63438 DIREC TIONS IN DE VELOPMENT Public Sector Governance The Political Economy of Decentralization Reforms Implications for Aid Effectiveness Kent Eaton, Kai Kaiser, and Paul J. Smoke THE POLITICAL ECONOMY OF DECENTRALIZATION REFORMS THE POLITICAL ECONOMY OF DECENTRALIZATION REFORMS KENT EATON KAI KAISER PAUL J. SMOKE THE WORLD BANK Washington, D.C. © 2011 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 All rights reserved 1 2 3 4 14 13 12 11 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. ISBN: 978-0-8213-8840-2 eISBN: 978-0-8213-8841-9 DOI: 10.1596/978-0-8213-8840-2 Library of Congress Cataloging-in-Publication data have been requested. Cover photo: Highway Crossing, © Jürgen Zeller Cover design: Naylor Design Contents Acronyms viii Abstract ix Acknowledgments xi Applied Political Economy of Decentralization: An Overview xiii The Premise and Purpose xiii The Approach xiv Initial Context and Motivation xv Key Actors and the Centrality of Incentives xvii Shifting Reform Dynamics xvii Roles and Incentives of Development Partners xix Operational Implications xix 1 Introduction 1 The Centrality of Politics 1 Analytical Gaps and Challenges: A Pragmatic Approach 2 Beyond “Political Will” 3 A Political Economy Approach to Decentralization 4 Note 5 2 Outlining a Framework 7 The Basic Challenge 7 Initial Context and Motivation 8 Key Actors and the Centrality of Incentives 9 v vi Contents Reform Trajectories 9 Roles and Incentives of Development Partners 10 Notes 11 3 Understanding Political Incentives and Behavior 13 Electoral Incentives 13 Electoral Incentives and the Design of Decentralization 15 Partisan Incentives 16 Intra-party Dynamics 17 Inter-party Dynamics 18 Common Institutional Incentives to Defend the Center 19 Presidents, Legislators, and Judges in Democratic Regimes 21 Informal Institutions 24 Coalitional Incentives 25 Labor Unions and Business Associations 26 Ethnic Cleavages and Customary Authorities 28 Notes 29 4 Understanding Bureaucratic Incentives and Behavior 35 Key Actors and Basic Roles 37 Government Agency Interrelationships 39 The Role of External Development Partners 41 The Challenge of Coordination: Mechanisms and Realities 43 Informal or Ad Hoc Mechanisms 44 More Formal Mechanisms 45 Development Partner Coordination 47 Notes 48 5 Understanding the Dynamism of Context and Incentives 49 Unintended Consequences 50 Where You Sit Is Where You Stand 53 Pressures to Modify Initial Frameworks 56 Notes 59 Contents vii 6 Applied Political Economy of Decentralization Diagnostics 61 Expectations for Political Economy Assessments 63 Commissioning Country-Specific PED Analysis 65 Political Economy of Decentralization Country Assessment 66 Political Economy of Decentralization Issue Analysis 68 Review and Dissemination 69 From Analysis to Action 71 Looking Ahead 79 Notes 79 Annex 1: Guidance for Political Economy of Decentralization Analysis 81 References 85 Acronyms ANC African National Congress (South Africa) CDF Constituency Development Fund (Kenya) CPP Cambodia People’s Party DSF Decentralization Support Facility (Indonesia) MAS Movement Toward Socialism – Movimiento al Socialismo (Bolivia) MinMecs Intergovernmental Relations Committees of Ministers and Members of Provincial Councils (South Africa) MNR National Revolutionary Movement – Movimiento Nacionalista Revolucionario (Bolivia) MOI Ministry of Interior (Cambodia) PED Political Economy of Decentralization PEDCA Political Economy of Decentralization Country Assessment PEDIA Political Economy of Decentralization Issue Analysis ToR Terms of Reference viii Abstract This volume presents a preliminary framework designed to help in- ternational development partners consider the relevance of political economy issues for their programmatic support to decentralization and local government reform. The intention is neither to advocate decentralization in general or in any particular form, nor to presume or privilege any particular decentralization objective. Instead, the purpose is to document the potential value of better understand- ing how (primarily national and intergovernmental) political and institutional dynamics do or could affect the scope for realizing decentralization reforms aligned with commonly advocated service delivery, governance, and poverty reduction goals. The underlying premise is that systematic analysis of these issues can productively complement the dominantly technical diagnostic work typically carried out by development partners. Specifically, development partners can benefit from better understanding the practical signifi- cance of motives that drive politicians and bureaucrats to support or oppose reform at various stages of the decentralization process, from making an initial reform decision to detailed design and implementation. In addition, the framework addresses how these incentives can weaken, strengthen, or shift in response to changes in political and economic conditions that arise after reform begins. A general approach to conducting political economy of decentraliza- tion analysis is outlined, recognizing the need to tailor such analysis to the particular country context. This volume is based on literature reviews and knowledge derived from selected country experiences. ix Acknowledgments We gratefully acknowledge support for this work from the Bank- Netherlands Partnership Program, PREM-Governance Window (TF090075) and the leadership of Barbara Nunberg and Naazneen Barma in organizing and managing this activity. Earlier versions of this volume were presented at workshops hosted by New York University/Robert F. Wagner Graduate School of Public Service on October 1–2, 2009, by the World Bank Office in Bangkok on November 16, 2009, and the World Bank headquar- ters in Washington, D.C., on April 12, 2010. We are particularly thankful to Asmeen Khan and Debbie Wetzel for their support in organization of these dissemination events and for their role as moderators. The authors are grateful to the many academics, practitioners, and government counterparts who participated in these workshops and whose excellent comments and thoughtful suggestions have informed this revised version of the volume. We thank the follow- ing for their comments and support: Oscar Calvo-González, Tyler Dickovick, Victor A. Dumas Silva, Robert Ebel, Tulia Falleti, Verena Fritz, Benjamin Goldfrank, Henrik Fredborg Larsen, Scott Leiper, Brian Levy, Linda Che-Lan Li, Michael Malley, James Manor, Hieu Huu Nguyen, Kim N. B. Ninh, Kimchouen Pak, Austere A. Panadero, Alice Poole, Janelle Plummer, Erman Avantgarda Rahman, Jesse Ribot, Leonardo Romeo, Aladeen Shawa, Jonathan R. Stromseth, Joel A. Turkewitz, and Roland White. The principal discussants at the final workshop in Washington, D.C. were Jonathan Rodden, Judith Tendler, and Lily Tsai. Peer reviewers included Junaid Ahmad, xi xii The Political Economy of Decentralization Reforms in Developing Countries Jonas Frank, Keith McClean, Agnes Soucat, Lawrence Tang, Serdar Yilmaz, and Yongmei Zhou. Amanda Green and Lorena Viñuela provided substantial assistance in the preparation and review of this volume. Adelma Bowrin and Angkanee Luangpenthong supplied important administrative support. All views expressed in this volume are those of the authors, and not necessarily those of their respective organizations. Applied Political Economy of Decentralization: An Overview The Premise and Purpose Decentralization has been an important reform in many developing countries and a major focus of the considerable support provided by development partners to public sector reform. Yet its record is mixed at best in terms of realizing many of the stated objectives of reform, such as enabling more efficient service delivery, advancing democratic reform, and promoting economic development and poverty reduction. The lack of correspondence between official public policy goals (supported by development partners) and the (often-shifting) goals of political and bureaucratic actors commonly result in a failure to fully meet the stated objectives of decentralization and in a host of unintended consequences. The situation is complicated by the fact that central governments that pursue decentralization are not mono- lithic entities—the goals and behaviors of various actors within the central governments that pursue decentralization reforms, and their relative ability to shape policy may vary considerably. Development partners themselves can greatly influence reforms. They, too, face diverse incentives that may not correspond with those of govern- ment actors or of other donors, and may even reinforce conflicts among key actors and exacerbate inconsistencies in resulting de- centralization reforms. Substantial attention has been devoted to defining the insti- tutional and fiscal requirements for meeting commonly stated, normatively desirable decentralization objectives. But much less