount of funds received teria can have access to capacity-building grants to upgrade on local council bulletin boards. Field assessments confirm that their institutions. Such an approach of direct support to local gov- such measures have helped enhance transparency, with citizens ernments is critical for local service delivery and is now reflected monitoring financial information on budget releases (how much and in the CASs of about half of the study countries. when). Although most projects are still open and it is therefore too early Bank support under this project has also supported the devel- to assess results, the experience in Uganda suggests that the opment of a cross-sectoral monitoring system of local government program has helped increase the number of district governments performance. However, the indicators will need to be strengthened meeting the minimum capacity standards, leading to substantial in- during the next phase of the support. Source: Field visits and Bank reports. 33 DECENTRALIZATION IN CLIENT COUNTRIES Figure 3.1: Distribution of State and Local Revenue Sources 100 92.6 90 e 83.0 nu 80 78.2 ever 70 63.0 60 54.0 local 49.6 dna 50 39.5 40 state 29.5 30 28.0 21.8 total 20 16.5 12.4 ot 9.0 9.9 10 8.3 7.4 % 4.5 0 Morocco Pakistan Peru Philipines Russian Tanzania (2004) (2003) (2005) (2001) Federation (2003) (2006) CG transfers Taxes Other sources Combined (taxes and others) Source: IMF-Government Finance Services and World Bank data. raising own-source revenue. Nevertheless, the however, local governments do not have discretion evaluation found that this first step was a over revenues and the Congress determines all tax necessary condition for macroeconomic stability. policy issues. In Sierra Leone, Bank assistance helped establish In 7 of the 20 countries, results have been incentives for local councils by providing them counterproductive. Own-source revenue of with a 50 percent matching grant for revenues districts in Tanzania and Uganda14 has diminished raised locally.13 However, for capacity significantly since the Bank began its support for Enhancing frameworks and institutional reasons, revenue decentralization. In Tanzania, as discussed in for own-source revenue generation is still negligible, even chapter 2, this happened as a direct result of Bank through Bank support has though the framework is in place. support that centralized revenue sources while been difficult. other Bank support was in parallel helping The evaluation found modest results in decentralize expenditure responsibilities. There this area in eight country cases, for reasons that is as yet no clear strategy for enhancing local include inappropriate assignment of revenue government revenue, and district governments sources, inefficient tax administrative systems, and continue to be dependent on central transfers for weak local capacity. In Indonesia, where authority more than 90 percent of their budget in both for local governments is granted under countries, although there is ongoing dialogue.15 There has been little decentralization, new local taxes have There is also some evidence that the high level of progress in increasing proliferated (sometimes without a legal transfers discourages districts from collecting own-source revenue by basis). Many of these taxes have little own-source revenue (World Bank 2006k, p. 6). local governments for economic rationale and impose various reasons, additional transaction costs because of In sum, own-source revenue can help encourage including inappropriate their arbitrary nature and wide variation autonomy and responsiveness of subnational assignment of revenue across localities. In Peru, municipal tax governments in service delivery. But it has been sources and weak local revenues increased by 11 percent in real difficult to progress on this dimension because of capacity in tax terms because of Bank support for tax central governments' reluctance to relinquish administration. policy and administration reforms; control over lucrative taxes, the need to 34 BANK SUPPORT FOR DECENTRALIZATION FRAMEWORKS AND SUBNATIONAL GOVERNMENT CAPACITY strengthen central tax collection in countries dialogue, reports, conferences, state- In Uganda and Tanzania, undergoing fiscal stress, and the weak capacity level ESW, and support for reform own-source revenue of and political will of many local governments to "competition" across states. With the district governments administer taxes. In the interim, however, passage of state fiscal responsibility decreased after reforms. enhancing the efficiency and capacity of local laws in Karnataka and Andhra Pradesh governments to use the resources they get from in and around 2003, a clear limit to state borrow- the central government and to collect local fees ing was established. Although it is too early to and taxes while engaging in policy dialogue on assess the impact of the laws and sort out causal- appropriate revenue assignments--as the Bank is ity, fiscal positions have improved in the two doing in some cases--may be the best alternative. states (see figure 3.2). Debt management and borrowing Adjustment lending to the Brazilian central The Bank has engaged in debt management and government, underpinned by good analysis, the regulatory framework for prudent borrowing supported a package of reforms to the intergov- at the subnational levels in only 8 of the 20 ernmental fiscal system. The center of countries. (In the other 12 countries, this was these is the Fiscal Responsibility Law, Raising efficiency and not an immediate issue, with local governments which has strong enforcement capacity of local receiving only grants and not being in a financial mechanisms through the legal governments along with position to access credit.) The framework for system.16 The fiscal performance of policy dialogue for subnational borrowing was substantially im- the states and municipalities has appropriate revenue proved in four of these eight countries: Brazil, improved since the 1990s, and the assignment may be the India, Russia, and Pakistan. aggregate subnational sector has had best alternative. primary surpluses since 2002. The In both India (Karnataka and Andhra Pradesh) approach in Brazil also represented good and Brazil, the Bank helped governments deepen practice in that the package included support for their understanding of center-state fiscal issues central government incentives to states for through a series of informal notes, policy adjusting their fiscal policies. Figure 3.2:Trends in Fiscal Deficits in Selected Indian States (as a % of gross state domestic product) 6.0 ctu 5.0 odrp 4.0 domestic 3.0 state 2.0 ossrg of % 1.0 0.0 1995­99 1996­97 1997­98 1998­99 1999­2000 2000­01 2001­02 2002­03 2003­04 2004­05 Andhra Pradesh Karnataka All states Source: Reserve Bank of India (various years). 35 DECENTRALIZATION IN CLIENT COUNTRIES In Brazil and India, In Russia, Bank support helped Thus, support for a strong framework to regulate states improved their establish a framework for subnational subnational debt has proven valuable, where it fiscal positions through government debt management. Since has been undertaken with commitment at the reforms, with strong 1999, Russia's regions have been able national level and accompanied by improved support at the to borrow externally only to finance financial management (including transparent national level. external debt incurred in the past. No debt recording) at the subnational level. In new external borrowing is permitted, countries where subnational borrowing is not yet although the regions can borrow internally common, it may be wise for the Bank to support under certain conditions. There are credit-rating establishment of mechanisms for hard budget agencies, and 20 regions (of a total of 89) have constraints and financial management processes received credit ratings. Bank support also helped for transparent debt recording. Pakistan to pass the Fiscal Responsibility Act and improve debt management of provinces by Administrative Capacity and substituting lower-interest for higher-interest Accountability debt. Weak local administrative capacity and account- abilities result in less efficient service delivery Ensuring transparency of Results of Bank support for and misuse of public resources (Shah 1998; debt accumulation is an subnational debt management were Litvack and Seddon 1999; Ahmad and others important first step in mixed in Albania, Bolivia, Peru, and 2005). To guard against this, subnational govern- debt management. the Philippines. For example, in Peru, ment capacity can be strengthened through despite recent legal measures,17 a administrative decentralization or free-standing Bank study to assess the magnitude of capacity-building activities. The Bank focused subnational debt stock and identify municipali- mainly on the latter route to support decentral- ties that might face debt stock problems showed ization, perhaps because of the complexity of that unregistered debt is significant and that debt civil service and administrative reform, particu- is highly concentrated. Moreover, a significant larly in a decentralized context.18 IEG therefore number of municipalities do not comply with the examined the extent to which Bank support Fiscal Responsibility and Transparency Law resulted in clearer definition of roles and respon- (Ahmad and García-Escribano 2006, p. 21). In the sibilities of different levels of government, Philippines, the Bank supported the develop- stronger subnational financial management and ment of a financing framework for local govern- procurement systems and capacity, and better ments in 1996, but the implementation remains systems to monitor and evaluate subnational weak. government performance. Box 3.3: Analytic Support to South Africa Successful interventions are not necessarily expensive provided the basis for significant aspects of the final or linked with lending, say Bank staff who worked in legislative enactment. First, the vertical division of rev- South Africa (not among the 20 country cases covered enues between national and subnational levels has pro- in depth, as there was no Bank lending to the country) gressively shifted in favor of the latter. Second, ad hoc as it embarked on establishing a decentralized form of and opaque transfers that were a vestige of the apartheid governance. era were phased out and replaced by more transparent Three international experts financed by the Bank en- and equitable intergovernmental transfers. Finally, the gaged key decision makers at workshop sessions and Bank supported the government in establishing a sound helped them prepare policy documents. Those papers framework for subnational borrowing. Source: Desk review and interviews with staff. 36 BANK SUPPORT FOR DECENTRALIZATION FRAMEWORKS AND SUBNATIONAL GOVERNMENT CAPACITY Bank support enhanced administrative capacity Even capacity building has been Support for prudent and accountability substantially in 7 countries, difficult where the responsibilities subnational borrowing modestly in another 12, and negligibly in 1 of the were unclear. Although the Bank has must be accompanied by 20 country cases. Box 3.3 illustrates how the addressed this issue in all 20 countries transparent debt Bank's analytic support helped the South African through analytical support or lending recording at the government design and implement intergovern- activities, substantial results are local level. mental reform. readily apparent only in Ethiopia, Russia, Pakistan, Rwanda, Sierra Leone, and Roles and responsibilities of different levels Uganda (see table 3.1 on page 30). of government It is important first to clarify the service delivery In Russia, Bank support helped clarify To strengthen responsibilities of each level of government, responsibilities at each government administration at the define the necessary competencies, and then level and strengthened institutions local level, Bank support establish the expenditure and revenue responsi- for implementing decentralization focused on institution and bilities of each level (see World Bank 2004j, p. (through improved coordination and capacity development. 185).19 Lack of clarity in responsibilities has linkages between central and local reduced accountability and hampered the institutions). Similarly, in Pakistan, terms of effective operation of subnational governments, partnership between different levels of govern- particularly where subnational governments ment have helped clarify the roles of various depend on central government transfers. actors in the health and education sectors. In Table 3.2: Assessing Results for Enhanced Administrative Capacity and Accountability Improved clarity of Strengthened financial administrative management and Improved monitoring Quality of responsibilities procurement capacity at of subnational results of different levels subnational levels government performance High Improved clarity in responsibilities Improved financial management A well-functioning monitoring system of of different levels of government institutions (external audits, local government performance with rele- established by law and being accounting standards and reports, vant data being available at all levels of implemented and internal controls) and government for improving policy decision improved transparency in procurement at subnational levels Substantial Clarification being improved by Strong and sustainable institutional An effective monitoring system initiated project-related arrangements changes with limited outcomes but at project level and planned to be generating improvement in evident or clear likelihood of replicated for a more general level overall system improvement in both cases Modest Some outputs (say responsibilities Improvement only in one area or M&E system of local government per- clarified for one level), but con- mainly outputs such as relevant formance is available in a sector (mainly fusion continues at other levels laws, but no evidence on improve- through project-level assistance) and in overall arrangements ment in external audits, accounting standards, or transparency in procurement Negligible No progress in this area Some initial outputs in both cases Some outputs, such as a framework for a monitoring system or a needs assess- ment, being undertaken Source: World Bank documents. 37 DECENTRALIZATION IN CLIENT COUNTRIES Lack of clarity about Uganda, Bank support to local govern- Achieving clarity in the roles and responsibilities responsibilities reduces ments helped clarify the roles of the of different levels of government can be politi- accountability and various local actors as well as strength- cally charged. A first step, therefore, would be hinders effective ening the links between different systematic and publicly discussed analytical work operation. levels of government (see box 3.4).20 to help stimulate demand and build consensus for reform within the country. Such analysis can Other countries in the sample had less success then guide the selection of entry points for clarifying the assignment of responsibilities lending support. The Bank may also, as it is among different government levels. Even with increasingly doing, improve clarity through Bank support, governments have been unable to project-specific mechanisms such as perform- reduce the ambiguity in the roles of different ance agreements or terms of partnerships actors involved in service delivery in Bolivia, between different levels of government. Burkina Faso, Indonesia, Morocco, Madagascar, Peru, and the Philippines. This is partly because Financial management and these countries have multiple players at the procurement capacity central and local levels, some with devolved The aim of Bank support in all 20 countries was to functions and others with deconcentrated strengthen financial management and procure- functions in the same sector. In several cases-- ment systems and capacity at the subnational Albania, Andhra Pradesh and Karnataka (India), levels through one or more activities. Such Peru, the Philippines, the Republic of Yemen, and capacity is critical to effective and transparent Tanzania--Bank support has sometimes local resource management, especially when contributed to confused allocation of responsi- demand-side accountability mechanisms such as bilities through its support for approaches that citizen participation and oversight are weak or create parallel institutions for development nonexistent. In 75 percent of the countries, activities at the local levels. assistance to strengthen financial management systems was based on Country Financial Account- ability Assessments and Country Procurement Box 3.4: Appointment of Chief Accounting Officers Assessment Reports that examined fiduciary in Uganda issues at the subnational levels. In the initial decentralization, the District Service Commission ap- Bank support contributed to strengthened pointed all technical staff at Uganda's district level, but a 2005 consti- financial management capacity of subnational tutional amendment gave responsibility for the appointment of the chief governments in a high or substantial way in 15 of accounting officer (CAO) back to the central government. Although the 20 country cases--making the subnational this was widely perceived as having derailed devolution, field visits and accounting and auditing systems more standard- interviews with government stakeholders and CAOs found that the lat- ized and transparent. In Pakistan, Bank support ter are able to act more independently and fearlessly than the local has helped improve the timeliness and accuracy commission. of financial statements, with annual accounts The centralization of such appointments has reduced the risks of col- being produced within six months of the end of lusion at the local level between the executive and the elected council the fiscal year and improved reconciliation levels and facilitated better governance. One CAO noted that she was able to across governments. Roles of provincial and monitor and bring to light without any resistance a case where misrep- district accounts committees have been clarified, resentation between a school and local district staff had led to payment although capacity development has been slow of per capita funds for 1,500 students instead of the 250 actually registered. because of the shortage of skilled staff for the Citizens do not hesitate to bring cases to the CAO, because he or she is district accounting offices. In Brazil, the quality of perceived as independent. This shows how selective division of author- fiscal accounts information has improved contin- ity can enhance accountability. uously, reducing reporting delays and expanding Source: Desk review and interviews with Bank staff. the number of government entities (especially municipalities) that send fiscal reports for 38 BANK SUPPORT FOR DECENTRALIZATION FRAMEWORKS AND SUBNATIONAL GOVERNMENT CAPACITY inclusion in the consolidated public accounts. contributed to reform of financial Bank support contributed management systems. These pro- to substantial results in The Bank was less successful in supporting grams, however, have relied com- financial management in procurement reform; substantial results were pletely on donor funds, and their the majority of sample achieved in only four countries, with modest sustainability with regular budget countries. results in the others.21 Substantial progress in financing is untested. public procurement occurred in Parana and Bahia (Brazil), the Philippines, Kazan (Russia), and Monitoring subnational Uganda. Even in these countries, there is signifi- government performance cant need for capacity building at the subnational Monitoring local government performance levels (see IEG 2006e). In others, such as Pakistan, requires (i) tracking the flow of resources from Tanzania, and the Republic of Yemen, there is little the central level to the local level and evidence of increased transparency and economy monitoring the fiscal performance of Bank support for in local public procurement, although several local governments (which is particu- improving procurement outputs (particularly laws, regulations, and larly justified when the former is was not often effective. procurement agencies) have been generated. providing the resources) and (ii) citizen oversight of local government perform- Two factors have reduced the effectiveness of ance in the actual service delivery. The second Bank-supported capacity-building programs. aspect is discussed in chapter 4. First, in many countries, personnel incentives and governance arrangements (such as levels of Monitoring the efficient use of public resources remuneration, transparency in appointment, and by local governments is essential, where local conditions of employment) have not been governments have a particularly fair degree of modernized. Trained staff, such as auditors and autonomy. It helps improve efficient allocation of accountants, continue to shift to private sector resources, increase transparency in the use of jobs, so it is not clear that public sector service public funds, and create accountability among delivery benefits from their enhanced skills. different levels of government as well as among citizens. In 5 of the 20 countries (Brazil, Russia, Second, the proliferation of local districts, often Rwanda, Sierra Leone, and Uganda), Bank for political reasons, has compounded capacity support led to improved systems to monitor the issues in countries such as Albania, Brazil, transfer of resources through the budgetary Indonesia, the Philippines, and Uganda. This processes as well as to initiate some monitoring tends to increase administrative expenditures, of local government performance. exacerbate weak administrative capacity, and reduce resources available for service delivery. In Brazil and Russia, stakeholders note that such The political difficulty of reform on these transparent monitoring has improved fiscal dimensions has often stymied reformers, even discipline. For example, as a result of the Fiscal with Bank support. Responsibility Act in Brazil, the national govern- ment improved its monitoring of state and The Bank also supported pilot mechanisms to municipality fiscal performance.22 Russia help strengthen financial management systems developed a good system to monitor regional and government processes (see box 3.2, page budgets. All regions are required to submit 33). These include accreditation systems in Peru, information on 10 indicators regularly which set out the objectives, criteria, and assign- to the federal treasury; penalties for More lucrative private ment of responsibilities that govern the accredi- noncompliance include reducing sector jobs and tation of subnational governments so they can transfers. proliferation of local assume the management of decentralized government structures programs. In Pakistan, contractual agreements Progress has been slow in this have reduced capacity at between different levels of government dimension in other countries receiv- local levels. 