xiii xiv The Political Economy of Decentralization Reforms in Developing Countries consideration has been given to identifying the political and institutional incentives that drive reform or to evaluating how these reform drivers may shift over time. Understanding decentralization requires appreciating its fundamental paradox: what motivates the central government to give up powers and resources to subnational governments? The underlying premise of this volume is that systematic analysis of po- litical economy issues can productively complement the dominantly technical diagnostic work carried out by development partners seeking to undertake decentralization reforms. Development partners can benefit from a better understanding of the practical significance of motives that drive politicians and bureaucrats to support or oppose reform at various stages of the pro- cess, from initial reform decisions to detailed design and implementation. Since development partners are central to this process in many countries, greater self-reflection with regard to how they affect reform individually, collectively, and in relation to government actors would also be beneficial. Finally, it is important to understand how the incentives of key actors may weaken, strengthen, or shift in response to changes in conditions that arise after reform begins. We acknowledge that some development partners have made considerable efforts in recent years to conduct governance assessments that help pinpoint challenges and obstacles to the realization of normatively desirable fiscal and institutional reforms related to decentralization. Such assessments, however, rarely go into sufficient depth in evaluating: (a) whether actors whose coop- eration is needed to define and implement decentralization face incentives (and have the capacity) to support reform; (b) whether these often-diverse ac- tors—including other development partners—are likely to work in harmony or at cross-purposes; and (c) the implications of (a) and (b) for whether and how to support decentralization in a productive and sustainable way. It is important to emphasize that this effort does not intend either to advocate decentralization in general or in any particular form, or to presume or privilege any particular decentralization objective. The central purpose is to better understand how political and institutional dynamics (primarily those related to national actors, intergovernmental relations, and develop- ment partners) do or could affect the scope for realizing decentralization reforms aligned with the commonly advocated objectives of improved service delivery, governance, economic development, and poverty reduction. The Approach This volume outlines a framework for bringing political economy into the analysis that development partners undertake when formulating decentral- Applied Political Economy of Decentralization: An Overview xv ization support activities. The framework is general in nature, recognizing that the politics of decentralization is too diverse and complex a topic to analyze or approach uniformly across countries. Thus, rather than present- ing a standardized instrument that is expected to provide clear solutions to specific problems in any setting, this volume reviews and illustrates common political and behavioral patterns and dilemmas that lead to decentraliza- tion and that are in turn generated by decentralization. The framework is presented in Section 2. The overarching approach is simply to illustrate the types of questions that development partners may wish to ask about the politics of decentral- ization. The answers to these questions are expected to help development partners think about how to better design activities that can pragmatically support decentralization and lead to sustainable results. At least four major elements of the national and intergovernmental political economy of decentralization should generally be considered, de- pending on the particular country: The initial context and motivations for decentralization reform; The key actors involved in decentralization—politicians, bureaucrats, and citizens—and the incentives that condition their behavior with respect to reform design and implementation; The current stage of reform and its trajectory as it has unfolded since the initial decision to decentralize; and The role and incentives of key external development partners with re- spect to decentralization, both individually and collectively. Figure 1 provides a stylized overview of the main issues covered in this guidance. Traditional donor analysis of decentralization typically focuses to varying degrees on the political, administrative, and fiscal dimensions of decentralization. These are generally assessed with a strong emphasis on how they affect development objectives such as service delivery performance. While Figure 1 treats subnational government as a single sector, it is impor- tant to note that the interrelationships between intermediate (provincial) and local (district, county) levels have meaningful effects on the prevailing state and evolution of decentralization. Initial Context and Motivation Decentralization reforms evolve in many different contexts. In some cases, as in the Philippines or Indonesia, they respond to urgent political and/ or economic crises that have led to a dramatic turnover in leadership and xvi The Political Economy of Decentralization Reforms in Developing Countries F I G U R E 1. Drivers of Intergovernmental Reform a perceived demand for meaningful change. In other cases, decentraliza- tion has been framed as a means to strengthen the legitimacy of the state throughout the national territory, as in Bolivia and Colombia, and/or in the context of challenging post-conflict situations, such as Cambodia, Rwanda, and Uganda. A range of countries, including Brazil, South Africa, and Mexico adopted decentralization as part of the transition from authoritarian to democratic rule. In still other cases, decentralization has been part—either through explicit policy or tolerance of subnational behavioral discretion— of a broader market transition or economic development strategy, as in Vietnam and China. Understanding the initial conditions under which decentralization arose is a useful starting point for assessing the most genuine and robust reasons for pursuing it, their likely implications for the shape and pace of reform, and the potential durability of resulting policies. For example, a country in crisis may decentralize too much too quickly and run into serious problems later. A country under no obviously strong pressure to decentralize, on the other hand, may become stalled at early stages of reform. The nature of the system being decentralized is another important piece of the puzzle. Countries with a long history of authoritarian rule may face special challenges in promoting robust decentralization. Countries emerging from military rule, but that have some earlier history of more democratic systems (especially in Latin America), seem to be able to make the transi- Applied Political Economy of Decentralization: An Overview xvii tion more successfully. Institutional legacies also shape the environment in which decentralization unfolds, including a country’s federal or unitary design and the parliamentary or presidential form of its government. Cultural and colonial traditions of authority and administration may condition how reforms proceed, although this is not always well recognized in the design of normatively inspired reforms. Key Actors and the Centrality of Incentives Whatever the starting context, those most responsible for decentralization— elected politicians and national-level bureaucrats—face diverse incentives to pursue, to appear to pursue, or to limit reform. For analytical purposes, we distinguish in this volume between the motivations facing elected politi- cians—electoral, partisan, institutional, and coalitional—and those facing appointed bureaucrats—consolidating institutional power, improving career trajectories, and checking rival agencies. Both sets of motivations heavily affect the shape of decentralization and may in turn be affected by it; these may or may not be closely related. Such incentives are far from uniform across major players and typically underlie important debates and conflicts about the nature, extent, and pace of decentralization. At the political level, elected officials, political parties, subnational gov- ernments, and key civil society players such as unions and business associa- tions act individually and/or in concert in ways that define how decentraliza- tion is conceived and implemented. The relative strength of various actors and the degree to which some of them have common interests ultimately determine outcomes, particularly with respect to the initial decision to un- dertake decentralization. At the organizational