39 DECENTRALIZATION IN CLIENT COUNTRIES Centralized monitoring of ing Bank support for decentralization. Bank support generated weaker results in areas local government Some monitoring systems are not fully that require effective collaboration among performance is essential, in place, or if in place, the systems still multiple sectors across different levels of especially if subnational do not generate adequate information. government: governments depend on In Peru, they are still confined to top- transfers. down monitoring at the central level. · Enhancing own-source revenue of subnational In Tanzania, monitoring is still only governments done for the donor-funded program, and it is still · Clarifying responsibilities of different levels of difficult to know how much funding local govern- government ments are supposed to receive and whether the · Establishing cross-sectoral monitoring systems funds have been released. In Indonesia and the at local levels. Philippines, monitoring or reporting systems do not yet provide information on local government Better results have occurred in countries where performance, although there is ongoing discus- Bank and other donor support was framed sion on how this can best be undertaken (IEG around a country-led strategy for implementing 2007h, chapter 3). In Pakistan, systems are in decentralization (see table 3.3). In the absence of place but are weak (World Bank 2007h, 2007i, a country-led strategy, Bank support often was 2007j).23 fragmented and failed to recognize the important linkages between different policy actions affect- In addition to supporting measure- ing decentralization. Without such a country There is a need to develop ment of access and quality of outcomes strategy up front, which may not be possible in all robust indicators to within each sector, the Bank needs to cases, Bank support for decentralization was monitor the costs of support the central and local govern- more effective when initiated as smaller-scale delivering services at the ments in developing monitoring interventions that allowed client stakeholders to local level. systems to measure the performance realize the benefits of such reform (Sierra Leone, of local governments and the costs of Russia, and Uganda) and strengthen their delivering basic services. Recent local government ownership of reform before scaling up. support activities provide an opportunity to develop and test such cross-sectoral systems (see Analysis of the case studies yields six main box 3.4). There is, however, still a need to develop lessons for improving Bank support for robust indicators and mechanisms for measuring decentralization frameworks: and monitoring LGU performance that take into account the diverse conditions and capacity levels 1. Undertake upstream analysis, as has been done of local government. Such monitoring systems in Russia and South Africa (box 3.3), and help would also facilitate provision of incentives for prepare an implementation strategy or plan that better performance and would increase trans- carefully considers all elements of decentral- parency and accountability of local government ization and their linkages. functioning. 2. View support for decentralization as a package of fiscal, administrative, and governance reforms Overall Findings and Lessons affecting different levels of government with the Bank support for decentralization has been potential to influence different sectors. The Bank associated with more successful results in areas may then decide to provide assistance selec- where the central government is a primary tively in one or two areas where it has a com- driver: parative advantage and may agree that development partners will address other critical · Enactment of laws areas. · Top-down improvement of financial systems 3. Prepare for a long-term commitment to im- · Imposition of hard budget constraints. plement successful decentralization strategies. 40 BANK SUPPORT FOR DECENTRALIZATION FRAMEWORKS AND SUBNATIONAL GOVERNMENT CAPACITY Table 3.3: Contribution of Bank Support for Decentralization Frameworks Country traits for most of Countries the evaluation period High Substantial Modest Negligible Countries where there was greater clarity in Russia, Brazil, Philippines policy objectives for decentralization and Uganda Ethiopia, Bank support was provided consistently Pakistan with country's stated objectives Post-conflict countries Sierra Leone, Rwanda Countries with less clarity in policy Albania, Bolivia, Rep. of Yemen objectives and/or where Bank support Burkina Faso, was not provided consistent with client India,a Indonesia, objectives for decentralization Madagascar, Morocco, Nepal, Nicaragua, Tanzania, Peru, Source: World Bank database and IEG review. a. In India, results of support for fiscal reform in the states of Andhra Pradesh and Karnataka would be considered substantial; support for local governments generated modest results in terms of decentralization. In India, although political decentralization is strong, there is considerable ambiguity in the devolution of functions because of resistance to implementing fiscal and administrative decentralization. In Indonesia and Tanzania, although the laws exist, there are evident weaknesses in terms of inconsistencies between laws and weak imple- mentation of administrative decentralization. As needed, package support into manageable reform. DPLs also tended to encourage a Better results are phases across CASs, reflecting a clear se- shift from a multiplicity of investment associated with Bank quencing of Bank activities with monitorable projects with different sectoral ap- support that was aligned indicators. proaches to more consistent, systemic with a coherent country 4. Form joint partnership strategies with other de- support. They also allowed for more strategy for velopment partners to ensure complementary effective collaboration within the Bank decentralization. roles. and among development partners than 5. Support capacity building at central and local other approaches would have done. They also levels, tailoring it to the responsibilities and allowed countries to take ownership. roles of each level of government. 6. Use a judicious mix of instruments to support Budget support was more effective in sustaining decentralization. policy reform when there were accompanying investment loans that had a framework consis- In 80 percent of the countries where Bank support tent with the DPLs. They helped build necessary generated substantial improvement in decentraliza- institutions and create ownership of the reform tion frameworks, DPLs helped encourage policy at local levels (Uganda and Russia). 41 Chapter 4 Evaluation Essentials · Support for decentralization in the education sector was more effective when the country had a clear frame- work for decentralization. · Understanding the country context was critical for successful out- comes. · Bank project assessments rarely monitor the effects of decentraliza- tion on service delivery. Fada school children, Chad. Photo courtesy of Kimberley Fletcher; FletcherGallery@Mac.com. Bank Support for Decentralization in Education Services: A Sector Example D id Bank support for decentralization of education services have in- termediate outcomes that are likely to improve service delivery? The education sector was selected for deeper review because of a re- cently completed IEG evaluation in this sector (IEG 2006a), which included a few in-depth country case studies.1 Additionally, there are ongoing IEG evaluations paucity of data on project outcomes at the local in health, water, and municipal management levels and the difficulty of attributing Regional or sectors, which are likely to examine some sector- national-level outcomes to Bank support for specific and decentralization issues. Six country decentralization. cases (Morocco,2 Pakistan [Province of Punjab], Peru, the Philippines, Russia, and Tanzania)3 In three of the six countries (Pakistan, Russia, were selected for analysis based on (i) regional and Tanzania), Bank support was provided representation, (ii) those that received Bank within the context of overall client agendas for support for decentralization, and (iii) those decentralization. Legislation in each country had where missions were undertaken and/or where transferred responsibilities for service there were recent IEG evaluations. delivery to lower levels of government In three of the six before the implementation of Bank countries, the This chapter points to some initial lessons in support reviewed for the evaluation. governments had designing Bank support for decentralization in Here, Bank support aimed to rational- enunciated the education sector. In each case, the evaluation ize the decentralization already decentralization policies focused on the three intermediate outcomes existing in the county, using its covering the education specified in the evaluation framework (figure 1.1 assistance as an entry point. sector; in the others, Bank and see table 4.1).4 It did not evaluate the support tried to develop a delivery of services itself (one of the final In the other three countries decentralization outcomes in the framework) because of the (Morocco, Peru, and the Philippines), framework. 45 DECENTRALIZATION IN CLIENT COUNTRIES Bank support explicitly attempted to create a fiscal transfer system resulted in schools competing normative framework to initiate decentralization, for enrollments by improving service quality. The using its assistance in the sector as an entry point. allocation of resources on a competitive basis to In Morocco5 and Peru,6 the government had not regional governments willing to initiate and yet decentralized education when Bank support implement reform has been noted as enhancing in the sector was prepared. In the Philippines, the local ownership and accountability for improving sector was already partially decentralized and service delivery. In contrast, in the Province of Bank support aimed to deepen this.7 Punjab (Pakistan), the allocation of funds to schools has remained arbitrary and opaque since Improving Intergovernmental Fiscal decentralization. Frameworks in the Education Sector Bank support helped improve intergovernmental In Tanzania, local governments acted more as fiscal frameworks for education in Russia, the agencies with delegated rather than devolved Province of Punjab (Pakistan), and responsibilities. Responsibilities are still shared In Punjab, Russia, and Tanzania, consistent with the prevailing between the Ministry of Education (which is Tanzania, the Bank framework within the country. In these unwilling to give up its control) and the Ministry of helped enhance three countries, local governments Local Government, with the Ministry of Finance predictability of receive funds in a more transparent being the conduit for external development transfers. and predictable manner than before assistance. Education funding for the district the Bank-supported reform, thereby arrived from multiple sources, and there was weak ensuring availability of resources for service coordination between central, regional, and delivery in this area. Field visits in Tanzania and district administration in the delivery of primary Chuvashia (Russia) confirmed that schools have education.8 District education officers had to been able to deliver services in a more competi- answer to the ministry representatives as well as to tive and efficient manner since decentralization the district executive director. Stakeholders identi- (see box 4.1). fied education funds with the central government and donors, rather than the district budget. Thus, Per capita fiscal Introducing an element of competition although Bank support helped increase the transfers to schools led to in fiscal transfers has helped improve the availability of resources at the local level, it only competition in improving delivery of education services in Russia modestly enhanced local governments' decision service delivery. andinitiatereformatthelocallevels.The making, responsibility, and accountability. Table 4.1: Assessing Results in the Education Sector Consistency with Enhanced administrative client's intergovernmental capacity of local Improved accountability fiscal framework governments in education to citizens · Improved predictability and timeliness · Strengthened capacity of local government · Improved school autonomy, consistent in receipt of funds, consistent with staff and school managers or with client's decentralization intergovernmental fiscal transfers administrators framework · Strengthened roles of different actors, · Increased control of teachers by local · Strengthened oversight by parents consistent with decentralization governments or their powers and citizens framework or stated objectives of delegated to facilities · Established minimum service stan- Bank support dards or outcome monitoring through development of national learning assessments Source: IEG desk review. 46 BANK SUPPORT FOR DECENTRALIZATION IN EDUCATION SERVICES: A SECTOR EXAMPLE In Punjab Province, with budget support from to LGUs proved to be overly Local government roles the Bank, a system of annual conditional grants optimistic, particularly because Bank under Bank support were for education has been initiated between the support attempted to introduce new not always consistent provincial and district governments, based on a roles at the local level in a sector not with government policy. needs-cum-performance formula. The grants are fully devolved under the 1991 Code. provided on signature of "terms of partnership" In Peru, political economy issues were not fully agreements signed by both governments. These analyzed, and Bank support for decentralization agreements help improve the transparency and of education included in the project was efficiency in the financing and management of dropped. the education sector by clarifying the responsi- bilities and obligations of districts and the Enhanced Administrative Capacity province in planning and implementing projects. of Local Governments However, subsequent transfers to schools The evaluation focused on two The Bank helped to remain opaque and are yet to be addressed. administrative aspects: strengthened clarify roles and systems and capacity for monitoring responsibilities through In Morocco, Peru, and the Philippines,9 where results at the local level, and capacity agreements signed Bank support did not have the benefit of a building of local government staff and between different levels of government-led decentralization framework, the school administrators to improve government. Bank supported delegation of some education their performance (see box 4.2). The responsibilities to local governments. However, Bank supported decentralizing the these efforts did not generate the expected results. appointment of school staff only in Punjab, In Morocco, school construction and some Pakistan (among this six cases in this chapter), decision making were decentralized to provincial where DPLs supported measures to enable local governments (when Bank-supported efforts to school councils to hire additional decentralize the responsibility to community "contract" teachers based on merit. In Bank support for groups failed), but even the transfer of project the other five countries, the focus was decentralization in the funds to provinces proved to be difficult. on improving capacity of teachers and sector was not as their efficient deployment. There was successful in countries In the Philippines, the assumptions underlying little systematic assessment of out- without a broad rapid decentralization of education management comes in this area. decentralization strategy. Box 4.1: Aligning FiscalTransfers with Local Sectoral Priorities Piloting fiscal transfer systems within the country's broad de- ceived success of the system, although other regions have the au- centralization framework was effective in Russia. Bank support tonomy to determine whether to implement the decree. helped pilot regional governments adopt the legal frameworks In contrast, in Tanzania, despite significant results in education necessary for per capita funding and replace funding accord- outcomes in quality and access, intermediate outcomes from a de- ing to cost or staffing levels. All schools in participating regions centralization point of view were not fully effective. Development now receive resources according to funding norms, linked to re- funds were transferred through formula-based grants directly to form in the sector. schools through local government budget accounts, but the cen- This has improved competition by creating demand among tral government continued to provide detailed instructions on their students for better-quality services. It also helped schools in- specific use, which limited scope for local government decision crease the own share of resources from 11 to 17 percent between making (except to a small degree in the procurement of textbooks). 2001 and 2005. A presidential directive has been issued to imple- Thus, the ability of local governments to play a meaningful role in ment per capita funding across the country, signaling the per- local sectoral reform was not strengthened. Source: Field visits and Bank reports. 