level, multiple government agencies typically become involved in decentralization with varying levels of interest and influence and often-diverse incentives, especially during the detailed design and implementation phases. Sections 3 and 4 review these political and bureaucratic incentives, respectively, in more detail, providing examples of their relative significance and how they have shaped decentral- ization in a number of countries. Shifting Reform Dynamics Decentralization is a process, not a one-time act, and the trajectory of reform is heavily influenced by how the often-conflicting incentives of different actors are pursued. At the initial design stage, the decision to decentralize is reflected in de jure changes in legal frameworks, including executive decrees, new statutes, and in some cases even constitutional reforms. Even if the legal xviii The Political Economy of Decentralization Reforms in Developing Countries framework defines a strong form of decentralization, however, its provisions may be subverted for political reasons in the more detailed design decisions that follow the adoption of an overarching framework. In the implementation phase, the transfer of resources and responsibili- ties often sets in motion intensely political struggles in the state bureaucracy, both between and within agencies and across levels of government. These struggles can affect how far and how well implementation proceeds. Although not the focus of this volume, once the new division of authority has been detailed and implementation has begun, political dynamics within disparate subnational jurisdictions may powerfully shape the use of new governing prerogatives on the ground. Different dimensions of decentralization can vary in importance at different stages. Administrative, fiscal, and political decentralization can be rolled out in different sequences on the basis of politically motivated strategies. Although strategic sequencing does not occur in all countries, it is important to be aware of where it is occurring, and to understand underlying political motivations and their implications for possible reforms and outcomes. Similarly, some decentralization reforms spe- cifically or differentially target individual public service sectors (such as health, education, transport, and water) and in ways that relate primarily to administrative or fiscal dimensions. This can occur due to legitimate conceptual, technical, and pragmatic concerns about appropriate subna- tional government functions and related levels of autonomy, but it may also reflect powerful political and institutional dynamics that enhance or constrain reform possibilities. Changing political and economic conditions, as well as emerging empirical evidence on the performance and unintended consequences of decentralization policies, can alter incentives and the overall reform trajec- tory. For example, the alleviation of immediate crises or the emergence of new ones can change attitudes toward decentralization and lead to modifi- cations in its form and function. Decentralization does not always yield its expected benefits quickly, and impatience with this time lag can lead to calls for policy adjustments. The often-diverging political incentives facing national and subnational politicians and bureaucrats influence—sometimes in very different ways— how these actors respond to evidence that reform is not working as planned. Whereas subnational actors might interpret post-decentralization problems as evidence of the need to provide them with additional resources, national actors may view these same problems as an argument for recentralization. Section 5 considers how the various incentives and behaviors of key actors may influence the evolution of decentralization. Applied Political Economy of Decentralization: An Overview xix Roles and Incentives of Development Partners The international development community has played a major role in pro- moting and supporting decentralization in developing countries, especially in those that are heavily dependent on foreign assistance. Yet development partners themselves face varying incentives, which can at times lead them to act in ways that are not entirely consistent with country policies, with other donors’ programs, or even with their own rhetoric. It is neither a secret nor a surprise that individual agencies often have a political need to distinguish the specific roles they play, and while donor competition is not necessar- ily or always undesirable, fragmented donor behavior can hinder even the best-intentioned reforms. Some development partners, including the World Bank, are large and complex institutions, with multiple departments that focus on different as- pects of development, service delivery, and institutional reform, all of which affect decentralization. These departments may have different objectives and incentives that further complicate the role of donors in decentraliza- tion reform. Section 4 discusses the role of development partners and their interactions with country actors involved in decentralization. Operational Implications The central premise of this volume is that the quality and sustainability of decentralization support provided by development partners could universally benefit from more thorough consideration of the political and institutional context and dynamics underlying reform in a given country and the endog- enous role that development partners themselves often play. This analysis is intended to complement rather than to replace more conventional technical analysis. By broadening conventional analysis to incorporate these issues, much can be learned about the preferences and likely behaviors of different political and bureaucratic actors. A greater understanding of these factors can in turn help development partners to better appreciate and respond to the historical factors behind decentralization as well as the current drivers of reform, in hopes of supporting the most pragmatic and sustainable approaches. There are a few caveats to our recommendation for more robust political economy assessment. First, our analysis is not based on a systematic examina- tion of the full range of factors and stakeholders. We did not conduct origi- nal research for this effort—we draw on a review of relevant literature and personal or available secondary knowledge of selective country experiences. In some countries, these sources told us a great deal about certain aspects of xx The Political Economy of Decentralization Reforms in Developing Countries the simple framework we present, but not very much about others. It would therefore be useful to conduct a more systematic analysis of a range of care- fully selected countries and to use the information derived from this exercise to develop a more fully articulated and illustrated framework identifying the political and institutional dynamics that motivate decentralization. Second, pursuing this type of analysis will almost certainly force devel- opment partners to consider making more difficult decisions and to con- front challenges that would complicate how they approach reform, in some cases perhaps even going against what they see as their own organizational, departmental, or individual objectives. Is decentralization support even worthwhile in a country where the dynamics underlying reforms are not consistent with the attainment of development partner objectives? Should a donor partner think about working with a different government agency than it originally intended to work with or that it has worked with historically on subnational government matters? Are there ways that development partners could help promote better coordination—among other development part- ners, among fractured government agencies, between development partners and government agencies, or within their own organization—even if at the expense of individual glory? What might create the incentives, support, and/ or momentum for such an approach? Third, we have not focused on local-level political and bureaucratic dynamics beyond how certain local actors interact with the national level in shaping the overall intergovernmental system, but we must recognize that they can be extremely important. Understanding the positions, at- titudes, and opportunities of local-level actors may provide important insights, including in cases where development partners might be more influential by focusing on subnational or nongovernmental interventions when the scope for meaningfully influencing decentralization reforms de- fined by the central government is limited. A more developed framework for helping development partners to better diagnose the subnational and nongovernmental situation and opportunities would be a useful addition to their analytical tool kit. Finally, the approach we outline to conducting more robust political economy analysis is only illustrative. In some cases, there will be little analysis of the relevant issues available, and new work will have to be commissioned. In other cases, existing work can be drawn on, so that additional work will be limited. If the political economy situation is relatively well understood by development partners, the main task will be to interpret its specific relevance for operational purposes rather than to conduct new in-depth analysis. Development partners preparing new decentralization support programs will need to use their judgment, with appropriate support when required, to Applied Political Economy of Decentralization: An Overview xxi determine the extent and depth of additional analysis that would be useful in a particular case. Some development partners may feel comfortable maintaining the status quo in how they approach decentralization support and with whom they work, and they may rightly see the approach advocated in this volume to be challenging and potentially onerous. The approach does, however, build on governance assessments and other analytical work that development partners themselves have already seen the need to undertake. Given the often-underwhelming performance of decentralization reform in developing countries—including the disappointing performance of donor support for decentralization relative to the substantial volume of resources devoted to it in the past couple of decades—broader-based and more careful analysis of whether and how to engage on decentralization reform in a particular country could be well worth the effort for development partners and the countries they support. Introduction The Centrality of Politics No matter what the official justification, decentralization is largely driven and continually shaped by politics and institutional dynam- ics. Politicians and public officials tend to cite lofty, normatively inspired, internationally advocated goals when they decide to de- centralize, including the promotion of democracy, development, and public security, and/or improved efficiency and equity in the delivery of government services. Yet other more immediate, nar- rowly political factors are often more centrally behind the decision to decentralize and the process of bringing decentralization to life. It is difficult, in fact, to imagine a more intensely political process than decentralization. Roughly since the emergence of the modern nation-state, struggles between and within national and subnational government entities over which levels do what, and with whose revenues, have been at the core of state formation. Decentralization comes in many shapes and sizes, but in every instance involves changing the institutional rules that divide re- sources and responsibilities among levels of government. Politicians and bureaucrats fight over decentralization for the same reason that they fight over the design of state institutions more generally: their power and authority are at stake. This reality holds regardless of: (a) the nature of decentralization (whether the full devolution of responsibilities to the local level or a more limited form of sub- national empowerment); (b) the substantive target(s) of reform (administrative, political, fiscal); (c) the generality or narrowness of reform (whether broad-based or activity/sector-specific); and (d) the type of transformation involved (creating new subnational 1 2 The Political Economy of Decentralization Reforms in Developing Countries administrations or governments where none existed, transforming existing local administrative units into elected governments, or strengthening elected local governments that have powers but function poorly). All of these facets of decentralization involve deeply political acts and processes, from the initial decision to decentralize through all subsequent stages of reform. Analytical Gaps and Challenges: A Pragmatic Approach Although the factors that drive and shape decentralization are undeniably political, development partners still tend to treat decentralization as a primar- ily technical exercise based on universally desirable norms. While donor staff might accept in principle the point that “politics matter,” in practice their interventions in the area of decentralization rarely reflect a well-developed sense of the political dynamics that set reforms in motion and in which they are embedded. In most cases, internal politics can, or perhaps should, lie beyond the reach of external development partners’ influence, but this does not mean that they should not seek to better understand these dynamics and their implications. This volume is based on the premise that a more thorough knowledge of the politics of decentralization can help development partners make better decisions, resulting in more appropriate and durable outcomes for country reforms and donor programs.1 While development partners who are interested in genuine and sustain- able reform cannot afford to ignore the politics that pervade decentralization, to date they have had little readily available information and few analytical tools to help them figure out exactly how “politics matter” or how politics might affect their projects. This stands in contrast to the availability of advice on the technical side of decentralization, including on the design of intergov- ernmental administrative and fiscal systems and service delivery models. In truth, the political side of decentralization is more difficult to systematically examine in a policy-relevant way, and it does not lend itself as readily to the identification of “best practices.” The fact that similar practices lead to such divergent outcomes in different settings speaks powerfully to the need to take politics more seriously from the beginning. The politics of decentralization is too diverse and complex a topic to analyze or approach uniformly across countries. Rather than presenting a standardized instrument that could be applied in any setting and be expected to identify clear solutions to specific problems, this volume reviews com- mon political patterns and dilemmas that bring about decentralization and that are in turn generated by decentralization. Our goal is to sketch a basic diagnostic framework that identifies the types of questions development partners may wish to ask about the politics of decentralization. Our hope is Introduction 3 that answers to these questions will help development partners think about how to better design activities that can pragmatically support decentral- ization and lead to sustainable results. In this effort, we draw heavily on a significant body of new scholarship on decentralization—political science research developed over the last decade that complements the older public administration and finance literature on decentralization, but focuses more on political and institutional dynamics. We also draw on material from a number of country cases. It is important to emphasize that we are not advocating that develop- ment partners support decentralization in general or as a means to particular ends. Our position is more agnostic. If donors are considering whether to introduce programming in support of decentralization, we propose that they consider paying closer attention to the incentive structures we describe in this volume. Our approach is similarly neutral about the form that decen- tralization takes. We also note that our focus here is on national and broad intergovernmental dynamics. We acknowledge the critical importance of local political and bureaucratic dynamics in how decentralization unfolds on the ground, and we hope that future work can explore these phenomena and link them back to higher-level dynamics. Beyond “Political Will” While we do not propose that development partner staff need to master all aspects of a country’s political system, we do maintain that they should go substantially beyond gauging the overused, misunderstood, and vague indicator of “political will.” In the past, attempts by development partners to factor politics into the analysis of decentralization tended to reflect an overly voluntarist set of assumptions. According to this thinking, decentralization would be defined in normatively state-of-the-art legal frameworks. It would then be implemented according to these frameworks, but only if national politicians and bureaucrats were sufficiently enlightened and committed to decentralization. A more positivist approach yields the potentially bad news (for advocates of decentralization) that national-level “political will” is often insufficient for decentralization and that national governments are rarely monolithic, uni- fied entities. A host of obstacles, including electoral challenges, bureaucratic opposition, and institutional and policy legacies, can hinder the efforts of national government advocates who are committed to adopting, designing, and implementing meaningful