47 DECENTRALIZATION IN CLIENT COUNTRIES Box 4.2: Gender and Capacity Building at the Local Level Decentralization in traditional societies may lead to a revival of · In the Philippines, IEG field assessments of Bank support in patriarchal approaches in governance at the local level. One way the water sector found that women felt empowered by their of mitigating this risk is to strengthen the capacity of both women participation in water user groups and noted the new skills and men to participate in local-level decision making or user and competencies they had developed as meter readers, groups and to ensure gender-balanced training in terms of both lay auditors, and lay accountants. content and access. · In Indonesia, Bank support has resulted in improved oppor- Research shows that mandated representation of women in tunities for women's participation in drafting local develop- local leadership can help policy decisions to be more gender ment proposals; self-assessments suggest that women have aware (Chattopadhyay and Duflo 2001); field assessments in the become more empowered. Fifty-three percent of the loan pro- Philippines and Uganda confirm the benefits. Such measures are posals approved were from female applicants. prevalent in the education and water sectors and in social fund and · In Uganda, under the Bank-supported Local Government De- community-driven development (CDD) activities: velopment Program, 28 percent of trainees and 30 percent of the beneficiaries of short-term jobs were women. The impact · In Tanzania, school committees are gender balanced and of these activities has not been measured. democratically constituted. Training includes all stakehold- ers, male and female. Issues of HIV/AIDS, gender, and gov- Much more could be done to better monitor whether decen- ernance are mainstreamed into all training. tralization is benefiting men and women equally through improv- · In Pakistan, the Bank supported close monitoring of gender- ing outcome indicators of monitoring systems. The lack of segregated education data. The appointment of teachers is gender-disaggregated data at all levels poses a significant re- based on a point system, which includes some affirmative ac- straint to understanding whether Bank support for decentraliza- tion for recruitment of women. School councils were reestab- tion has resulted in different results for men and women. lished in fiscal 2004 based on majority membership of parents, There is also need for a gender focus in middle-income coun- with mothers represented in councils of girls' schools, although tries outside East Asia. Bank support in these areas often seems significant capacity constraints limited their participation. to take a gender-blind or--at best--a gender-neutral approach. Source: Field visits and Bank reports. Assessment of capacity- In Russia and Tanzania, Bank support straightforward and noted that "council and building support was led to enhanced capacity at the local school-level institutions have, to a large extent, mostly focused on level in ways that improved service performed as planned." It found evidence that outputs. delivery. Field visits in Tanzania suggested "improved efficiency and effective- confirmed conclusions in Bank ness, and also that capacity-building activities reports that training communities and local were also effective" (Hakielimu 2005). In governments had improved their ability to play Chuvashia (Russia), school managers and local their new decentralized roles in the education government staff noted that training workshops sector. In addition, an assessment by an organized by the Bank, in particular the World independent nongovernmental organization Bank Institute, had raised their capacity to (Hakielimu 2005) found that about 90 percent of deliver services. schools reported receiving some training for school committee members, although In the four other countries, results were not as In Russia and Tanzania, they received only 50 percent of the evident. In the Philippines, although there was field visits suggest that intended amount allocated to each progress in training sector staff at the local levels, there has been school. the IEG mission found multiple agencies substantial capacity mandated to strengthen local government strengthening at the The assessment found that school capacity; training tended to be supply driven and local level. recordkeeping was well organized and agency specific. Given the multitude of players in 48 BANK SUPPORT FOR DECENTRALIZATION IN EDUCATION SERVICES: A SECTOR EXAMPLE the sector, both at the central and subnational although the partnership agreements Where multiple agencies levels, and the lack of clarity in their roles, include specific and annually updated are mandated to provide appropriate capacity building was not possible. targets for each district, including training at the local input, output, and outcome targets. level, coordination is Despite the creation of a training department at This is likely to lead to better data in important for effective the provincial level in Punjab, there is no the future. capacity development. evidence that these efforts led to more independent or responsive local decisions. In contrast, in Tanzania, despite significant Bank Government employees said that although the support for capacity building, there is still no federal government rushed to spend and raise systematic institutional monitoring of education social indicators, large programs could not be results at any level--central, district, or school. delivered successfully in the absence of adequate In Peru, Bank support financed the publication capacity and accountability.10 In Morocco, in- of both international and national assessment service training of teachers and school principals results; established monitoring and supervision was shifted to regional governments, but there is systems, including the creation of a payroll no information on outcomes. system to track the problem of ghost teachers; and developed an M&E system designed to Strengthening local systems and capacity to provide transparency of information during the monitor performance and outcomes in their decentralization process. Despite this, the efforts limited jurisdiction helped enhance not only for institutional strengthening and accountability of local governments and service capacity building have remained Monitoring local providers, but also local administration (see table focused at the national and central outcomes and 4.2). In Chuvashia (Russia), the regional govern- ministry level and have not affected performance helps ment prepared and published an annual report the local levels (IEG 2007e). enhance demand for on education, based on formats developed with reform. Bank support. These annual reports generated Enhanced Accountability of local demand for reform and helped schools Local Governments/Schools to Citizens obtain local funds to pay for reform. This experi- The evaluation examined the extent to which Bank ence is being disseminated through inter- support helped increase accountability at the local regional workshops, and other regions may level by enhancing the voice of citizens in service replicate the preparation of such reports. delivery, increasing transparency in local govern- ment functioning, and supporting systems for In Punjab, Pakistan, local efforts to monitor holding local governments accountable to citizens. progress in education are still at an early stage, In terms of improved citizen voice and Table 4.2: School Autonomy and Accountability inThree Pilot Regions in Russia Supported by the Bank Pilot regions Samara Chuvashia Voronezh Progress Planned (%) Actual (%) Planned (%) Actual (%) Planned (%) Actual (%) Independent bank accounts opened 100 100 55 61 70 70 School boards established 90 90 70 84 65 100 Introduction of PCF Done Done Done Other achievements · PCF was also introduced in 22 "grant" regions against 7 planned. · A legal framework for implementation of PCF has been adopted. Source: World Bank 2007i. Note: PCF = per capita funding. 49 DECENTRALIZATION IN CLIENT COUNTRIES Effective accountability transparency of local functioning, the viders and reportedly has led to some quality involves enhancing evaluation found better results in the improvement. citizens' voice, increasing Philippines, Tanzania, and Russia, but transparency of local such results were not evident yet in Enhanced access to relevant information is government performance, Morocco,11 Punjab (Pakistan),12 or Peru important for citizen oversight and voice, as well as and establishing systems (despite significant efforts) (IEG 2007e, to enhance transparent functioning. In Tanzania, to hold local governments pp. viii, 10, 27).13 Results were more school and district authorities are expected to accountable. mixed in the last area in all six countries. publish on public notice boards their respective budgets, the funds received, and their expenses. Bank support sought to strengthen citizen partic- Field assessments in the Philippines and Tanzania ipation and oversight of service delivery in the found that citizens knew that information was education sector in all six countries. In Tanzania, available. Field visits to a few schools found that Russia, and the province of Punjab (Pakistan), there was some information on several notice school boards or councils have been constituted boards, although it was not always posted in a user- with Bank support, and these have been strength- friendly format. The complexity of the transfer ened through capacity building and provision of processes and multiplicity of funding sources some operational resources. Field visits in reduced transparency and the accountability of Tanzania found a relative increase in community local governments to citizens. support for such decentralized citizen oversight systems. In Russia, however, this was not a strong Improved transparency in financial management element (perhaps because of the need to change at the school level in Russia and Tanzania helped mindsets, as explained by a Russian nongovern- increase accountability of school managers both mental organization). A 2006 USAID field assess- to governments and to citizens. Independent ment in Pakistan (USAID 2006) found that bank accounts were opened by all participating significant capacity constraints and lack of schools in these countries. In contexts of weak awareness hindered effective citizen oversight.14 budgetary control mechanisms, the accounts also acted as a basic financial control mechanism, In the Philippines, Bank support included the helping ensure that funds were used for the design of a five-year school improvement plan in intended purposes. partnership with the parents and the Involving citizens in community. It used systematic data Finally, the Bank has also helped client stakehold- school management and such as student achievement (based ers pilot several instruments, such as Public procurement improved on school, division, or national tests) Expenditure Tracking Surveys and citizen report accountability in the and language needs. This contributed cards, which have successfully helped identify Philippines. an impetus for local planning and leakages in the system and increase understanding programming and helped empower of the perceptions of citizens on service delivery school heads to manage a portion of the school (the Philippines and Tanzania). These instruments resources and oversee capital improvements of have engaged poor people in an assessment of their buildings. An innovative initiative for school service provisioning and quality. They have heads and civil society groups to monitor typically generated valuable information that has procurement and delivery of school led to some strengthening of accountability Enhancing access to textbooks (textbook counts) and measures (see IEG evaluation of how to build information for citizen furniture also led to better economy better systems of M&E [Mackay 2007, box 9.1]).16 oversight has generated and efficiency in the procurement of positive results, though school supplies; it also engaged Bank efforts were not as prominent in attempt- they will be fully students and parents in oversight.15 ing to strengthen systems for citizens to hold exploited only with Although sustainability is a concern, governments accountable and to seek redress in improved capacity these measures helped facilitate case of grievances.17 Where they were addressed, and awareness. citizen monitoring of service pro- Bank support was not effective. 50 BANK SUPPORT FOR DECENTRALIZATION IN EDUCATION SERVICES: A SECTOR EXAMPLE Bank efforts helped establish a standard of Russia and Tanzania was more success- School accountability minimum service in the education sector only in ful because the interventions were was strengthened through Tanzania, but this has not yet influenced the designed within broader government a variety of internal accountability of local governments because of decentralization strategies that in- control mechanisms. the lack of systematic monitoring of their cluded the education sector. performance. In Peru, the 2007 Bank study on education identifies the lack of standards as a key In contrast, in Morocco and Peru, the absence of a reason for ineffectual accountability of schools. broad decentralization strategy hindered Bank National learning assessment tests also help efforts to support sector-specific measures to citizens and others to rank and compare the improve decentralization of educa- performance of local governments. tion.19 In the Philippines, although In implementing school- Bank support helped introduce greater based management in Bank support facilitated the participation of school autonomy, the role of the local decentralized countries, Russia in international learning assessments, government in delivering services is stillthe roles of local helping compare the quality of Russian education unclear and yet to be effective. In the governments should with that of other countries. In Peru and Pakistan, Punjab, although there was a client be clarified. the Bank supported national education assess- strategy, the pace of the reform initiated ment systems that could have contributed to by Bank support, combined with low capacity at increased government accountability. However, the local levels, may have reduced the effective- in both countries, specific responsibilities of ness of the support. different levels of government were unclear, leaving accountability difficult to establish. Based on a variety of research findings, the Bank Moreover, these assessments were donor has a preference for school-based management financed, raising concerns about sustainability. with active citizen or parent oversight and some monitoring by local governments. Unless such Bank support for mechanisms to seek citizen decentralization is planned carefully and the redress of grievances has not been enough to roles of different actors at the local levels are increase local government accountability in any clearly defined, such school-based autonomy can significant manner. Except in Lima, Peru, where a cause tensions. This has been the case where national toll-free complaint line has been responsibility for primary education has been established for the education sector, such devolved to local governments, as in Tanzania, or systems have not been a common element of even where such decentralization is partial, as in Bank support in the evaluation country cases. In the Philippines. all other countries, citizens use informal systems, such as complaint boxes or raising grievances at Efforts were usually less successful when the community meetings. Interviews in Tanzania Bank team set overly ambitious objectives and with local government officials suggest that such did not fully consider the specific challenges of complaints have led to some action (such as the country context. In some countries, for dismissal of errant teachers), but none of these instance, the targeted spending levels required systems has been systematically assessed.18 for Bank support exceeded the institutional capability, particularly at the local levels, even Conclusion when the appraisal document noted the lack of Although generalizations are difficult from such a implementation capacity. Analysis of the merits limited sample, the above results do reiterate the and demerits of decentralization in education finding in chapter 3 that Bank support for was also weak. decentralization is likely to be more successful when it is aligned with a client strategy for In Peru, the Primary Education Quality Project decentralization. For example, Bank support for included a component originally designed as a decentralization for the education sector in key tool to promote the decentralization of the 51 DECENTRALIZATION IN CLIENT COUNTRIES Bank support needs to education sector, but it was modified and for such information to be available consis- increase attention to before loan signature by a new tently across all local government jurisdictions. strengthen the administration. The new administra- First, it is difficult to assess the impact of accountability of local tion focused on improving school decentralization at the local level by measuring governments to citizens. management while maintaining a national or even regional outcome indicators in centralized system. The Bank moved the sector. Policy makers need such data to fine- forward anyway, dropping the decentralization- tune policies. Second, availability of local related activities but not modifying the project information can enhance responsiveness and objectives. The lack of strong technical analysis accountability of local governments as well as seems to have constrained the Bank from generate local demand for reform when capacity convincing the new government of the benefits is built up. of a decentralized system. It is difficult to link decentralization to improved In strengthening the accountability framework at service delivery in the areas of education. For the local level, Bank assistance has been more decentralization to work, the WDR (World Bank successful in enhancing voice and transparency. 2004j) notes, "There must be relevant informa- However, it will need to improve its attention to tion about performance across jurisdictions so mechanisms for holding local governments that citizens can bring justified pressure to bear accountable. Strengthening focus on monitoring on politicians and policy makers if their area is systems that provide comparative information lagging. Second, there must be an environment across localities, establishing service standard in which local jurisdictions can experiment and benchmarks, publishing outcome data evaluate new approaches." Local data on service at the local level, establishing systems delivery are important; to redress complaints, strengthening By these criteria (of the cases reviewed in this measuring regional or local-level justice and accountability chapter), decentralization in the education national outcomes is institutions, and encouraging some sector has taken root only in Russia, in the sense inadequate. competition between governments that information across regions is being built up can help increase such accountability. and regional governments and schools have This is a complex task and will need close coordi- significantly improved discretion over fiscal and nation among different units within the client administrative matters. In other countries, country and the Bank. decentralization in the education sector has been initiated, but the central government continues Finally, even this limited review stresses the to make the key administrative and fiscal need both for sector-level data at the local level decisions. 52 Chapter 5 Village meeting in Ilha Ibo, Mozambique. ©Atlantide Phototravel/Corbis; reproduced by permission. Findings and Recommendations F or reasons as diverse as the countries themselves, decentralization is oc- curring in a broad cross-section of Bank-borrower countries. In several of the country cases reviewed, democratic governance and increased participation of local citizens in development planning through elected local governments are among the several objectives. Findings populations (in Pakistan and Peru). Thus, from a The election of local governments is the first step development perspective, decentralization is in this direction, and in all 20 countries this has neither inevitable nor inherently good or bad, occurred. From a development perspective, and Bank support must depend on country however, neither decades of theoretical and context, political will, and government commit- empirical research by academics nor the rich ment (see table 5.1). experiences of the Bank have shown that decentralization either facilitates or undermines Additionally, decentralization was largely driven the main goals of the Bank's assistance: to by political motivations in many client countries. improve governance and service delivery while This fact led to the Bank's initial combating poverty and encouraging growth. hesitation in supporting broad-based From a development decentralization--including support point of view, The evaluation found that the process of for strengthening local governments. decentralization is implementation is difficult in several countries This in turn led to a sector-by-sector neither inevitable nor that are under fiscal stress and that have regional approach for decentralization, with a inherently good or bad. inequities because of trade-offs that need to be focus on efficiency. As a new wave of made. Field visits indicated that decentralization decentralization persisted, the Bank started to has led to a sense of empowerment among local provide support more regularly. communities (Tanzania, the Philippines, and Uganda). But the evaluation also found that This is true especially since the mid 1990s, mainly devolution under certain conditions exacerbated because the process of decentralization resulted interregional inequities, as in Bolivia and Russia. in shifting responsibilities to local governments Moreover, given the potential for elite capture, in many critical sectors or subsectors, such as enhanced discretion for local governments was primary education, basic health, and water. It not always consistent with enhancing service then became essential to ensure that decentral- provision for the poorest and most peripheral ization was effectively managed, that resources 55 DECENTRALIZATION IN CLIENT COUNTRIES Table 5.1: Support for Decentralization Needs to Be Country Specific Levels of decentralization Possible Bank responses The client has a legal framework · Align Bank portfolio with client policies on decentralization. for decentralization that has · Support the client in analyzing the implications of decentralization across relevant sectors and differ- devolved or delegated ent levels of government (covering macroeconomic fiscal effects, equity implications, administrative responsibility and authority rationalization and capacity, accountability frameworks, important trade-offs, and critical political for service delivery to local economy issues) if such analysis is not available. governments and has · Initiate dialogue on areas for further policy reform and support the client's development of an action established institutional plan for deepening reform around which development partners may provide coordinated support. arrangements for decentrali- · Help local governments strengthen their capacity for planning, budgeting, financial management, and zation, although these may procurement. be weak. · Strengthen accountability mechanisms of local governments to citizens and upward to higher levels of government as well. · Help the central government establish a cross-sectoral monitoring system to review local govern- ment performance and service delivery, with information made available to policy makers. · Ensure clarity in roles and responsibilities both vertically and horizontally, and strengthen administra- tive decentralization (whatever the mode may be). · Support testing or piloting of fiscal transfer systems to local governments, with some flexibility to respond to local needs and priorities. · Encourage appropriate assignment of revenue to local governments and strengthen capacity and systems to collect own-source revenue. · Ensure that hard budget constraints and transparent debt recording systems exist and work before encouraging local governments to borrow. The client has decided to · Help the client analyze the costs and benefits of fiscal and administrative decentralization and exam- decentralize and/or has a policy ine different options. or law but does not have · Help local governments strengthen capacity for planning, budgeting, and financial management, as institutional arrangements well as accountability mechanisms. in place. · Help build a cross-sectoral system to monitor the results of service delivery at the local level. · Strengthen mechanisms for citizen participation in the design and implementation of development activities and strengthen the capacity of local governments to support such development. · Ensure coordination in any Bank-supported sector-by-sector decentralization and coherence across relevant laws. · Ensure hard budget constraints and transparent debt-recording systems before encouraging local governments to borrow. Countries that are focusing on · Avail themselves of existing sector-level or local level entry points that help the government administrative deconcentration strengthen human and institutional capacity at local levels. · Strengthen accountability of service deliverers to the government. · Support community participation in design, implementation, and monitoring of development activities. · Strengthen civil society monitoring at local levels to provide information on citizen perceptions of service delivery. Source: IEG evaluation findings. 