decentralization. At the same time, this less voluntarist approach provides the potentially good news that “political will” as traditionally defined—enlightened, united, 4 The Political Economy of Decentralization Reforms in Developing Countries and strongly proactive political leadership—may not always be necessary to realize decentralizing reforms. Development partner support for decentral- ization reform has tended to focus on a narrow range of central government actors, overlooking opportunities to engage in national- and subnational- level interventions that might be productive even in the absence of broad- based political commitment to decentralization. Rather than emphasizing the degree of political will to decentralize, we believe the analysis of political and bureaucratic incentives offers a more useful diagnostic entry point for development partners wishing to support decentralization. If politicians and bureaucrats support decentralization when they face incentives to do so, we need to assess not only the depth of their normative commitment to decentralization, but also how it is likely to affect their career paths, electoral ambitions, support coalitions, and secu- rity of tenure. As decentralization unfolds, the incentives of key actors and institutions change, creating new pressures and positions. Political incentives can shift rapidly, suddenly opening up critical opportunities to advance decentralization and just as quickly closing the reform window. As a result, development partner support for decentralization needs to operate with a high degree of flexibility and pragmatism. Even as we focus in this volume on the political and bureaucratic incen- tives that open up “windows of opportunity” for development partners who wish to support decentralization, we acknowledge the powerful structural factors that are beyond the control of individual bureaucrats and politicians but profoundly shape decentralization. For example, central governments are in a structurally advantageous position for collecting public resources relative to subnational governments; hence the need for transfers. Subna- tional regions within the same country can face stark territorial imbalances in levels of development and natural resource endowments. Conflicts over decentralization and recentralization cannot be separated from histori- cal grievances between regions seen to be winners and losers in the long struggle for economic development. Major economic shocks or exogenous price changes can affect a country’s ability to pursue decentralization policy effectively, at least for a time. Finally, structural changes like rapid urbaniza- tion can strongly condition debates over decentralization, helping to explain why it is occurring in so many otherwise disparate settings. A Political Economy Approach to Decentralization Section 2 of this volume sets out the elements of a framework for consider- ing how national political economy factors may shape intergovernmental systems and reform trajectories. Section 3 reviews the types of political ac- Introduction 5 tors typically involved in decentralization and the incentive structures that guide them. Section 4 reviews the bureaucratic landscape of decentraliza- tion, highlighting key actors (including development partners) and their relationships. Section 5 illustrates how these actors and motives may affect the often-lengthy process of decentralization, from its adoption to its design and implementation, and even to the modifications that are undertaken as political and economic conditions change over time. Section 6 offers some guidance for engaging in pragmatic country-level political economy analysis and applying it for operational engagement. Note 1. Governance and political economy analysis has a crucial part to play in informing agencies such as the World Bank, and can be applied to a range of challenges with the aim of enhancing development effectiveness (Fritz, Kaiser, and Levy 2009). Broadly defined, political economy analysis is concerned with the interaction of political and economic processes in a society: the distribution of power and wealth among dif- ferent groups and individuals, and the processes that create, sustain, and transform these relationships over time. 2 Outlining a Framework The Basic Challenge A significant portion of mainstream decentralization diagnostics and advisory work conducted by international development partners focuses on the nature and performance of a country’s subnational government or intergovernmental system. This analysis is often con- ducted from the perspective of performance in organizational and procedural design, service delivery, and development outcomes. Questions typically include whether subnational governments are assigned appropriate functions, whether they have sufficient resources to carry out these functions, and the degree to which the incentives and accountability for service delivery are adequate. To a great extent, this diagnosis is either explicitly or implicitly shaped by normative principles and considerations of “best” or good international practice. Although there have been meaningful and productive decen- tralization reforms and development partner support programs in some countries, it is widely accepted that many perform well below expectations.1 There are many reasons for this, but one factor is the limited attention paid to the true (as opposed to official) motives for decentralization and the incentives for key actors to behave in cer- tain ways as decentralization unfolds. As noted above, some factors are beyond development partner control, but knowing more about them provides valuable information about potentially productive types of support and how these might be strategically identified, designed, and implemented. At least four major issues regarding the national and inter- governmental political economy of decentralization should be considered, depending on the country context: 7 8 The Political Economy of Decentralization Reforms in Developing Countries The initial context and motivations for decentralization reform; The key actors involved in decentralization—politicians, bureaucrats, and citizens—and the incentives that condition their behavior with respect to reform design and implementation; The current stage of reform and its trajectory as it has unfolded since the initial decision to decentralize; and The role and incentives of key external development partners with re- spect to decentralization, both individually and collectively. Initial Context and Motivation Countries decentralize or reform already decentralized systems in many different contexts. In some cases, as in the Philippines and Indonesia, they are responding to urgent political and/or economic crises that have brought about dramatic leadership changes and a perceived demand for drastic reform. In other cases, decentralization has been framed in terms of bol- stering the legitimacy of the state and its presence throughout the national territory, as in Bolivia and Colombia, and/or in challenging post-conflict situations such as Cambodia, Rwanda, and Uganda. A range of countries, including Brazil, South Africa, and Mexico, adopted decentralization as part of the transition from authoritarian to democratic rule. In still other cases, decentralization has been part—through either explicit policy or tolerance of subnational behavioral discretion—of a broader market transition or economic development strategy, as in Vietnam and China. Understanding the initial conditions under which decentralization arose in a given country is a useful starting point for assessing the most genuine and robust reasons for pursuing it, their likely implications for the shape and pace of reform, and the likely durability of resulting policies.2 For example, a country in crisis may decentralize too much too quickly and run into serious problems later. A country under no obviously strong pressure to decentralize, on the other hand, may become stalled at early stages of reform, frustrating those who expected more. The nature of the system being decentralized is another important piece of the puzzle. Countries with a long history of authoritarian rule may face special challenges in promoting robust decentralization. Countries emerging from military rule, but that have some earlier history of more democratic systems (especially in Latin America), seem to be able to make the transition more successfully. Institutional legacies also strongly shape the environment in which decentralization unfolds, including a country’s federal or unitary design and the parliamentary or presidential form of its government. Cultural and colonial traditions of authority and administration may also condition Outlining a Framework 9 how reforms proceed, although this is not always well recognized in norma- tively inspired reform designs. Key Actors and the Centrality of Incentives Whatever the starting context, those most responsible for decentralization— elected politicians and national-level bureaucrats—face diverse incentives to pursue, to appear to pursue, or to limit reform. For analytical purposes, we distinguish in this volume between the motivations facing elected politicians (electoral, partisan, institutional, and coalitional) and the motivations facing appointed bureaucrats (consolidating institutional power and improving ca- reer trajectories, among others). Both sets of incentives are far from uniform across major players and typically underlie important debates and conflicts about the nature, degree, and pace of decentralization. At the political level, elected officials, political parties, subnational governments, and key civil society players (such as unions and business associations) act individually and/or in concert in ways that define how decentralization is conceived and implemented. The relative strength of various actors, and the degree to which some of them have common inter- ests, ultimately determine how decentralization evolves—particularly with respect to the initial decision to undertake decentralizing reforms. At the organizational level, multiple government agencies typically become involved with varying levels of interest and influence and often-diverse incentives, usually during the detailed design and implementation phases rather than at the adoption stage. Reform Trajectories Decentralization is a process, not a one-time act, and the trajectory of reform is heavily influenced by the degree to which the often-conflicting incentives of different actors play out. At the initial design stage, the decision to decen- tralize is reflected in de jure changes in legal frameworks, including executive decrees, new statutes, and in some cases even constitutional reforms. Even if the legal framework defines a strong form of decentralization, however, its provisions may be subverted for political reasons in the more detailed design decisions that follow the adoption of an overarching framework. In the implementation phase, the actual transfer of resources and responsi- bilities often sets in motion intensely political struggles in the public sector bureaucracy, both among and within agencies. These struggles can affect the extent and nature of implementation. Although not the focus of this volume, once the new division of authority has been detailed and implementation 10 The Political Economy of Decentralization Reforms in Developing Countries begun, political dynamics within subnational jurisdictions may powerfully shape the use of new governing prerogatives on the ground. It is important to recognize that different dimensions of decentraliza- tion can vary in importance at different stages. Falleti (2005; 2010) examines how political, administrative, and fiscal decentralization can be rolled out in different sequences on the basis of politically motivated strategies. Although strategic sequencing does not occur in all countries, it is important to be aware of where it is occurring and to understand the underlying political motiva- tions and their implications for possible reforms and outcomes. Similarly, some decentralization processes specifically or differentially target individual public service sectors (such as health, education, transport, and water), and in ways that relate primarily to administrative or fiscal dimensions. This is due sometimes to legitimate conceptual, technical, and pragmatic concerns about appropriate subnational government functions and related levels of autonomy, but it may also reflect powerful political and institutional dynam- ics that enhance or constrain the possibilities for reform. As discussed above, changing political and economic conditions, as well as emerging empirical evidence on the performance and unintended consequences of policies, can alter reform incentives and trajectories. The alleviation of an immediate crisis or the emergence of a new one, for example, can change attitudes toward decentralization and lead to modifications. Not uncommonly, decentralization is slow to bring about its officially expected benefits, and impatience with this can lead to calls for policy adjustments. It is important to note that the often-diverging political incentives facing national and subnational politicians and bureaucrats influence—sometimes in very different ways—how these actors respond to evidence that reform is not working as planned (Connerley, Eaton, and Smoke 2010). Thus, whereas subnational actors might interpret post-decentralization problems as evidence of the need to provide them with additional resources, national actors may view these same problems as an argument for recentralization. Roles and Incentives of Development Partners The international development community has played a major role in pro- moting and supporting decentralization in developing countries, especially those that are heavily dependent on foreign assistance. Development partners do not form a monolithic community, however, and they do not always work in harmony or act in ways that are entirely consistent with country policies or even their own rhetoric.3 It is neither a secret nor a surprise that individual agencies often have a political need to distinguish the roles they are playing from each other, and they may be obligated or encouraged to Outlining a Framework 11 follow their own institutional policies even when the resulting programs do not fully support official government decentralization policies or are at odds with the established programs of other development partners. While donor competition is not always undesirable, there is ample evidence that fragmented donor behavior can wreak havoc in the area of decentralization.4 An important overarching question for the international development community is whether decentralization (and, more specifically, a particular form of it) is inherently desirable. A number of development partners do advocate particular forms of decentralization; for example, democratic local governance is a favored objective of several major bilateral agencies. Other development organizations, including the World Bank, tend to emphasize system development, service delivery, and poverty reduction rather than more explicitly political objectives. Whatever goals they may be pursuing in a broad sense, some development partners, including the World Bank, are large and complex institutions with multiple departments that focus on different aspects of development that affect decentralization. The varying objectives and incentives of these departments can further complicate the role of donors in decentralization reform. Notes 1. This is the position taken by a wide variety of literature published over the last two decades, including: Bahl and Linn (1992); Prud’homme (1995); Ter-Minassian (1997); Bird and Vaillancourt (1998); Litvack, Ahmad, and Bird (1998); Smoke (2001); Tanzi (2001); Ahmad and Tanzi (2002); Ebel and Yilmaz (2002); Shah (2004); World Bank (2005); Bardhan and Mookherjee (2006); Cheema and Rondinelli (2007); World Bank (2008); and Connerly, Eaton, and Smoke (2010). 2. See Smoke, Gomez, and Peterson (2007) for a literature review and discussion about initial motivations and reform trajectories (covered below). 3. Donor behavior and coordination on decentralization are discussed in Romeo (2003) and Smoke (2007; 2010). 4. See, for example, the review of the World Bank’s decentralization support experi- ences (World Bank 2008). 