56 FINDINGS AND RECOMMENDATIONS were available to fulfill the decentralized service · Bank support brought better results The Bank's support for delivery responsibilities, that capacity of local where there was consensus around decentralization was, at governments and other partners was strength- the reform within the country prior a sectoral level, driven ened to undertake the new responsibilities, and to Bank engagement. largely by considerations that local government accountability for service · The achievement of results is ham- of efficiency. delivery was enhanced. Thus, Bank support for pered by the fragmentation that decentralization at the country level was highly comes with a sector-by-sector approach. relevant because it was responsive to client · Assistance aimed at bolstering the basic de- demands, consistent with their policies, and centralization framework in client countries is consistent with the Bank mandate to improve important, but equal attention is needed to delivery of basic services to the poor. strengthen capacity and accountability mech- anisms of local governments. The quality of Bank support for decentralization · Analysis of political economy issues The quality of Bank was mixed during the 1990s and was not always is critical to the prospects for success support improved in two- initiated in a timely fashion. However, the quality and sustainability of decentraliza- thirds of the 20 countries of support improved significantly in more than tion and, therefore, to Bank in the last five years of two-thirds of the countries in the last five years of effectiveness. the evaluation period. the evaluation period. In these countries, the · Systematic monitoring and evalua- Bank has helped to undertake or update relevant tion of local outcomes is critical for under- and comprehensive analysis, identify fiscal costs standing the impact of decentralization on and equity implications of decentralization, and local service delivery to the poor. support necessary policy reform. At the same time it has provided technical assistance and Anchoring Bank support to country-led support to local governments to strengthen their decentralization institutions and their capacity to play their new Results have been better in countries where roles. There is still a need to strengthen the Bank and other donor support was framed Bank's results-based approach for decentraliza- around a country-led strategy for implementing tion, with a focus on assessing results at the local decentralization (Uganda and Russia). In the level and establishing causal links among inputs, absence of a country-led strategy, Bank support outputs, and desired outcomes. often was fragmented and failed to recognize the important linkages between different policy The evaluation found that Bank support actions affecting decentralization. contributed to improved effectiveness of decentralization in a manner that is likely to Thus, for example, if a country implemented enhance service delivery, substantially in 7 of the fiscal decentralization without clarify- 20 countries, modestly in 12, and negligibly in 1. ing authority and responsibilities at Bank support was The evaluation did not attempt to aggregate the different government levels, results effective when it was indicators for quality and decentralization were weaker (Albania and Nicaragua). linked to a country-led frameworks into a single rating. However, a In some countries, after several years decentralization strategy. comparison of the ratings for quality and results indicates that when the quality of Bank support Table 5.2: Better Quality of Bank Support Associated improves, the results also get better (see table with Better Results 5.2). This suggests that closer monitoring of the quality of Bank support for decentralization will Percentage of ratings for likely improve the Bank's contribution to overall Quality of decentralization frameworks results in the country. Bank support High or satisfactory Moderate or negligible High (6) 5 1 Much has been achieved, but this evaluation Medium (14) 3 11 identified five broad areas of concern: Source: IEG evaluation. 57 DECENTRALIZATION IN CLIENT COUNTRIES of weak results, the Bank supported the prepara- Regional vice president has placed one unit in tion of a strategy for implementing its decentral- charge of developing a strategy to support client ization policy (Madagascar and Nicaragua). countries as they seek to strengthen decentral- ization in consultation and collaboration with Overcoming fragmentation relevant sector units. The Uganda country team Better collaboration across Bank units would holds weekly cross-sectoral meetings to discuss enhance the development effectiveness of Bank Bank support and to enhance consistency. Such support for strengthening decentralization. For practices need to be encouraged as governments example, in Tanzania, Bank support for the move forward with their decentralization decentralization of expenditures by the regional agendas; this could avoid harm that Bank urban unit and parallel support by PREM for support could cause without such coordination. centralization of revenue sources limited the effectiveness of overall Bank support for Strengthening local governments decentralization. In Nicaragua, collaboration The evaluation found that a broader agenda of between the agricultural and rural units, which intergovernmental institution building that supported a law of fiscal transfers to municipal encompassed local governments was lacking in government and PREM--which reportedly almost half of the countries evaluated-- cautioned the government against the law--may especially those without decentralization plans have avoided some setbacks. or political consensus for innovative reforms. The evaluation found a need for support to Separate analysis and interventions for strength- strengthen local government systems. Once a ening urban local governments as opposed to country puts into place elected local govern- rural local governments further fragmented ments and provides them with some responsibil- support for decentralization in states in India. ity for service delivery, it is important for the Improved collaboration between units working Bank to recognize them as legitimate agents of on different aspects of decentralization through the local development process. Additionally, sectoral interventions could have enhanced Bank support for local institutional reform needs development effectiveness in Albania, the Philip- to have solid intermediate benchmarks that will pines, and the Republic of Yemen. encourage results-based strengthening of local governments, both rural and urban. Future Bank support needs to be designed as part of a broader framework for decentralization. The Bank needs to proceed with caution in client Such support may be incremental and focused countries where some form of political on specific institutions or on a particular level of decentralization has taken place but the adminis- government. The framework need not always be trative and fiscal details remain murky. Local developed in depth in every sector, and different governments should receive Bank funds as elements need not be pursued simultaneously. credits or block transfers only when that support However, it should ensure that different sectoral is underpinned by thorough analysis. It should interventions (including urban and rural) can be also be linked to technical assistance for local coordinated to avoid internal contradictions and governments, strategic but incremental, and to complement one another. should be included within a strategy for decentralization reform. Additionally, as the Bank Bank country directors and vice implements its local government program, it Better coordination presidents have the span of vision and should mitigate the risk of softening budget between different units control to avoid an isolated sector-by- constraints that have been carefully built up. within the Bank that are sector approach. A few country providing support for directors have already begun to Understanding political economy issues decentralization require more cross-sectoral collabora- Finally, the evaluation found that the Bank's would be helpful. tion in providing support. One support was often insufficiently grounded in an 58 FINDINGS AND RECOMMENDATIONS understanding of the country's political highly unequal countries. Much of the Bank's economy, particularly at the local government work on decentralization sidesteps this crucial level. Analytical work did not ask basic questions issue. about the country context, such as, Which actors and interest groups have an interest in Recommendations undermining reforms? Who has an interest in In many of its country programs, the Strengthening local success? What are the ways in which actors will Bank has made a de facto strategic governments in both attempt to game the system in the future? decision to support decentralization rural and urban areas is and subnational government capacity critical when they are By comparing proposed or ongoing reforms with development. In a few cases, notably responsible for delivering ideal best practice benchmarks, analytical work where the client country has made basic services. often did not assess whether the proposed reforms decentralization a cornerstone of its could create a sustainable political equilibrium. development strategy and has demonstrated What looks like a sensible recommendation--such political commitment to decentralizing, Bank as encouraging local credit markets with limited support has been built on an explicitly cross- central oversight or voluntary coordination among cutting approach. regional health agencies or local mayors, and so forth--might be unrealistic. Political economy In most cases, however, Bank support has taken factors must be reviewed from the beginning, a sector-specific route, targeting decentralization which calls for an analytical framework that is not and/or subnational government capacity de- exclusively sectoral. velopment as a logical way of supporting more effective and responsive service delivery. In these M&E at the local level latter cases, the various Bank sector units have The impact of decentralization cannot be not always provided consistent and coherent measured by Regional or national outcome support for decentralization. indicators or by assessing inputs such as percent- age of fiscal transfers or of own-source revenue Looking forward, IEG offers the following of local government budgets. Robust local-level recommendations. They are applicable to every indicators and systematic assessment methods client country that has transferred at least some are urgently needed before one can judge the responsibility for service delivery to subnational effectiveness of decentralization to improve governments, where the Bank makes a de facto equitable service delivery. strategic decision to provide support for decentralization through either a cross-cutting This is not an easy task. There is a paucity of data or a sector-specific approach: on resources needed for effective service delivery in different localities. Furthermore, the absence of · Ensure that Bank support, particularly lending, institutional mechanisms for establishing and is underpinned by genuine client commitment monitoring affordable minimum service standards to decentralized service delivery, given its im- even in basic education or health care also makes portancetothesuccessofBankinterventions.Oc- the task difficult. casionally, a role for the Bank may be justified in the absence of client commitment (for exam- In closing, it is also important to remember that ple, to forestall potentially adverse measures), al- much of the optimism about decentralization, in though the evaluation finds that Bank particular devolution and enhanced accountabil- interventions under such circum- Improved understanding ity, presupposes some degree of local revenue stances are not usually effective. of political economy autonomy. Although generating local revenue is · Encourage the adoption of a more issues is essential to help linked to the wider economy and is a long-term results-based approach to decen- frame more realistic and strategy, there is a need for a realistic understand- tralization by helping develop in- less ambitious objectives ing about how this will be achieved, especially in country and Bank capacity for M&E in this area. 59 DECENTRALIZATION IN CLIENT COUNTRIES that focuses on local outcomes (such as en- tors at different levels of government and hanced accountability, greater citizen partici- across sectors affected by decentralization pation, and improved service delivery) rather ­ Accompanied by support (from the Bank or than only on the process of decentralization. others) to develop and maintain local gov- · Ensure that Bank support at the country level ernment capacity to the extent feasible. is--among other things-- · Strengthen institutional arrangements within ­ Founded on a clear analytical framework the Bank to ensure that an integrative view based on an integrative understanding of underpins Bank interventions, particularly economic, political, and institutional fac- those based on sector-specific entry points. 60 Appendixes Meeting of Mexican farmers. Photo by Giuseppe Franchini, courtesy of World Bank Photo Library. APPENDIX A: TREATMENT OF DECENTRALIZATION IN BANK STRATEGIES AND POLICIES IEG reviewed Bank strategies and Operational that this may be done by decentralizing water Policies to assess their treatment of decen- management. In fact, there is more of a sugges- tralization-related issues.1 Although several tion that effective management may need strategies discussed decentralization, no policy decisions to be made at a higher level. addressed the issue substantively. OP 8.60 on Development Policy Loans (DPLs) states, "The Water management must make a series of Bank may provide development policy lending important transitions, including from local to to a member country or to a subnational division regional and international management. Water of a member country." OP 8.30 on financial management is moving from being just a local intermediaries is discussed because it was found issue to being a national and an international to have some relevance for municipal develop- issue, requiring new approaches to financing, ment funds. All safeguard policies apply to dispute prevention, and resource management. subnational lending, as do some other policies, such as OP 4.20 on gender. B. Sustaining Forests--A Development Strategy (2003) Bank strategies have taken a variety of A major Bank concern in the natural forest approaches to decentralization. Some acknowl- component of its program has been the edge that decentralization is occurring in client prevalence of the concession system of manage- countries and suggest the need to ensure ment, which has produced great pressure on the involvement of lower-level institutions in resource and social and environmental problems capacity building and in making choices. Others, in many places. Governance issues are at the such as rural development, the previous Health core of this problem. In addition, the rapid trend Strategy, the Public Sector Strategy, and the Cities toward decentralization in the countries will and Local Government Strategy, strongly encour- have consequences on managing the compli- age and support decentralization within their cated, interconnected natural resource systems sectors while noting the risks. that will not necessarily be positive. The remainder of this appendix provides brief The policy seems to focus more on people and descriptions of these approaches. It is merely a communities: "cut and paste" of relevant paragraphs from the strategies and policies with some care to ensure To ensure that women, the poor, and other that the overall messages are not distorted. marginalized groups in society are able to take a more active role in formulating and A. Water Resources Sector Strategy: implementing rural forest policies and Strategic Directions for World Bank programs; to support the scaling up of Engagement (2003) collaborative and community forest This strategy focuses on local government and management so that local people can citizen participation in determining strategies manage their own resources, freely market and building capacity but does not suggest at all forest products, and benefit from security 63 DECENTRALIZATION IN CLIENT COUNTRIES of tenure; to work with local groups, D. Education Sector Strategy Paper 1999 NGOs, and other partners to integrate and 2007 Draft Update forest, agro forestry, and small enterprise Overall the 1999 strategy is noncommittal: activities in rural development strategies. Governance and decentralization: Vir- C. Making Sustainable Commitments: tually all of the Bank's client countries are An Environment Strategy for the World tackling education reforms that often Bank (2001) involve decentralizing management and This strategy focuses on the ongoing decentral- accountability for results. The Bank plans ization in many countries and stresses the (with many partners and leading academic implications. A few examples: institutions) to develop a training course for policymakers and international agency In parallel with the changing relative roles of staff on what works and what doesn't and the public and private sectors, the ongoing how to implement education reforms in decentralization of regulatory functions politically sustainable ways. from central to local government levels worldwide has increased the need for local Arguments about the equity of local partic- government involvement in many areas of ipation and the effectiveness of local environmental regulation, and enhanced decision making are often advanced to the role of civil society in influencing push for greater decentralization of the decision-making. The new challenges management of education systems. Central created by decentralization for effective governments around the world have environmental regulation and management decentralized education management to at the local levels deserve special attention varying degrees--and with different ends in capacity-building efforts. in mind. Some have delegated responsibil- ities to their own local representatives, A sample objective is the following: while others have transferred authority to locally elected governments and, in some Supporting policy, regulatory, and institu- cases, to parent-elected school boards. tional frameworks for sustainable environ- Although they have a major role to play in mental management. In cooperation with education, governments cannot do client countries and development partners, everything. Fiscal considerations, includ- we will help client countries (a) strengthen ing competing claims on the public purse, their environmental policy, regulatory, and make it difficult for most governments-- institutional frameworks with a special even those whose philosophies might focus on local environmental institutions; push them in this direction--to be the sole (b) strengthen environmental assessment provider of "free" education to all who seek systems and practices; (c) reinforce the it at every level. There are many areas of positive role of markets and the environ- education service provision (such as mental benefits of sectoral and macroeco- textbooks and vocational training) where nomic reforms; and (d) support good actors other than the government tend to governance, institutions for collective be more effective and efficient. action, increased transparency, access to environmental information, and public The trend seems to be toward school-based participation in decision making. management rather than a focus on strengthen- ing local governments: Its framework, however, indicates an increasing role of citizens, but not necessarily for local Partnerships amongst central government, governments or decentralized institutions. local government and communities, within a 64 APPENDIX A: TREATMENT OF DECENTRALIZATION IN BANK STRATEGIES AND POLICIES more decentralized form of management, political voice at the national level. The can improve service delivery; and poor Bank will encourage governments to communities and rural non-government concentrate on: providing public goods; providers can be effective partners in upgrad- establishing legal, administrative, and ing the quality of education. Local partners, regulatory systems that correct for market in particular, have the local knowledge and failures; facilitating efficient operation of the understanding of local values, culture, the private sector; and protecting the and traditions that are an essential feature of interests of the disadvantaged. Decentral- sustainable development. ization offers great scope for improving delivery of public sector functions. Facilitat- Many others have important roles to play ing further decentralization in rural areas is alongside the various levels of government. an important part of the policy agenda These other players include students, outlined in the strategy. To promote the parents, families, communities, local and development of effective institutions for non-governmental organizations and rural development, the Bank will support: foundations, teacher groups, various forms of private and public-private ventures, and · Strengthening of local administrative numerous international organizations. The capacity more this rich array of partners can work · Transfer of responsibility for services to together effectively, the better the results the administrative level closest to the will be -- and the faster education will users improve. Many partnerships are exemplary · Enhanced accountability for public now, but some are not and most could be administration at every level strengthened further. · Participatory approaches, including in- creased political space and participation 2007 Draft Update in decision-making bodies for women The trend toward decentralized manage- · Economies of scale in government ment and financing of education has been functions gaining momentum. Decentralized manage- · Appropriate private sector involvement ment can improve local accountability but in the delivery of public services, with also lead to inconsistent roles of central, public accountability. regional, and local authorities, as well as inequitable financing across communities. It also adopts community-driven development Policy reform toward effective decentralized (CDD) as part of its overall strategy: "Decentral- service delivery is essential. ized development efforts such as CDD offer the potential for increased community participation E. Reaching the Poor: A Renewed in all aspects of rural development as well as Strategy for Rural Development (2003) offering greater inclusion of all social groups in This strategy strongly supports decentralization rural decision-making." (See pages 35, 36, and 37 as part of good governance. in particular.) In the decentralization process, local F. Reforming Public Institutions and governments must be given sufficient fiscal Strengthening Governance: A World resources to discharge their new responsi- Bank Strategy (2000) (referred to as the bilities. Political decentralization is also Public Sector Strategy) necessary, as it promotes accountability and Decentralization is one of the eight key elements governance reforms at the local level. This of the Bank's public sector strategy, and the is especially important for rural areas strategy states its links with poverty alleviation as because most rural people have had a weak follows: "(i) increased resources for develop- 65 DECENTRALIZATION IN CLIENT COUNTRIES ment purposes; (ii) improved service delivery; complexity has sometimes been over- and (iii) empowerment of the poor to direct the looked by the Bank and other donors in the use of government resources." haste to offer policy advice. Fragmentation of policy advice has also sometimes been a We must prioritize our activities in order to problem. Decentralization is a cross-cutting staff effectively under current resource issue that affects most topics in which we constraints. We aim for the Bank to be engage our clients--from macroeconomic considered one of a very few leading stability to service delivery. Although authorities worldwide in several core areas country teams are increasingly trying to where we have a track record or a compar- bring a comprehensive, coordinated ative advantage, including (a) the role of perspective, the traditional Bank approach the public sector, (b) the broad structure has handled public finance and sectoral of government (including decentralization issues separately and has sometimes and intergovernmental fiscal relations), (c) focused disproportionately on the fiscal core systemwide administrative and civil aspects of decentralization without consid- service reform and capacity building, (d) ering the political and administrative public expenditure analysis and manage- aspects that are critical to success. ment, and (e) sectoral institution-building (including regulation of private service See box 8 of the strategy for the Bank approach delivery). in the public sector. "This strategy paper does not go into detail in evaluating institution Decentralization activity--an accelerating building efforts in particular sectors. That task trend in many countries and the focus of should be a primary concern of strategy papers extensive work in the Bank--entails a for the individual sectors." change in the legal and regulatory framework for government activity, and it G. Health, Nutrition, and Population often holds the promise of increasing both (1997) and 2007 Healthy Development: "voice" and participation (by moving the The World Bank Strategy for Health, administration of public services closer to Nutrition, and Population Results the citizenry) and competitive pressures The previous Health, Nutrition, and Population (including competition among levels of (HNP) Strategy (1997) recommended that government and between subnational governments "be encouraged to address each of entities). But while decentralization holds the three HNP priorities outlined above, through many promises, it also entails risks that decentralization, greater partnerships with non- must be addressed, including the risk of governmental providers, and a more direct "capture" by local elites or lack of capacity public involvement in securing sustainable within local government. financing." It states the need for sequencing: Designing decentralization policy within a ...where institutional capacity for financing particular country context is a complex and regulation is weak, a gradual approach task. There is no right or wrong degree of emphasizing decentralization and internal decentralization or standard "best practice" markets is better than actively transferring that can be applied across countries, ownership of public facilities, with all its although most developing countries tend attendant employment and political to be relatively centralized public sectors consequences. In stronger institutional and could benefit significantly from greater settings--when there is an appropriate and decentralization if well-designed and effective regulatory environment--a more implemented. The best design will vary active participation by nongovernmental depending on circumstances, and this providers can be encouraged. 66 APPENDIX A: TREATMENT OF DECENTRALIZATION IN BANK STRATEGIES AND POLICIES The 2007 HNP Strategy takes a cautious approach governments. Ensuring well-functioning to decentralization and only notes "taking urban areas requires support to a spectrum advantage of decentralized decision-making and of activities, both national and local, that management while, simultaneously, putting in affect urban outcomes. Skills and resources place compensatory mechanisms for capacity and for this effort must be mobilized across equity issues." It suggests in its results framework sectors, thematic groups, and professional that decentralization is one of the many innovative clusters in the Bank Group. The strategy ways of making the system work. therefore calls for a commitment by a wide coalition of forces within the institution H. Cities in Transition: Urban and Local and among external partners to working Government Strategy (2000) together in new ways on the urban frontier, This strategy notes that decentralization is under with a newly empowered set of clients. way in all Regions and provides new ways to work together within the urban arena and with newly The urban strategy is also geared toward energized local governments. It stresses the helping government at all levels, the need to focus on local governments: private sector (for profit and nonprofit), community groups, and citizens function The role of national governments is being in the urban economy in ways best suited refocused to facilitate markets, promote to them. This means, for example, promot- economic and social stability, and ensure ing effective competition among land equity. But reforms of public sector developers and service providers; making management or private sector develop- local government budgets more transpar- ment will not do what is desired for ent and thereby reducing the perceived national development until they are risks of partnership for private financiers; adapted and implemented appropriately at increasing channels of information and the municipal level. Local government collaboration among community groups, remains the everyday face of the public informal sector operators, and local sector--the level of government where government agencies; and refining policy essential public services are delivered to tools such as targeted subsidies, basic land individuals and businesses, and where use planning, and urban transport manage- policy meets the people. ment to address social and environmental externalities in the urban economy. The new urban and local government strategy does more than simply retool the With formal fiscal decentralization in many urban development portfolio or seek countries, the Bank's urban staff can bring stronger performance from it, although the perspective of municipalities, and an both are required. Rather, it argues for the understanding of the needs and con- Bank to recognize cities and towns as a straints facing different kinds of local dynamic development arena where the governments, into the national macroeco- convergence of sectoral activities and nomic and fiscal dialogue to promote more collaboration among communities, levels effective design and implementation of of government, and other private and these reforms. public sector institutions can create a microcosm of sustainable development for The strategy clarifies that "an important feature the country. The Bank would therefore of these efforts is the emphasis on strong apply to urban economies and local underlying incentives for local government governments the same quality and rigor in accountability, and performance, which are analysis, advice, and strategizing that it essential to combat problems of corruption." It applies to national economies and central notes that assistance for urban development has 67 DECENTRALIZATION IN CLIENT COUNTRIES also "traditionally devoted attention to the Such FIs aim to remove or substantially intergovernmental assignment of functions, reduce subsidies, whether provided expenditures, and revenues that determines the through interest rates, directed credit, system of incentives and the availability of institution-building grants, or otherwise. resources for local governments." Para 8 states certain conditions for poverty reduction programs including that they (a) I. Empowering People by Transforming are transparent, targeted, and capped; (b) Institutions: Social Development in are funded explicitly through the govern- World Bank Operations (2005) ment budget or other sources subject to The World Bank has consolidated its approach to effective control and regular review; (c) are social development into a single Bank-wide fiscally sustainable; and (d) do not give an Strategy and Implementation Plan, entitled unfair advantage to some FIs vis-à-vis other "Empowering People by Transforming Institu- qualified and directly competing insti- tions: Social Development in World Bank tutions. Operations" (World Bank 2005d), which focuses on efforts to empower poor women and men K. Good Practice Note for through enhanced Bank support for social Development Policy Lending: inclusion, cohesive societies, and accountable Subnational Development Policy institutions. Social development is defined as Lending (2005) transformation of institutions and, as such, This Good Practice Note is intended to provide promotes better growth, better projects, and guidance to staff in the preparation of develop- better quality of life. The paper sets a vision, ment policy operations and is not mandatory objectives, and a course of action for the longer operational policy. It provides advice to Bank term and suggests specific actions, targets, and staff and to government counterparts on institutional measures for the next five years. It development policy operations in support of clearly notes the need to link CDD activities with state-level programs of fiscal and sector policy macro policies of decentralization and to and institutional actions, with a focus on individ- integrate local governments into the CDD activi- ual state development policy operations. The ties. Consequently, it proposes to reduce the by- note complements the OPCS series of Good passing of legitimate local governments and to Practice Notes (World Bank 2004e) for develop- ensure that Bank support for CDD activities is ment policy operations, particularly Good part of Country Assistance Strategy (CAS)­ Practice Note 1, "Designing Development Policy programmed strategies. Operations," and it also draws on earlier staff guidelines issued by the Poverty Reduction and J. OP 8.30 on Financial Intermediaries Economic Management (PREM) Network. (revised in 1998) Development policy lending support for efforts This OP is applicable to Municipal Development to improve subnational finances, policies, and Funds: institutions takes different forms, depending on the roles of the central and the state-level author- Financial intermediaries are inter alia to ities: fiscal federalism operations, multistate support private sector lending or to support development policy operations, and individual the country's poverty reduction objectives. state development policy operations. This Good Financial Intermediary Loans are provided Practice Note provides no advice on develop- in the context of sound analytical work on ment policy operations supporting fiscal federal- sector issues, appropriate technical ism reforms or on subnational investment assistance, and, as relevant, development operations (except when discussing the choice policy operations to address policy issues. of lending instruments). 68 APPENDIX A: TREATMENT OF DECENTRALIZATION IN BANK STRATEGIES AND POLICIES Section IV discusses the necessary assessments maintained over the medium term.... OP and criteria for determining when to lend to states 8.60 requires "satisfactory fiscal relations or provinces, to whom to lend, and what lending with the central government" for instrument to use. With regard to intergovern- subnational DPL operations to go ahead. mental fiscal relations, the note states: The standard for "satisfactory" is not a textbook model of fiscal federalism--state In the continued presence of a weak responsibilities matching own revenues budget constraint in intergovernmental and rule-based transfers and binding fiscal relations, a program to reform constraints on borrowing--but a practice subnational fiscal policies is unlikely to be of fiscal relations that does not undermine fully implemented or its outcome the incentive for state fiscal policy reform. 69 APPENDIX B: METHODOLOGY Selection of Countries, Lending Twenty of the 89 countries were selected for the Activities, and Economic and evaluation (table B.1), ensuring Regional Sector Work representation (table B.2), as well as a mix of federal and unitary government structures. Some Selection of country cases smaller countries and some postconflict All closed projects (as of December 2006) classi- countries were also selected. Some countries fied as having a decentralization or subnational were selected in coordination with the IEG public theme were selected. These projects were then sector reform and judicial reform evaluations. classified by country and Region. Bank- Together, lending to these 20 countries consti- supported decentralization activities were found tuted about 47 percent of total lending classified in 89 countries. as having a decentralization theme (table B.3). Table B.1: Country Cases Evaluated Albania Madagascar Russia Bolivia Morocco Rwanda Brazil (states: Parana and Bahia) Nepal Sierra Leone Burkina Faso Nicaragua Tanzania Ethiopia Pakistan (Provinces of Sindh and NWFP) Uganda India (states: Andhra Pradesh, Peru Rep. of Yemen Karnataka, and Uttar Pradesh) Philippines Indonesia Table B.2: Number of Focus Countries Number of Percentage of total commitment containing Region focus countries decentralization activities reviewed Sub-Saharan Africa 7 42 East Asia and Pacific 2 59 Europe and Central Asia 2 36 Latin America and the Caribbean 4 41 Middle East and North Africa 2 50 South Asia 3 84 Source: World Bank data. 71 DECENTRALIZATION IN CLIENT COUNTRIES Table B.3: Activities with Some Decentralization Support, by Commitment, Fiscal 1990­2007 Amount of commitment Percentage Instrument for decentralization of total Number of Percentage ($ million) commitment activities of total activities Development Policy Loan 10.498 47.6 63 31.0 INV 11.572 52.4 140 69.0 22.07 100 203 100 Source: IEG sample activity list. In the 20 countries, all lending activities · Was the work undertaken in consultation with approved after fiscal 1990 and closed by fiscal other donors? 