3 Understanding Political Incentives and Behavior M uch of the recent political science literature recognizes that decentralization poses a fundamental puzzle. On the one hand, any decentralization measure worthy of the name tends to reduce the power and authority that national politicians enjoy rela- tive to subnational actors. On the other hand, national politicians formally control the decision to decentralize. What incentives do national politicians have to endorse changes that appear, at least at first glance, to diminish their political prerogatives? While political incentives take multiple forms, here we focus on the four most im- portant types that politicians face: electoral, partisan, institutional, and coalitional.1 By surveying a broad range of incentives, the discussion pre- sented in this section aspires to be broadly relevant to development partners working in a range of countries, including long-standing democracies, stable authoritarian systems, and countries on the continuum between these extremes. Different incentive structures are more or less salient in a particular country depending on regime type. For example, electoral incentives are not likely to be as im- portant in authoritarian regimes, and the nature of such incentives will vary in other cases with the degree of electoral competition. Electoral Incentives In the growing set of countries where relatively free and fair elections determine who gains access to decision-making positions within 13 14 The Political Economy of Decentralization Reforms in Developing Countries the state, development partners would benefit from paying attention to the electoral dynamics of decentralization. Two dynamics have received special attention in the literature: the increasingly common selection of subnational officials via election rather than appointment, and the strategic calculations that ambitious politicians make about the likelihood of winning national versus subnational elections. The advent of subnational elections marks one of the most significant new features of the third wave of democracy that began in the 1970s (Hun- tington 1991).2 Regularly electing subnational officials (and eliminating the prerogative of the national government to appoint them) opens up the possibility that national and subnational offices will be won by candidates representing different political parties (or different political groupings in countries where parties are insignificant). In these situations, subnational officials may be able to use their separate democratic legitimacy to push for greater resources and responsibilities from below.3 In Mexico, the declining legitimacy of the governing party in the 1980s led the national government to accept the state-level electoral victories of a right-of-center opposition party, which then used its new governing au- thority to demand additional decentralization (Beer 2004). Brazil provides a similar example, though on the left. Subnational elections created an opening in the 1980s and 1990s for the Brazilian Worker’s Party, whose innovative approaches in such municipalities as Porto Alegre led to repeated electoral victories, thus offering a powerful example of the merits of decentraliza- tion in a period marked by national government attempts at recentraliza- tion (Samuels 2004). In these and other cases, the emergence of politically noncompliant local officials has made it harder for national governments to monopolize the design of intergovernmental systems. Subnational elections can also have the opposite impact. National officials can use the perquisites of incumbency to support their favored candidates in subnational races, thus widening and deepening the networks that sustain governing and/or hegemonic parties. In Cambodia, subnational elections began at a very local level—the mostly rural communes—where Prime Minister Hun Sen and his ruling Cambodia People’s Party (CPP) were very popular. After the coup d’état that he led in 1997 and the national elections that he narrowly won in 1998, widespread CPP victories in 2002 commune elections further consolidated his popularity in rural areas and helped reinforce his national hold on power (Smoke and Morrison 2009). In Indonesia after the fall of Suharto, the push for subnational elections created an arena through which President Habibi sought to bolster the fortunes of the Golkar party and his own success in the next election—unsuccessfully, as it turned out (Hofman and Kaiser 2004). Understanding Political Incentives and Behavior 15 The relative timing of subnational and national elections may have an additional impact on political struggles over decentralization. Simultaneous elections tend to nationalize electoral contests and diminish the likelihood that local races will be decided on the basis of how well local officials have performed during their period in office (Carey and Shugart 1992). Noncon- current elections, on the other hand, can help to keep subnational elections focused on subnational outcomes, which most of the literature sees as critical for decentralization to improve downward accountability to voters.4 Electoral Incentives and the Design of Decentralization Political scientists have used theories of progressive ambition to study the nexus between electoral incentives and decentralization. According to an influential argument by O’Neill (2003; 2005), ambitious politicians will view decentralization through the lens of its likely impact on their own electoral futures. When national parties calculate that their electoral fortunes are more promising at the subnational level, and consequently that holding subnational offices will be important to them in the future, they are more likely to sup- port decentralization to make these offices more powerful. Thus, although national politicians generally share an interest in defending the prerogatives of the national government, the dynamics of electoral competition can open up divisions among national politicians vis-à-vis decentralization. In Bolivia, the combination of a fragmented party system and consti- tutional rules allowing Congress to pick the president from among the top three vote winners led the National Revolutionary Movement (MNR) to believe that it would be unlikely to hold on to national power for long after its 1993 electoral victory. Because the MNR had a more developed organi- zational apparatus and more stable support bases than other parties at the subnational level, MNR leaders endorsed a bold approach to decentraliza- tion in 1994. In Colombia, a temporary split within the otherwise dominant Liberal Party enabled the Conservative Party to defy expectations and win the presidential election of 1982, after which party leaders used their control of the national government to implement a series of decentralizing changes before leaving power in 1986. Even before elections are held, politicians’ attempts to anticipate the outcome of electoral contests can encourage them to decentralize. These calculations are especially critical in the context of negotiated transitions from authoritarian to democratic rule, which are often characterized by high levels of uncertainty about likely winners and losers.5 In South Africa, the transition to a multiparty democracy in the 1990s set the stage for significant negotiations over decentralization. Anticipating a national electoral defeat at 16 The Political Economy of Decentralization Reforms in Developing Countries the hands of the African National Congress (ANC), the National Party and the Inkatha Freedom Party insisted on strengthening provincial governments.6 Similarly, partisan and electoral incentives overlapped in the struggle over decentralization in post-Pinochet Chile. Whereas parties on the left sought to reintroduce municipal elections, which they correctly calculated they would win overwhelmingly, the right preferred to hold elections in the new administrative regions that Pinochet had introduced.7 In nondemocratic regimes, the absence of electoral pressures can be an important factor in understanding how politicians respond to decentraliza- tion. Consider the military regimes that replaced elected civilian governments in Argentina, Brazil, and Chile in the 1960s and 1970s. In these cases, after lengthy periods during which competitive elections had been held regularly for subnational offices, military-led governments opted either to terminate (Argentina and Chile) or manipulate (Brazil) subnational elections. Once they had eliminated electoral threats from subnational officials, military governments adopted policies that substantially expanded the authority of subnational governments, transferring schools and hospitals and in some cases additional fiscal resources. Authoritarianism thus can generate signifi- cant subnational reforms, even if these fall short of political decentralization.8 This new research on electoral incentives has several implications for those seeking to advance decentralizing reforms. First, it suggests that it is not enough to monitor the outcomes of national and subnational elections to determine whether politicians who are normatively committed to decen- tralization have prevailed. Electoral and political cycles can be significant for decentralization even when it is not a salient electoral or political issue. Second, when electoral returns open up potential space for decentraliza- tion, development partners face a number of options. On the one hand, they could decide to act quickly to support decentralization, as changes introduced while a reform window is open can sometimes survive adverse political circumstances after the window closes. On the other hand, develop- ment partners may be more cautious about supporting decentralization in response to limited and/or brief reform windows, which could lock in partial reforms and thus impede better service delivery and poverty reduction in the years ahead. Attempts to navigate this political space do not come with easy answers, but a better reading of political economy drivers can help assess the potential evolution and sustainability of decentralization developments. Partisan Incentives In countries where parties operate as the main gatekeepers