2007 were reviewed in the public, health, education, urban, and water sectors to confirm The evaluation team reviewed 222 projects'/ whether they supported some aspects of programs' (with significant decentralization decentralization (using the definition adopted components) appraisal and completion reports for this study, see chapter 1). This resulted in a to understand how the Bank supported sample of 203 projects. About one-third of the decentralization in the 20 countries. The analysis: identified activities involved Development Policy Loans (DPLs). The high percentage of DPLs can · Separated decentralization-related objectives be explained by the increased use of the instru- (and classified them as fiscal, administrative, po- ment for decentralization reform since the late litical, or purely capacity building at the sub- 1990s and by the inclusion of some large loans to national or local government level) the Russian Federation, Indonesia, and Brazil. · Examined design issues, including (i) consider- ation of and addressing capacity-related risks; Analytic Approach (ii) consideration of equity, gender, and envi- All decentralization economic and sector work ronmental risks; (iii) consultation with local stake- (ESW) (with significant analysis of decentraliza- holdersandwithotherdonors;and(iv)treatment tion issues) prepared for the 20 country cases of monitoring and evaluation indicators was identified (more than 40 pieces). An external · Examined achievement of stated objectives re- evaluator reviewed a third of the cases; the rest lated to decentralization (used only for trian- were reviewed by IEG evaluators. Indicators used gulating the country-level findings), based on for the assessment were as follows: information in the Implementation Completion Reports (ICRs), IEG reviews of ICRs and Pro- · Did the work discuss all aspects of decentral- ject Performance Assessment Reports. ization (as relevant)? Did it discuss country- specific issues (political economy issues, fiscal The evaluation reviewed all country strategies costs of decentralization)? Did the ESW focus during the evaluation period, decentralization- on issues related to monitoring and evaluation? focused ESW, and other activities to better · Did the work reflect an understanding of what understand how the Bank supported decentral- the Bank (and other key donors or actors) was ization (decentralization frameworks and stated doing in the area and recommend how the sectors) at a country level. This also helped the effectiveness of Bank support could be evaluators better understand the timeliness of improved? response (mainly in terms of an ESW), framing of · Did the ESW provide clear and phased rec- objectives at the country level, choice of instru- ommendations linked to the analysis? ments as reflected in country strategies, and · Were client stakeholders closely involved in whether the Bank focused on monitoring and the preparation of the ESW? evaluation in this area. 72 APPENDIX B: METHODOLOGY From the above analysis, the evaluation ii. Enhanced administrative capacity and established objectives of Bank support in each accountability country (thus mitigating the attribution issue). b. Education and urban service sectors i. Generated resources for service delivery at The evaluation assessed the quality of support, the local levels, reducing reliance on local using a three-point scale (high, medium, and government transfers low) to rate: ii. Strengthened institutions, capacity, and in- centives for improved service delivery · Quality of ESW during evaluation period iii. Improved accountability to citizens. · Bank support reflecting country-specific cir- cumstances, particularly client participation in The four-point scale used to rate progress in each the design of Bank support, which was reflec- case was high, substantial, modest, and negligi- tive of country context (identification of risks and ble. The rating was based on a before-and-after how they were addressed) and of Bank diagnosis analysis. Indicators were aggregated to a country · Internal consistency of Bank support at a coun- level for each area. try level · Results orientation of Bank support The evaluation did not aggregate the ratings for · Donor collaboration at country level. quality and decentralization frameworks to arrive at a single country-level rating for results, given The evaluation assessed results in each country that the final results of projects and programs are against stated objectives in two areas, using a four- influenced by extraneous variables (such as point scale to rate: wavering political commitment, for instance, caused by emerging political opposition) that are a. Decentralization frameworks beyond the Bank's control. i. Strengthened intergovernmental relations 73 ENDNOTES Chapter 1 by design not consistent with the Decentralization Act. 1. Subsidiarity is the principle that decisions about The Tanzanian authorities designed it to explore alter- resource management should be made and executed native mechanisms for increasing local participation at the lowest possible level if it can be shown that local and better service delivery, notably in education. For that capacity at this level is competent enough. reason, the project was clearly labeled as a pilot. At 2. Management notes that, in its sector and the- completion, IEG rated it was satisfactory in terms of out- matic coding framework, sectors indicate which part of come, with sustainability rated as likely. The Tanzanian the economy receives support. Themes correspond to authorities adopted many of the elements of the pilot the goals of Bank activities. Both sector and thematic and integrated them into their existing decentralized project coding are reviewed and validated centrally. structure countrywide. Management maintains that decentralization, like en- 4. See the project ICR (World Bank 2005a, p. 3), vironment, is a theme, that there are no known biases which notes that the decision to pursue the establish- in assigning the decentralization thematic code to Bank ment of a regional health authority structure in the activities, and that there are no pure investments in de- Tirana region (as part of a decentralization strategy) centralization. proved to be potentially problematic, especially given the 3. Activities focused on strengthening revenue lack of prior diagnostic or sector work. The Bank also sources and collection ability of urban governments, im- supported the Health Insurance Institute that was even- proving their efficiency, and strengthening their ca- tually to develop into the sole purchaser of health care pacity for borrowing. services from providers. See also the 2006 Albania Health 4. For example, in Tanzania during the 1980s, the IMF Sector Note (World Bank 2006a), which concludes that and the World Bank supported decentralization in favor such decentralization in Tirana calls for careful evalua- of urban districts, particularly in Dar es Salaam, through tion and corrective action based on the evaluation results the economic recovery program that supported policy (p. xii) This said, although IEG's ICR Review noted the reform for this purpose and through the water and confusion at the institutional level and proposed an sanitation project that provided support to the sector. audit, it also prematurely added that the project had laid the groundwork for a revised institutional architecture Chapter 2 for a decentralized health system with the Ministry of 1. A relevant ESW piece is one that discusses the im- Health as policy body and regulator, the Health Insur- plications of decentralization for the client's develop- ance Institute as insurer and purchaser of services and ment agenda and makes appropriate recommendations. regional health authorities and planning (based on the 2. The United Nations Capital Development Fund had Tirana Regional Health Authority model). financed a project that piloted the provision of support 5. Similarly, the 2004 ESW report (World Bank 2004g) to local governments in a limited number of districts. notes little progress in enabling local government access The Bank, along with other donors, supported gov- to private capital based on creditworthiness and notes ernment actions to replicate the pilot in a larger num- that such finances are mainly sourced from development ber of districts and subsequently at the national level. assistance. 3. Management notes that Tanzania's Human Re- 6. The inability of private financial intermediaries to sources Development I Project, which the International become depository banks for LGUs is a key structural Development Association supported with a credit, was impediment to their entry. 75 DECENTRALIZATION IN CLIENT COUNTRIES 7. See the Project Appraisal Document (World Bank development projects from a menu of choices. Local gov- 2006i, p. 13), which notes that "loans and borrowing only ernments were limited to providing the recurrent costs account for 1.8 percent of LGUs' total revenue, and of the completed activity and fulfilling some monitoring debt service only accounts for 2.8 percent of their total responsibilities. The funding went directly from the expenditure." More recently, a Joint IFC-World Bank project office to the local community's bank account, and Subnational Finance Department was established in community groups undertook the procurement. Thus, 2007 to provide states, provinces, municipalities, and although it is not inconceivable that the project had a their enterprises with financing and access to capital mar- positive, if indirect, impact on the capacity of local gov- kets--without sovereign guarantees. Such an initiative ernments, it also provided for a lesser role for local gov- can perhaps provide a framework for lending to local ernments in determining the allocation of resources governments within which both GFIs and private fi- than would otherwise have been given to them. nancial intermediaries can have a role. 13. See, for example, Saxena (2003), which notes that 8. In the Philippines, at the time of the IEG mission, although there may be benefits to both systems, there decentralization was led by the regional urban unit, is need to sharpen the institutional links between po- without the full benefit of PREM experts. litical decentralization, in the form of Panchayati Raj, and 9. The Municipalities Law of 1997 lists a set of re- administrative decentralization, in the form of user sponsibilities of municipal governments but does not committees promoted by the external donor projects. explain what is mandatory or specify a minimum set of Saxena adds that this is important because in several de- obligations for municipal governments (World Bank velopment projects, such as drinking water, health, wa- 2003a, p. 11). tershed development, and primary education, 10. The Poverty Reduction Support Credit (PRSC) I government has relied on committees that are inde- included a prior action for the second tranche that pendent of panchayats. Often in the same Region, the would have required amendments to the Law of Mu- World Bank has promoted different committees, one for nicipalities to balance municipal expenditures with re- forests, the second for drinking water, and the third for sources transferred to municipalities. The proposed education, and most of these are distinct from pan- presidential decree, however, could not amend the chayats. See also IEG's report on the effectiveness of previous law, and the condition was therefore revised World Bank support for community-based and-driven to allow a decree that would clarify some of the budget development (IEG 2005b, pp. 31­32). responsibilities of municipalities and help to neutralize 14. The projects referred to in this paragraph are ac- the transfers to the municipalities. The PRSC II program tive and the results have not been assessed by IEG. document indicates that further progress in this area Management in its response also pointed out that "a re- would be a priority for future PRSCs, including the pres- cent review of CDD and social fund activities in Africa entation of a new law proposal to the National Assem- showed that 71 percent of active operations transferred bly, as part of a revised strategy and policy. resources to activities that are undertaken through local 11. Management notes that support from Social De- government planning. In addition, two-thirds of these velopment and PREM staff to Nicaragua on decentral- have local governments manage these CDD resources ization was well coordinated across operations. Staff from themselves." these networks worked together in supporting 15. The framework, developed by a group of devel- Nicaragua's Municipal Development Project, as well as opment partners including the Bank, provides 28 high- the PRSC. The unexpected size of the transfers in the level indicators to measure progress and monitor 2003 law did raise fiscal issues; strong cross-network in- performance of PFM systems, processes, and institutions. tegration was critical for the policy dialogue with gov- They measure the extent to which the PFM system is an ernment authorities to address not only fiscal issues but enabling factor for achieving development outcomes. also local development priorities. The Bank has integrated the use of the indicators in PFM 12. For example, the Ethiopia Rehabilitation and So- analytic work, as they facilitate structured measure- cial Development Fund Project established a central- ment of progress. ized project appraisal and approval system under which 16. Such joint diagnostic work has the potential to local communities established community development lead to harmonization of procedures, as is occurring in groups to identify, seek financing for, and implement local some countries, including Albania. 76 ENDNOTES 17. See the survey on support to local governance Uganda, district chief financial officers reported that if and decentralization (DEGE Consult and NCG 2006). the center realizes less tax revenue than expected, or The report notes that the dominant approach for de- if actual donor disbursements fall short of that bud- livering donor support for decentralization and local gov- geted, releases can often be delayed or less that that bud- ernance is still a project approach, where each donor geted. Local government stakeholders noted the supports a discrete project, with its own institutional importance of monitoring fiscal gaps at the local level, arrangements for project implementation. which is the difference between the budgeted amount and that received. Chapter 3 9. See "Philippines Decentralization and Service De- 1. For example, the evaluation does not assess livery" (ADB and World Bank 2004); see also 1993 Fis- whether the control of administrative staff was decen- cal Decentralization Study and Devolution and Health tralized, which while an important area for decentral- Services (World Bank 1994a). The Philippines also par- ization, is one where the Bank has not provided much ticipated in an ASEM grant for strengthening the policy assistance. framework for decentralization. In 2000, a public sec- 2. In Burkina Faso, the Community-Based Rural De- tor reform loan was initiated to help the government velopment Project and Urban Project II helped revise reassess the scope of decentralization in view and sup- existing laws. In Rwanda, the Community Reintegra- port the revision of the Internal Revenue Allotment tion and Development Project laid the foundation for formula with a view to improve its equity, incentives for the government's approach to decentralization and the local revenue mobilization, as well as its overall fiscal sus- subsequent legislation. tainability. Consensus was difficult to achieve on the in- 3. Four regions and Addis Ababa have passed enabling tractable issues surrounding the Internal Revenue legislation for municipalities with support from the Allotment, and this activity was dropped. Bank's PRSCs. 10. See Fiscal Federalism in Theory and Practice 4. The effectiveness of the law will depend on how (Ter-Minassian, 1997, p. 49). See Taliercio (2005): "The it is implemented. For example, the state of Karnataka fact that subnational governments have both limited issued a notification in 2005 to allow citizens to seek in- powers to raise revenues and limited capacity to collect formation henceforth only on one subject through a sin- them poses profound questions about the actual ben- gle application of less than 150 words. What constitutes efits of decentralization." See also http://www.world "one subject" is left to the discretion of the Right to In- bank.org/wbi/mdf/mdf1/revenue.htm: "Own-source rev- formation Commission. At the same time, there is a move enues should constitute a very significant share of local to allow requests through electronic mail in the state of finance. Governments spend more wisely and are more Bihar. accountable for money they are responsible for raising." 5. The Bank's 2004 ESW recommended as a medium- At a macroeconomic level, as discussed in chapter 2, en- term measure "designing and implementing basic laws couraging own-source revenue at the local levels is also and regulations on local government." As a monitoring noted as being not without risks in contexts of fiscal indicator for Bank PRSCs, "Basic laws and regulations stress (and where revenue generation is critical for the on local government administration and fiscal autonomy central government), weak capacity, or poor account- passed." IEG understands that these laws were passed ability (Raich 2005). in February 2008. 11. Such support has involved assisting Morocco 6. The findings on Bank support for Indonesia are and India (at a state level) to strengthen and implement based on IEG's Country Assistance Evaluation: The the value-added tax framework and share the revenues, World Bank in Indonesia, 1999­2006 (2007g). improving the policy framework for tax assignment in 7. The first phase of the program was with Bank Albania, Bolivia, Ethiopia, Pakistan, Sierra Leone, Rus- support (2001­04), but the government has since ex- sia, and Uganda, and enhancing the own-source revenue tended the program with its own funds. of urban and rural municipalities through project-level 8. There has been resistance from the education assistance in several other countries. sector, where discretion is not allowed on the grounds 12. Own-source revenues of subnational govern- that it may reduce the chances of meeting the Millen- ments overall have not increased in Brazil. nium Development Goals. During field assessments in 13. The new fiscal decentralization policy introduces 77 DECENTRALIZATION IN CLIENT COUNTRIES criteria tax sharing with local government, for empow- on strengthening meritocracy in the civil service as a ering local authorities to set tax and license rates and whole. See IEG's public sector evaluation (IEG 2008), levy service fees, and for devolving property tax to local which examines results of Bank support for civil service government. Legislation is pending to allow districts to reform in general. collect revenue on property. 19. See WDR (World Bank 2004j, p. 185). Laws typ- 14. Field assessments in Uganda show that district ically define broad responsibilities, leaving regulations governments are expected, for example, to cofinance to define competencies of each level of government and the costs of several investment activities through their their expenditure assignments. own-source revenue. When such revenue is not avail- 20. Field assessments in Uganda indicate that service able, they divert transfers meant for lower levels of gov- quality is constrained by the ceilings imposed on hiring ernment to fulfill such obligations, and lower levels of and offering competitive remuneration. government then get decreased funding. 21. In several countries, Bank support has helped to 15. A formula-based allocation of the General-Purpose pass or modernize procurement laws or establish pro- Grant was initiated in fiscal 2005­06, although the final curement agencies. IEG, however, looked for evidence implementation of a formula for the General-Purpose of results at the subnational level, without which the re- Grant and the Local Administration Grant will not take sult was rated as modest. place until June 2007. The Local Government Fiscal Re- 22. As a result, states are regularly meeting their view 2005 notes that the General-Purpose Grant has not debt service obligations to the National Treasury and the compensated local governments adequately for the de- latter continues to apply penalties to any states that fail cline in own revenue sources. to comply. The number of individual states generating 16. Between 2000 and 2001, the Bank provided two primary deficits decreased from seven in 2001 to three adjustment loans in support of Brazil's fiscal perform- in 2002 and to only one in 2003. In January 2003, for ex- ance and performance of the states in particular. The ample, constitutional federal transfers were blocked Program for Fiscal and Administrative Reform Special when the state of Rio de Janeiro failed to meet its debt Sector Adjustment Loan--$505 million--consisted of obligations. That state responded by restarting its sched- policy actions that introduced and initiated the gov- uled payments. ernment's fiscal stability plan, implementation of the 19th 23. See public financial management and accounta- amendment to the constitution. The second loan that bility assessments conducted for different provinces in the Bank supported for $750 million intended to deepen Pakistan (World Bank 2007h, 2007i, 2007j). and improve fiscal reforms and supported the passing of the Fiscal Stability Plan and the Law of Fiscal Re- Chapter 4 sponsibility among other actions. 1. IEG's primary education evaluation (IEG 2006a) re- 17. With Bank support, the government adopted viewed decentralization and school accountability/com- and published a supreme decree requiring independ- munity accountability, focusing on their impact on ent and publicized credit ratings of municipal and re- educational access, quality, equity, and learning out- gional governments for them to borrow amounts above comes. It found that decentralization and school ac- the ceiling established in the borrower's annual public countability are not necessarily associated with improved indebtedness law. access, quality, or learning outcomes, and may lead to in- 18. In several countries, administrative devolution oc- equities between poor and non-poor communities. In curred (that is, subnational governments were given con- contrast, this evaluation of decentralization focuses on in- trol over local staff, but the Bank was not involved in this termediate outcomes likely to improve service delivery decision [Indonesia, Pakistan, the Philippines, and (generation of discretionary resources for local govern- Uganda]). In a few countries such as Morocco, Pak- ments for service delivery; strengthened institutions, ca- istan, and Nepal and sometimes through specific sec- pacity, and incentives for better service delivery; enhanced tors such as in Bolivia, the Bank has tried to strengthen accountability of local governments to citizens for service administrative deconcentration or decentralization, but delivery). It points to the lack of documentation of links overall such support has been highly limited, given the between decentralization and better service delivery and political economy issues surrounding civil service reform, finds that the desired intermediate outcomes were not with focus being more on containing the wage bill and achieved in three of the six cases reviewed. 78 ENDNOTES 2. Morocco, selected to ensure some Regional rep- centralizationto26provincescoveredbyTEEPwillbegiven resentation, is the only sample country where the find- top priority. However, the decentralization was focused ings are based entirely on a desk review. on a school-autonomy model for the most part. 3. Field visits also were conducted in Peru, the Philip- 10. See IEG (2007d, pp. 21­22), and IEG (2006c), pines, Russia, and Tanzania. The evaluation also drew which notes concerns of significant misuse of funds in on field-based case studies of Bank support for primary the sector, pp. 6, 35­36, and footnotes 43 and 49. education in Pakistan and Peru, conducted in 2006 as 11. See Basic Education Project ICR (World Bank background for the IEG evaluation of primary education 2004f). The focus of the project was to involve com- (IEG 2007d, 2007e). munities in the management of schools, but this was not 4. This framework is consistent with the 1999 Edu- achieved. cation Sector Strategy and its 2005 Sector Strategy 12. See Punjab Education Second Policy Development Update. Credit ICR (World Bank 2006h). 5. The 2001 CAS for Morocco relies on decentral- 13. See IEG (2007e, pp. viii, 10, and p. 27, footnote ization as a process that will help to reduce poverty, and 24) and ESW on education in Peru (World Bank 2007m, specifically speaks about regional decentralization and chapter 9), which confirms that although there is great community schools. The 1999 Policy Reform Support potential for voice of, and accountability to, parents, in Loan included a prior action that required ministerial reality accountability has not been built up. instructions to decentralize decision making in school 14. See also USAID (2006), which found that capac- mapping and project implementation to regional au- ity of local government staff had not been strengthened. thorities. Bank investment projects supported decen- 15. See the evaluation of the Bank's fiduciary in- tralization of in-service training of teachers and school struments (IEG 2006e), which discusses the progress principals, budgeting and decision making to provinces, made in this area in the Philippines. and school construction to communities. 16. Some nongovernmental organizations in Tanza- 6. The PSRL IV dealt with the decentralization of nia protested to the IEG mission that steps taken to iden- the social programs, and, inter alia, aimed to support tify and address such leakages were inadequate. the establishment of a clear legal basis for the decen- 17. This could involve measures such as establishing tralization of education and improve clarity in the roles minimum service standards and holding national assess- and responsibilities of different levels of government. ment tests to provide comparative benchmarks against Other investment projects also supported decentral- which citizens can compare the performance of their gov- ization in the sector. ernment.Itcouldalsoinvolvehelpingtoestablishpilotdis- 7. In the Philippines, the 1991 Code did not devolve pute-resolutionmechanismsinthesectorortostrengthen the education sector, although the construction and local-level accountability or justice institutions. maintenance of school buildings was devolved to local 18. Strengthening local-level Parliamentary Accounts government units. A Governance of Education Act fo- Committees has been tried in a few countries such as Pak- cused on decentralization in the sector, namely through istan, but has not yet had an evident impact on local gov- school-based management systems (see note 12 also). ernments. See also EU (2007, p. 73), where it notes that 8. These varying project and donor requirements led assessments have shown that citizen participation is im- to one weak district in Tanzania (Kinodoni) having 950 portantinextractingresourcesfromlocalgovernmentsbut bank accounts. less important in holding local governments accountable. 9. See endnote 3. The Third Elementary Education Pro- 19. See Peru ESW (World Bank 2007m, note 17, p. ject (TEEP) aimed to "introduce systems and procedures 127), which concludes that "the state of decentralization to decentralize selected education functions and the cor- in practice (and even in regulation) is fluid and confused responding resources, to the Divisions and, where feasi- to such an extent that it is impossible to make forecasts as ble, the schools"; and build capacities in support of to whether decentralization is likely to make all of the decentralized implementation and effective schools. Dur- problems listed here worse or better." See also the Philip- ing negotiations, assurances were obtained from the gov- pines PAD for National Program Support for Basic Educa- ernment that it would carry out the project in accordance tion Project (World Bank 2006i, p. 2), which says that with the principles set forth in the Memorandum of Pol- education remains "largely regulated by central prescrip- icy and with the TEEP Guidelines. It was agreed that de- tions- administrative memos and orders." See also the 79 DECENTRALIZATION IN CLIENT COUNTRIES ICR of the Pakistan Second Social Action Program, which approaches. It provides nuanced guidance on the dif- notes that "decentralization of powers could not be insti- ferent actions that should be considered in different cir- tutedinthesocialsectorsaloneinanotherwisehighlycen- cumstances. An Operational Policy is intended as a tralized system of administration, in which lower level more concise statement of the Bank's obligations to all government functionaries were reluctant to exercise pow- Bank operational activities covered by the policy. It pro- ers already delegated to them" (World Bank 2006h, p. 5). vides formal and accountable safeguard policies that are approved by the Board. The Operational Policy is ac- Appendix A companied by a definitions appendix and a Bank pro- 1. The sector strategy outlines intent, discusses op- cedures document, which elaborate on some of the tions, and presents arguments for adopting particular implementation requirements raised by the policy. 80 BIBLIOGRAPHY Acosta, Olga Lucia, and Richard M. Bird. 2003. "The erment, Pluralism and Social Capital." Paper pre- Dilemma of Decentralization in Colombia." ITP Paper sented at the International Seminar of Decentral- 0404, University of Toronto, Toronto. ization and Devolution of Forest Management in ADB (Asian Development Bank) and World Bank. 2004. Asia and the Pacific, Davao, Philippines. "Philippines: Decentralization and Service Delivery: Bahl, Roy, and Jorge Martinez-Vazquez. 2006. "Se- From Promise to Performance--Cross-Cutting quencing Fiscal Decentralization." Policy Research Themes in Local Government Functioning." Report Working Paper 3914, World Bank, Washington, DC. No. 26104-PH, World Bank and Asian Development Baker, Jonathan, Hege Wallevik, James Obama, and Bank, Washington, DC, and Manila. Nazar Sola. 2002. "Local Government Reform in Tan- AfroBarometer. 2006. "Delivery of Social Services on zania: Towards a Greater Interdependency between Mainland Tanzania: Are People Satisfied?" Briefing Local Government and Civil Society at the Local Paper No. 34, AfroBarometer, University of Michigan. Level." http://www.tanzaniagateway.org/contents Ahmad, Ehtisham, and Mercedes García-Escribano. .asp?cpg=8&topicId=17. 2006. "Fiscal Decentralization and Public Subna- ------. 2002. "Decentralization of Governance and De- tional Financial Management in Peru." Working Paper velopment." Journal of Economic Perspectives 16 No. 06/120, IMF, Washington, DC. (4): 185­205. Ahmad, E., G. Brosio, and M. González. 2006. "Uganda: Ballivian, Ampara, and A. Spillmbergo. 2004. "Municipal Managing More Effective Decentralization." IMF Decentralization in Nicaragua: Fiscal and Institu- Working Paper 06/279, Washington, DC. tional Issues." Policy Note, World Bank, Washing- Ahmad, Junaid, Shantayanan Devarajan, Stuti Kehmai, ton, DC. and Shekhar Shah. 2005. "Decentralization and Ser- Bardhan, Pranab, and Dilip Mookherjee. 2000. "De- vice Delivery." Policy Research Working Paper 3603, centralizing Anti-Poverty Program Delivery in De- World Bank, Washington, DC. veloping Countries." University of California Center Aitken, I. 1998. Decentralization and Reproductive for International and Development Economics Re- Health. Boston, MA: Harvard University School of search (CIDER) Working Paper No. C98-104.1, Berke- Public Health. ley, CA. Albana, Dhimitri, Belinda Ikonomi, and Majlinda Dhuka. Bird, Richard M., and François Vaillancourt. 1994. Fiscal 2007. "The Fiscal Decentralization Initiative for Cen- Decentralization in Developing Countries. Cam- tral and Eastern Europe. Regional Development Pol- bridge, UK: Cambridge University Press. icy Performance in Albania." http://lgi.osi.hu/publi Bird, Richard M., Robert D. Ebel, and Christine I. Wal- cations/2007/368/FDI_Regional_Development_ lich. 1995. "Decentralization of the Socialist State: Albania.pdf. Inter-governmental Finance in Transition Economies." Alonso, Rosa, and Robert Utz. 2003. "Fiscal Aspects of World Bank Regional and Sectoral Studies Series, PRSP Implementation in Tanzania." http://sitere- Washington, DC. sources.worldbank.org/INTPRS1/Re Blaser, Jeremias, Dirk Besdziek, and Sarah Byrne. 2003. sources/383606-1119904390686/bbl101204_tan "Lessons Learned on Decentralization: A Literature zania.pdf. Review." http://www.federalism.ch/files/categories/ Anderson, Jon. 1998. "Four Considerations for Decen- IRCC/. tralized Forest Management: Subsidiarity, Empow- Boadway, Robin. 2006. Intergovernmental Fiscal Trans- 81 DECENTRALIZATION IN CLIENT COUNTRIES fers: Principles and Practice. Washington, DC: World Dhimitri, Albana, Belinda Ikonomi, and Majlinda Dhuka. Bank. 2006. "Regional Development Policy Performance in Braathen, Einar, Amon Chaligha, and Odd-Helge Fjeld- Albania." Fiscal Decentralization Initiative for Central stad. 2005. "Local Governance, Finances and Ser- and Eastern Europe Paper, World Bank, Washington, vice Delivery in Tanzania: A Summary of Findings DC. from Six Councils." Joint Report of NIBR, CMI, and Dickovick, J. Tyler. 2003. "Centralism and Decentraliza- REPOA, Dar es Salaam, Tanzania. tion." In Unitary States: A Comparative Analysis of Brosio, Giorgio. 2002. "An Evaluation of the World Bank Peru and Senegal, ed. J. Tyler Dickovick. Princeton, Support for Decentralization in the Middle East and NJ: Princeton University. North Africa Countries of Algeria, Egypt, Morocco, Dillinger, William. 1994 "Decentralization and Its Im- Tunisia, West Bank." Background Paper for Indepen- plications for Service Delivery." Urban Management dent Evaluation Group, World Bank, Washington, DC. Institute Paper, World Bank, Washington, DC. Center for Democracy and Governance. 2000. "De- Dillinger, William, and Steven B. Webb. 1999. "Fiscal centralization and Democratic Local Governance Management in Federal Democracies: Argentina and Programming Handbook." Technical Publication Se- Brazil." Policy Research Working Paper 2121, World ries, CDG, Washington, DC. Bank, Washington, DC. Chattopadhyay, R., and Esther Duflo. 2004. "Impact of Dunn, Jonathan, and Deborah Wetzel. 1998. "Fiscal De- Reservation in Panchayati Raj: Evidence from a Na- centralization in Former Socialist Economies: tionwide Randomized Experiment." Economic and Progress and Prospects." Paper presented at NTA Political Weekly 39 (9): 979­86. Conference, Atlanta. ------. 2001. "Women as Policy Makers: Evidence from Echeverri-Gent, John. 1993. The State and the Poor: Pub- an India-Wide Randomized Policy Experiment." Na- lic Policy and Political Development in India and tional Bureau of Economic Research Working Paper the United States. Berkeley: University of California 8615. http://www.nber.org/papers/w8615. Press. Cheema, G. Shabbir, and Dennis A. Rondinelli, eds. Eskeland, G., and D. Filmer. 2002. "Autonomy, Partici- 2007. Decentralizing Governance: Emerging Con- pation, and Learning in Argentine Schools: Findings cepts and Practices. Washington, DC, and Boston, and Their Implications for Decentralization." Policy MA: Brookings Institution Press and Harvard Research Working Paper 2766, World Bank, Wash- University. ington, DC. Crook, Richard, and Alan Sverrisson. 2001. "Decentral- Evans, Anne, and Nick Manning. 2004. "Administrative isation and Poverty-Alleviation in Developing Coun- Decentralization: A Review of Staffing Practices dur- tries: A Comparative Analysis or Is West Bengal ing Decentralization in Eight Countries." Draft Work- Unique?" Institute of Development Studies Working ing Paper, World Bank, Washington, DC. Paper 130, Brighton, U.K. EU (European Commission). 2007. "Supporting De- DEGE Consult and NCG (Nordic Consulting Group). centralization and Local Governance in Third Coun- 2006. "Survey on Support to Local Governance and tries." Tools and Methods Series, Reference Decentralization." For the Informal Donor Working Document No. 2, Brussels. Group on Local Governance and Decentralisation. Faguet, Jean-Paul. 2001. "Does Decentralization In- http://www.euromedina.org/bibliotheque_fichiers/Do crease Government Responsiveness to Local Needs? c_SurveyLocGovDec.pdf. Evidence from Bolivia." Policy Research Working Devarajan, S., V. Swaroop, and H.F. Zou. 1998. "Fiscal De- Paper 2516, World Bank, Washington, DC. centralization and Economic Growth: A Cross- Faguet, Jean-Paul, and Fabio Sanchez. 2006. "Decen- Country Study." Journal of Urban Economics, 37 (2): tralization's Effects on Educational Outcomes in Bo- 313­44. livia and Colombia." LSE Research Online Working Development Partners Working Group on Decentral- Paper, World Bank, Washington, DC. ization and Local Governance. 2007. "Alignment Fanthorpe, Richard. 2006. "On the Limits of Liberal Strategies in the Field of Decentralisation and Local Peace: Chiefs and Democratic Decentralization in Governance: A Review of Country Practices and Ex- Post-War Sierra Leone." African Affairs 105 (418): periences." Philippines Development Forum, Manila. 27­49. 82 BIBLIOGRAPHY Fernandez, Eduardo. 2007. "Assess Bank Support for Im- Hauge, Arild. 2001. "Strengthening Capacity for Moni- proving Governance, Including Capacity Building, toring and Evaluation in Uganda: A Results-Based Public Sector Reform and Decentralization." Back- Management Perspective." ECD Working Series No. ground Paper for Independent Evaluation Group, 8, World Bank, Washington, DC. World Bank, Washington, DC. Hesselbarth, Susanne, Finn Hansen, and Hans Olsen. Fjeldstad, Odd-Helge, Einar Braathen, and Amon Cha- 2007. "Alignment Strategies in the Field of Decen- ligha. 2006. "Local Government Reform in Tanzania tralization and Local Governance: A Review of Coun- 2002­2005: Summary of Research Findings on Gov- try Practices and Experiences." International ernance, Finance, and Service Delivery." REPOA Brief Development Partner Harmonization for Enhanced 6 (October) 1­8. Aid Effectiveness and Development Partners Work- Fleiner, Thomas, and Flora Musonda. 2002 "Evaluating ing Group on Local Governance and Decentralisa- Bank Support for Improving Governance through Lo- tion. Philippines Development Forum, Manila. calization/Decentralization." Background Paper for IADB (Inter-American Development Bank). 2001. "Sum- Independent Evaluation Group, World Bank, Wash- mary of Findings--Decentralization and Effective ington, DC. Citizen Participation: Six Cautionary Tales." Office of Galiani, Sebastián, and Ernesto Schargrodsky. 2001."Eval- Evaluation and Oversight Report RE-250, Inter- uating the Impact of School Decentralization on Ed- American Development Bank, Washington, DC. ucation Quality." http://www.utdt.edu/Upload/_ ------. 2000. "Subnational Development Strategy." Sus- 115332118904928800.pdf. tainable Development Department Strategy Paper, Gershberg, Alec Ian. 1998. "Decentralization and Re- Inter-American Development Bank, Washington, DC. centralization: Lessons from the Social Sectors in IEG (Independent Evaluation Group). 2008. Public Sec- Mexico and Nicaragua." Working Paper No. 379, Inter- tor Reform: What Works and Why? IEG Study Series. American Development Bank, Washington, DC. Washington, DC: World Bank. Girishankar, Navin, David DeGroot, and T.V. Pillay. 2006. ------. 2007a. "Albania: Public Administration Reform "Measuring Intergovernmental Fiscal Performance in Project." Implementation and Completion Report Re- South Africa: Issues in Municipal Grant Monitoring." view, World Bank, Washington, DC. http://wbln1023 Africa Region Working Paper Series No. 98, World .worldbank.org/oed/oeddoclib.nsf/DocUNIDView Bank, Washington, DC. ForJavaSearch/8525682E006860378525730500691A70 Government of Albania and the World Bank. 2006. "Al- ?opendocument. bania: Public Expenditure and Financial Account- ------. 2007b. "Country Financial Accountability As- ability (PEFA) Public Financial Management sessments and Country Procurement Assessment Assessment." Working Paper No. 39483, World Bank, Reports: How Effective Are World Bank Fiduciary Di- Washington, DC. agnostics?" World Bank, Washington, DC. Government of Punjab (Pakistan) and the World Bank. ------. 2007c. "An Evaluation of Bank Support for De- 2005. "Islamic Republic of Pakistan: Punjab Public Fi- centralization and Strengthening Intergovernmen- nancial Management and Accountability (PEFA) As- tal Relations." IEG Working Paper Series, World Bank, sessment." South Asia Financial Management Report Washington, DC. No. 32569-PAK, World Bank, Washington, DC. ------. 2007d. "Evaluation of the World Bank's Assis- Guadagni, Maurizio, Gabriela Boyer, Aidan Gulliver, tance for Primary Education to Pakistan: A Country Paola Perez-Alemán, and Dagoberto Rivera. 2000. Case Study." IEG Working Paper Series, World Bank, "Local Governments and Sustainable Rural Devel- Washington, DC. opment: Effects and Potential Sustainability of the Mu- ------. 2007e. "Evaluation of the World Bank's Assis- nicipalities Project (Protierra) in Nicaragua." tance for Primary Education to Peru: A Country Case Sustainable Development Network Working Paper, Study." IEG Working Paper Series, World Bank, Wash- World Bank, Washington, DC. ington, DC. Hakielimu. 2005. Three Years of PEDP Implementa- ------. 2007f. "India Tamil Nadu Urban Project (Credit tion: Key Findings from Government Reviews. 1923-IND), Second Tamil Nadu Urban Development http://www. Hakielimu.org/ThreeYearsofPEDPImple Project (Loan 4478-IND)." Project Performance As- mentation2005.pdf. sessment Report, World Bank, Washington, DC. 83 DECENTRALIZATION IN CLIENT COUNTRIES ------. 2007g. The World Bank in Indonesia, oping Countries: Exploring the Impact." OECD-DAC 1999­2006: Country Assistance Evaluation. Wash- Working Paper No. 236, Issy-les-Moulineaux, France. ington, DC: World Bank. Kaiser, Kai. 2006. "Decentralization Reforms." In Ana- ------. 2006a. From Schooling Access to Learning Out- lyzing the Distributional Impact of Reforms: A Prac- comes: An Unfinished Agenda--An Evaluation of titioner's Guide to Pension, Health, Labor Market, World Bank Support to Primary Education. IEG Public Sector Downsizing, Taxation, Decentral- Study Series. Washington, DC: World Bank. ization and Macroeconomic Modeling, eds. Aline ------. 2006b. Madagascar: Country Assistance Eval- Coudouel and Stefano Paternostro, 313­54. Wash- uation. Washington, DC: World Bank. ington, DC: World Bank. ------. 2006c. Pakistan: An Evaluation of the World Katsiaouni, Olympios. 2003. "Decentralization and Bank's Assistance. IEG Study Series. Washington, DC: Poverty Reduction: Does it Work?" Paper presented World Bank. at the Fifth Global Forum on Reinventing Govern- ------. 2006d. Republic of Yemen: Country Assistance ment, Mexico City, November 3­7. Evaluation. Washington, DC: World Bank. Khaleghian, Peyvand. 2003. "Decentralization and Public ------. 2006e. Review of the Bank's Fiduciary Instru- Services: The Case of Immunization." Policy Research ments: An OED Evaluation. Washington, DC: World Working Paper 2989, World Bank, Washington, DC. Bank. King, E., and B. Ozler. 1998. "What's Decentralization ------. 2005a. Albania: Country Assistance Evaluation. Got to Do with Learning? The Case of Nicaragua's Washington, DC: World Bank. School Autonomy Reform." Working Paper, World ------. 2005b. The Effectiveness of World Bank Support Bank, Washington, DC. for Community-Based and -Driven Development. Lakshminarayanan, R. 2003. "Decentralization and Its Im- IEG Study Series. Washington, DC: World Bank. plications for Reproductive Health: The Philippines ------. 2004a. Brazil: Forging a Strategic Partnership Experience." Reproductive Health Matters 11 (21): for Results. An OED Evaluation of World Bank As- 96­107. sistance. Washington, DC: World Bank. Lamchovichina, Elena, Lili Liu, and Mohan Nagarajan. ------. No date. "Progress Report 2000/01.The United 2006. "Subnational Fiscal Sustainability Analysis: Republic of Tanzania." http://povlibrary.worldbank What Can We Learn from Tamil Nadu?" Policy Re- .org/files/Tanzania_PRSP_ProgRep.pdf. search Working Paper No. 3947, World Bank, Wash- ------. 2004c. The World Bank's Assistance to Pri- ington, DC. mary Education: An OED Portfolio Review. Wash- Levy, Brian. 2004. "The Governance of Service Delivery ington, DC: World Bank. in Europe and Central Asia." Working Paper No. Institute of Development Studies. 2007. "Decentraliz- 36141, World Bank, Washington, DC. ing Service Delivery?" ed. Mark Robinson, special Lewis, Blane D. 2003. Minimum Local Public Service issue. IDS Bulletin 38 (1). Delivery Standards in Indonesia: Fiscal Implications Jeanrenaud, Claude. 2003. "Evaluation of World Bank and Affordability Concerns. Research Triangle Park, Support for Decentralization in Sub-Saharan African NC: Research Triangle Institute International. Countries." IEG Background Paper, World Bank, Litvack, Jennie, and Jessica Seddon. 1999. "Decentral- Washington, DC. ization Briefing Notes." Working Paper, World Bank Jette, Christian. 2005. "Democratic Decentralization Institute, Washington, DC. and Poverty Reduction: The Bolivian Case." New Mackay, Keith. 2007. How to Build M&E Systems to Sup- York: United Nations Development Programme. port Better Government. Washington, DC: World Jiminez, E., and Y. Sawada. 1999. "Do Community- Bank. Managed Schools Work? An Evaluation of El Sal- Mahar, Dennis J., and William R. Dillinger. 1983. "Fi- vador's EDUCO Program." World Bank Economic nancing State and Local Government in Brazil Recent Review 13 (3): 415­41. Trends and Issues." Working Paper 612, World Bank, Jutting, Johannes, Céline Kauffmann, Ida McDonnell, Washington, DC. Holger Osterrieder, Nicolas Pinaud, and Lucia Weg- Manor, James. 2003. Local Governance. New Delhi, ner. 2004. "Decentralization and Poverty in Devel- India: Swedish International Development Authority. 84 BIBLIOGRAPHY Marques, José, and Ian Bannon. 2003. "Central Amer- Shah, Anwar, ed. 2006a. Fiscal Management. Public ica: Education Reform in a Post-Conflict Setting-- Sector Governance and Accountability Series. Wash- Opportunities and Challenges." CPR Working Paper ington, DC: World Bank. No. 4, World Bank, Washington, DC. ------. 2006b. Local Governance in Developing Coun- Momoniat, Ismail. 2002. Understanding Fiscal Decen- tries: Public Sector Governance and Accountabil- tralization in South Africa. Washington, DC: World ity Services. Washington, DC: World Bank. Bank. ------. 2004. "Fiscal Decentralization in Developing and OECD-DAC. 2004. "Lessons Learned on Donor Sup- Transition Economies: Progress, Problems, and the port to Decentralization and Local Governance." Promise." Policy Research Working Paper Series 3282, DAC Evaluation Series, OECD, Paris. World Bank, Washington, DC. OECD Donor Working Group on Local Governance ------. 1998. "Balance, Accountability, and Respon- and Decentralization. 2006. Survey on Support to siveness Lessons about Decentralization." Policy Re- Local Governance and Decentralization. Oslo: search Working Paper, World Bank, Washington, DC. Danish Environmental Cconsulting Group and Smoke, Paul. 2001. "Fiscal Decentralization in Developing Nordic Consulting Group. Countries: A Review of Current Concepts and Prac- Okidi, John A., and Madina Guloba. 2006. Decentral- tice, Democracy." Governance and Human Rights ization and Development Issues: Emerging Issues Programme Paper Number 2, United Nations Re- from Uganda's Experience. Kampala, Uganda: Eco- search Institute for Social Development, New York. nomic Policy Research Center. Smoke, Paul, and Yun-Hwan Kim, eds. 2002. Intergov- Peterson, George E., and Particia Clarke Annez. 2007. Fi- ernmental Fiscal Transfers in Asia: Current Prac- nancing Cities: Fiscal Responsibility and Urban In- tice and Challenges for the Future. Manila: Asian frastructure in Brazil, China, India, Poland and Development Bank. South Africa. New Delhi, India, and Washington, Spahn, Paul Bernd. 2002. "World Bank Support for De- DC: Sage Publications and World Bank. centralization/Localization in the ECA Region--Second Prud'homme, Remy. 1995. "The Dangers of Decentral- Report: Lessons Learned and Recommendations." ization." World Bank Economic Observer 10 (2): Background Paper for Independent Evaluation Group, 201­20. World Bank, Washington, DC. Raich, Uri. 2005. "Fiscal Determinants of Empower- Steiner, Susan. 2006. "Decentralization in Uganda: Ex- ment." Policy Research Working Paper No. 3705, ploring the Constraints for Poverty Reduction." Ger- World Bank, Washington, DC man Institute of Global and Area Studies Working REPOA. No date. "Is Government Managing Money Paper No. 31, Hamburg. Well? Findings from Recent Audit Reports." http:// Steinich, Markus. 2000. "Monitoring and Evaluating www.repoa.or.tz/tgn/index.php. Support to Decentralization: Challenges and Dilem- Reserve Bank of India. Various years. State Finances-- mas." Paper presented at European Support for De- A Study of Budget. Mumbai: Reserve Bank of India. mocratic Decentralization and Municipal Rodden, Jonathan A., and Gunnar S. Eskeland. 2003. Fis- Development--A Contribution to Local Develop- cal Decentralization and the Challenge of Hard ment and Poverty Reduction, Maastricht, June 14­15. Budget Constraints. Cambridge, MA: MIT Press. Taliercio, Robert R. 2005. "Substantial Own-Source Rev- Rondinelli, Dennis, and Shabbir Cheema, eds. 2007. De- enue: Getting Policy and Administration Right." In centralizing Governance Emerging Concepts and East Asia Decentralizes: Making Local Government Practices. Washington, DC: Brookings Institution. Work, 107­28. Washington, DC: World Bank. Saxena, N.C. 2003. "Issues in Panchayats." Tanzi, V. 1996. "Fiscal Federalism and Decentralization: http://www.panchayats.org/downloads/Panchay- A Review of Some Efficiency and Macroeconomics ats%20NC%20Sax.PDF. Aspects." In Annual World Bank Conference on Schneider, Aaron. 2003. "Who Gets What from Whom? Development Economics, eds. M. Bruno and B. The Impact of Decentralisation on Tax Capacity and Pleskovic, 295­316. Washington, DC: World Bank. Pro-Poor Policy." IDS Working Paper 179, Institute of Ter-Minassian, Teresa. 1997. Fiscal Federalism in The- Development Studies, Brighton, U.K. ory and Practice. Washington, DC: IMF. 85 DECENTRALIZATION IN CLIENT COUNTRIES Tidrick, Gene. 2007. "Bank Support for `Pro-Poor' for Development Studies. Oslo: Norwegian Insti- Growth and Investment Climate." Background Paper tute of International Affairs. for Independent Evaluation Group, World Bank, Verheijen, Tony. 2007. "Administrative Capacity in the EU Washington, DC. Member States: The Limits of Innovation?" World Tommasi, Mariano. 2002. "Evaluation of World Bank Bank Working Paper No. 40559, World Bank, Wash- Support for Decentralization in the LAC Region." ington, DC. Background Paper for Independent Evaluation von Braun, Joachim, and Ulrike Grote. 2000. "Does De- Group, World Bank, Washington, DC. centralization Serve the Poor?" Paper presented at the United Nations. 2005. UN Poverty Report. New York: IMF Conference on Fiscal Decentralization, Wash- United Nations. ington, DC, November 20­21. United Nations Economic and Social Council. 2005. "De- Webb, Steven B. 2004. "Fiscal Responsibility Laws for Sub- centralization for Poverty Reduction." Paper pre- national Discipline: The Latin American Experience." sented at meeting of Economic and Social Policy Research Working Paper 3309, World Bank, Commission for Asia and the Pacific: Committee on Washington, DC. Poverty Reduction, Bangkok, Thailand, November Weber-Venghaus, Svenja. 2007. "Bank Assistance to Ed- 23­25. ucation, Health and Gender in Ethiopia." Back- UN-Habitat. 2007. International Guidelines on De- ground Paper for Independent Evaluation Group, centralization and the Strengthening of Local Au- World Bank, Washington, DC. thorities. New York: United Nations. Wetzel, Deborah. 2001. "Decentralization in the Tran- United Republic of Tanzania. 2007. "Local Government sition Economies: Challenges and the Road Ahead." Reform Program--Embedding Decentralization by Working Paper, World Bank, Washington, DC. Devolution across the Government." Main Assess- Wiesner, Eduardo. 2003. "Bank Lending for Subnational ment Report, Prime Minister's Office, Regional Ad- Development: The Policy and Institutional Chal- ministration and Local Government, Dodoma. lenges." IADB Office of Evaluation and Oversight ------. 2005. "Local Government Fiscal Review." Pres- Working Paper Series, Washington, DC. ident's Office, Regional Administration and Local Winkler, Donald R. 2005. Understanding Decentral- Government & Ministry of Finance. ization. Washington, DC: USAID. United Republic of Tanzania, International Monetary Winkler, Donald, and Alec Ian Gershberg. 2000. "Edu- Fund, and World Bank. 1999. "Tanzania Enhanced cation Decentralization in Latin America: The Ef- Structural Adjustment Facility Policy Framework fects on the Quality of Schooling." LCSHD Paper Paper for 1998/99­2000/01." World Bank, Washing- Series No. 59, World Bank, Washington, DC. ton, DC. World Bank. 2007a. "Burkina Faso: Seventh Poverty Re- USAID (United States Agency for International Devel- duction Support Credit Project." Project Information opment). 2006. Assessing the Impact of Devolution Document No. AB2729, World Bank, Washington, DC. on Healthcare and Education in Pakistan. Wash- ------. 2007b. "Decentralization and Local Governance ington, DC: The Urban Institute. in MENA: A Survey of Policies, Institutions, and ------. 2006. "Stocktaking on Indonesia's Recent De- Practices--A Review of Decentralization Experience centralization Reforms--Summary Report." Wash- in Eight Middle East and North Africa Countries." ington, DC: USAID. ------. 2007c. "Education Sector Strategy Update." ------. 2000. Decentralization and Democratic Local http://web.worldbank.org/WBSITE/EXTERNAL/TOP Governance Programming Handbook. Technical ICS/EXTEDUCATION/0,contentMDK:20262538~me Publication Series. Washington, DC: Center for nuPK:282402~pagePK:148956~piPK:216618~the Democracy and Governance. SitePK:282386,00.html. van Zyl, Johan, Tulio Barbosa, Andrew N. Parker, and ------. 2007d. "Healthy Development: The World Bank Loretta Sonn. 1995. "A Case Study from Northeast Strategy for Health, Nutrition, and Population Re- Brazil." Policy Research Working Paper 1498, World sults." Board Report No. 39485, World Bank, Wash- Bank, Washington, DC. ington, DC. Vedeld, Trond. 2003. "Democratic Decentralization and ------. 2007e. "Implementation Plan for Strengthening Poverty Reduction: Exploring the Linkages." Forum World Bank Group Engagement on Governance and 86 BIBLIOGRAPHY Anti-Corruption." Operations Policy and Country ------. 2006f. "Colombia: Municipal Development and Services (SecM2007-0425), World Bank, Washing- Urban Infrastructure Services Development Pro- ton, DC. jects." Project Performance Assessment Report No. ------. 2007f. "Local Government Discretion and Ac- 36435, World Bank, Washington, DC. countability: A Local Governance Framework." So- ------. 2006g. Making Services Work for the Poor in cial Development Department in collaboration with Indonesia: Focusing on Achieving Results on the the Finance, Economics and Urban Department and Ground. Washington, DC: World Bank. the Social Protection Team. Forthcoming. ------. 2006h. "Pakistan: Second Development Policy ------. 2007g. "Morocco Municipal Finance Project II." Credit Project." Implementation Completion and Re- Implementation Completion Report, World Bank, sults Report No. 35465, World Bank, Washington, DC. Washington, DC. ------. 2006i. "Philippines: National Program Support ------. 2007h. "Pakistan--Balochistan Province: Pub- for Basic Education Project." Project Appraisal Doc- lic Financial Management and Accountability As- ument No. 35445-PH, World Bank, Washington, DC. sessment." Integrative Fiduciary Assessment No. ------. 2006j. "Philippines: Support for Strategic Local 39764, World Bank, Washington, DC. Development and Investment Project." Project Ap- ------. 2007i. "Pakistan--Northwest Frontier Province: praisal Document No. 36317, World Bank, Wash- Public Financial Management and Accountability As- ington, DC. sessment." Integrative Fiduciary Assessment No. ------. 2006k. "Tanzania Public Expenditure and Fi- 39759, World Bank, Washington, DC. nancial Accountability Review, Local Government ------. 2007j. "Pakistan--Punjab Province: Public Fi- Fiscal Review--FY05." Report No. 36642-TZ, World nancial Management and Accountability Assessment." Bank, Washington, DC. Integrative Fiduciary Assessment No. 39761, World ------. 2006l. "Tanzania: Public Expenditure Review." Bank, Washington, DC. Local Goernment Fiduciary Assessment from the ------. 2007k. "Russia Education Reform Project." Im- Annual Consultative Meeting, Dar es Salaam, May plementation Completion and Results Report No. 12­13. ICR0000393, World Bank, Washington, DC. ------. 2005a. "Albania: Health System Recovery and De- ------. 2007l. "Strengthening the World Bank En- velopment Project." Implementation Completion gagement on Governance and Anticorruption." Re- and Results Report No. 32718, World Bank, Wash- port No. 39055, World Bank, Washington, DC. ington, DC. ------. 2007m. "Toward High-Quality Education in ------. 2005b. "Bosnia and Herzegovina: Transition Peru: Standards, Accountability, and Capacity Build- Assistance Credit, Public Finance Structural Adjust- ing." World Bank Country Study No. 40557, World ment First and Second Credits, and Social Sector Ad- Bank, Washington, DC. justment Credit Projects." Project Performance ------. 2006a. "Albania Health Sector Note." Report Assessment Report No. 31343, World Bank, Wash- No. 32612, World Bank, Washington, DC. ington, DC. ------. 2006b. "Bolivia: Towards an Inclusive Decen- ------. 2005c. East Asia Decentralizes: Making Local tralization." Institutional and Governance Review Government Work. Washington, DC: World Bank. No. 36285, World Bank, Washington, DC. ------. 2005d. "Empowering People by Transforming ------. 2006c. "Bolivia: Towards a New Social Contract: Institutions: Social Development in World Bank Op- Options for the Constituent Assembly." Social Analy- erations." Board Report No. 31494, World Bank, sis No. 38271, World Bank, Washington, DC. Washington, DC. ------. 2006d. "Bosnia and Herzegovina: Addressing Fis- ------. 2005e. "Fiscal Costs of Structural Reforms." cal Challenges and Enhancing Growth Prospects." Public Expenditure Review No. 30741, World Bank, Public Expenditure and Institutional Review, World Washington, DC. Bank, Washington, DC. ------. 2005f. "Philippines: Economic and Sector Work ------. 2006e. "Brief Local Government Reform in Tan- on Decentralization." Working Paper, World Bank, zania 2002­2005: Summary of Research Findings in Washington, DC. Governance and Service Delivery." World Bank, ------. 2005g. "Reforming Public Institutions and Washington, DC. Strengthening Governance: A World Bank Strategy, 87 DECENTRALIZATION IN CLIENT COUNTRIES Implementation Update: FY05." Public Sector Gov- ------. 2003a. "Morocco--Second Municipal Finance ernance Board Report, World Bank, Washington, DC. Project." Implementation Completion and Results Re- ------. 2005h. "Russia: Fiscal Costs of Structural Re- port No. 26153, World Bank, Washington, DC. forms." Public Expenditure Review No. 30741, World ------. 2003b. "Nicaragua--Poverty Reduction Strategy Bank, Washington, DC. Paper and Joint IDA-IMF Assessment: Second Annual ------. 2005i."Subnational Policy Lending." Good Prac- Progress Report." Poverty Reduction Strategy Re- tice Note for Development Policy Lending, World port No. 27500, World Bank, Washington, DC. Bank, Washington, DC. ------. 2003c. "Reaching the Rural Poor: A Renewed ------. 2005j. "What Determines the Extent of Fiscal De- Strategy for Rural Development." Working Paper centralization? The Russian Paradox." Policy Research No. 30411, World Bank, Washington, DC. Working Paper No. WPS3710, World Bank, Wash- ------. 2003d. "Sustaining Forests: A World Bank Strat- ington, DC. egy." World Bank Strategy Paper No. 35535, World ------. 2004a. "Assessment Report on Minimum Con- Bank, Washington, DC. ditions and Performance Measures of Local Gov- ------. 2003e. Water Resources Sector Strategy. Strate- ernment Authorities, Sikonge District Council." gic Directions for World Bank Engagement. Wash- President's Office, Regional Administration and Local ington, DC: World Bank. http://siteresources Government, LGSP. .worldbank.org/INTINFNETWORK/Resources/water.pdf. ------. 2004b. "Bolivia--Second Programmatic Struc- ------. 2002. "Brazil: Issues in Fiscal Federalism." Re- tural Adjustment Credit for Decentralization." Im- port No. 22523-BR. World Bank. Washington, DC. plementation Completion and Results Report No. ------. 2001a. "Decentralization and Governance: Does 31034, World Bank, Washington, DC. Decentralization Improve Public Service Delivery?" ------. 2004c. Decentralization in Madagascar. Wash- PREM Notes No. 55, World Bank, Washington, DC. ington, DC: World Bank. ------. 2001b. Making Sustainable Commitments: An ------. 2004d. "Designing Development Policy Oper- Environmental Strategy for the World Bank. Wash- ations." Good Practice Note for Development Policy ington, DC: World Bank. Lending, World Bank, Washington, DC. ------. 2001c. "Morocco Country Assistance Strategy." ------. 2004e. "Education Decentralization and Ac- Report No. 22115, World Bank, Washington, DC. countability Relationships in Latin America." Policy ------. 2001d. "Peru Institutional Governance Review," Research Working Paper 3453, World Bank, Wash- Report No. 22637, World Bank, Washington, DC. ington, DC. ------. 2001e. "Tanzania: Public Expenditure Review." ------. 2004f. "Morocco: Basic Education Project." Im- Report No. 22078, World Bank, Washington, DC. plementation Completion Report, World Bank, Wash- ------. 2000a. Cities in Transition: World Bank Urban ington, DC. and Local Government Strategy. Washington, DC: ------. 2004g. "Philippines--Decentralization in the World Bank. Philippines: Strengthening Local Government Fi- ------. 2000b. "Pakistan: Reforming the Provincial Fi- nancing and Resource Management in the Short nances in the Context of Devolution--An Eight- Term." Public Sector Study Report No. 26104, World Point Agenda." PREM Working Paper No. 21362, Bank, Washington, DC. World Bank, Washington, DC. ------. 2004h. "Review of Fiscal Developments and ------. 2000c. Reforming Public Institutions and Budget Management Issues FY03­FY04. Joint Eval- Strengthening Governance: A World Bank Strat- uation Report." Report presented at the Public Ex- egy. Washington, DC: World Bank. penditure Review FY04 Consultative Meeting, Dar es ------. 1999a. "Education Sector Strategy Paper." Sec- Salaam, May 14. tor Strategy Paper, World Bank, Washington, DC. ------. 2004i. "Uganda: Local Government Develop- ------. 1999b. "Philippines: The Challenge of Eco- ment Program." Implementation Completion Re- nomic Recovery." Sector Report No. 18895, World port No. 30284, World Bank, Washington, DC. Bank, Washington, DC. ------. 2004j. World Development Report: Making Ser- ------. 1998a. "Building Institutions and Financing Local vices Work for Poor People. Washington, DC: World Development: Lessons from Brazil and the Philip- Bank. pines. Municipal Development Project in the State of 88 BIBLIOGRAPHY Parana (Loan 3100-BR) Municipal Development Pro- ------. 1994b. "Philippines--Devolution and Health ject in the State of Rio Grande do Sul (Loan 31 29-BR) Services: Managing Risks and Opportunities." Sector Municipal Development Project in the Philippines Report No. 12343, World Bank, Washington, DC. (Loan 2435-PH) Second Municipal Development Pro- ------. 1992a. "Brazil: The Challenge of Municipal Sec- ject in the Philippines (Loan 3146-PHI)." Report No. tor Development in the 1990s. Volume l: Policy Sum- 18727, World Bank, Washington, DC. mary and Reform Proposals." LAI Department ------. 1998b. "Rethinking Decentralization in Devel- Infrastructure Operations Division Latin America oping Countries." Poverty Reduction and Economic and the Caribbean Regional Office, World Bank, Management Report, World Bank, Washington, DC. Washington, DC. ------. 1997a. "Health, Nutrition and Population." Sector ------. 1992b. "Public Sector Debt, Fiscal Deficits, and StrategyPaperNo.17092.Washington,DC,WorldBank. Economic Adjustment: A Comparative Study of Six ------. 1997b. World Development Report: The State EMENA Countries." Policy Research Working Paper in a Changing World. Washington, DC: World Bank. No. WPS840, World Bank, Washington, DC. ------. 1994a. "Fiscal Decentralization and Intergov- Wunsch, James S. 2005. "Decentralization, Local Gov- ernmental Finances in the Republic of Albania." Pol- ernance and the Democratic Transition in Southern icy Research Working Paper No. WPS1384, World Africa: A Comparative Analysis." African Studies Bank, Washington, DC. Quarterly. 89 THE WORLD BANK GROUP WORKING FOR A WORLD FREE OF POVERTY The World Bank Group consists of five institutions--the International Bank for Reconstruction and Development (IBRD), the International Finance Corporation (IFC), the International Development Association (IDA), the Multilateral Investment Guarantee Agency (MIGA), and the International Centre for the Settlement of Investment Disputes (ICSID). Its mission is to fight poverty for lasting results and to help people help themselves and their envi- ronment by providing resources, sharing knowledge, building capacity, and forging partnerships in the public and private sectors. THE INDEPENDENT EVALUATION GROUP ENHANCING DEVELOPMENT EFFECTIVENESS THROUGH EXCELLENCE AND INDEPENDENCE IN EVALUATION The Independent Evaluation Group (IEG) is an independent, three-part unit within the World Bank Group. IEG-World Bank is charged with evaluating the activities of the IBRD (The World Bank) and IDA, IEG-IFC focuses on assessment of IFC's work toward private sector development, and IEG-MIGA evaluates the contributions of MIGA guarantee projects and services. IEG reports directly to the Bank's Board of Directors through the Director-General, Evaluation. The goals of evaluation are to learn from experience, to provide an objective basis for assessing the results of the Bank Group's work, and to provide accountability in the achievement of its objectives. It also improves Bank Group work by identifying and disseminating the lessons learned from experience and by framing recommendations drawn from evaluation findings. ISBN 978-0-8213-7635-5 THEWORLD BANK SKU 17635