to political office, struggles over decentralization can reflect internal party structure as well as Understanding Political Incentives and Behavior 17 the logic of inter-party competition.9 Politicians will tend to view decentral- ization through the lens of the partisan incentives that govern their careers. While the party system is especially relevant in democratic contexts, partisan incentives and party-mediated relationships may matter even in countries with single-party dominant systems that do not qualify as democracies. Intra-party Dynamics Particularly critical in understanding support for decentralization is whether national or subnational politicians dominate the parties to which they be- long.10 In a pioneering study, Willis, Garman, and Haggard (1999) explained diverging levels of decentralization in five Latin American countries as a function of internal party structures. They found that legislators who are be- holden to subnational officials within their parties are more likely to support decentralization than legislators who are responsive to national party leaders. For example, if the careers of national legislators are substantially controlled by governors in their home states or provinces, then these governors are in a good position to demand the passage of decentralizing legislation “from below.”11 If, by contrast, it is national party leaders who control the careers of national legislators, then legislators are likely to face fewer pressures to decentralize and/or enjoy greater scope to design decentralizing legislation that reflects the preferences of national actors. Scholars have built on this approach more recently by emphasizing the importance of determining whether the lines of authority within parties flow upward to national party leaders or downward to subnational party leaders.12 The party system matters for decentralization in single-party systems, too. Here, the centralization of power within governing parties typically meant that decentralization conforms to the interests of the national leader- ship, leaving subnational party leaders little room to influence the design of decentralization. In Cambodia, Prime Minister Hun Sen’s tight control over the governing CPP has enabled him to decentralize without increasing the likelihood that local councils would challenge his dominance.13 In South Africa, where inter-party struggles dominated the design of decentralization in the mid-1990s, the shift to a single-party dominant system means that the logic of decentralization and recentralization can now be best understood through the prism of intra-African National Congress struggles between national and provincial party leaders (Dickovick 2007; 2010). In Uganda, President Yoweri Museveni pushed decentralization from the top down to build a strong base of public support for his National Resistance Movement, which was framed as a “nonparty” governance arrangement but functioned liked a political party. His use of village-level National Resistance Councils helped bring him to 18 The Political Economy of Decentralization Reforms in Developing Countries power, and the success of this local approach triggered the creation of a local government system that was highly empowered, at least initially.14 Decentralization between levels of government can also explain decen- tralization within parties.15 Chhibber and Kollman (2004) argue that it is the fiscal relationship between national and subnational governments that determines the attributes of the party system, and not vice versa. According to their comparative analysis of Canada, Great Britain, India, and the United States, centralized fiscal practices generate the formation of national party systems and fiscal decentralization generates regionalized parties. Inter-party Dynamics What happens between parties in their struggles to benefit and/or avoid losses from decentralization is also important in how decentralization unfolds. Political scientists compare party systems along three main dimensions: fragmentation (the number of parties), polarization (the degree of ideologi- cal distance between them), and institutionalization (the degree to which parties exist as organizations separate from their leaders and embedded in society). Each of these dimensions can shape decentralization. For example, fragmented party systems represent a broader array of interests, including (possibly small) regional parties that strongly favor decentralization. Polariza- tion may preclude anything other than a gradual approach to decentralization for the simple reason that parties are too ideologically divided to agree on major legislative changes. Institutionalized parties are sufficiently powerful to insist that changes as potentially significant as decentralization be fully vetted through established party channels. In competitive multiparty systems, it is safe for development partners to assume that jockeying between parties for political advantage will dominate the design of decentralization. Although decentralization frequently pits national parties against one another, it can also be designed in ways that benefit all national parties. In Indonesia, until 2007, legislation stipulated that only nationally registered parties could run candidates in subnational elections—a major impediment to the emergence of regional parties that could have challenged the national party system more effectively.16 Considering the regularity with which intra-party and inter-party incen- tives have shaped decentralization, development partners would benefit from paying more attention to political parties as well as how the party system itself conditions the struggle over decentralization. Partisan dynamics can be com- plicated, but asking basic questions about how a political party is structured and how it competes with other parties for power can uncover sources of support for and opposition to decentralization.17 When development part- Understanding Political Incentives and Behavior 19 ners engage with local officials like mayors and municipal councilors, they are likely to gain greater analytical leverage if they simultaneously approach these officials as actors who are engaged in partisan struggles—both with national leaders of their own parties and with competitors in rival parties. Common Institutional Incentives to Defend the Center In both democratic and nondemocratic regimes, national authorities can be expected to approach decentralizing changes with caution. Notwithstanding the partisan and electoral dynamics highlighted above, national officials typi- cally face strong pressures to defend the national institutions they occupy and represent in interactions with subnational governments. Two different strains in the recent political science literature on decentralization demonstrate the salience of institutional incentives, despite differences in regime type. Consider first recent scholarship on how national officials approach the sequencing of administrative, fiscal, and political decentralization. The traditional public administration literature on decentralization has offered definitions of these three main dimensions that are now generally accepted, if still considered somewhat problematic (Schneider 2003; Ebel and Yilmaz 2003; International Monetary Fund 2009). More recent political science literature has begun to explore the logical possibility that these three forms of decentralization do not necessarily need to be adopted at the same time, or in the same measure. Given the institutional incentives faced by national politicians, there are good reasons to expect nonsimultaneous approaches.18 Two sequencing patterns seem to hold special appeal to national poli- ticians who are worried about the loss of political control. Both produce unbalanced decentralization outcomes. First is the adoption of political decentralization without administrative or fiscal decentralization, through which subnational politicians are elected in their own right, but are given few opportunities to use their offices in ways that might challenge the interests of national politicians.19 Second is the opposite pattern, where national politi- cians give subnational officials meaningful responsibilities and resources, but not political independence.20 Falleti (2005, p. 330) argues that national actors prefer administrative to fiscal decentralization and fiscal to politi- cal decentralization, whereas subnational governments prefer the reverse. Each of these sequences regularly produces different levels of autonomy for subnational governments.21 For example, whereas Falleti argues that mov- ing from political to fiscal and administrative decentralization generated a significant increase in subnational government autonomy in Colombia, the ability of the national government in Argentina to pursue the opposite sequence produced little change in the discretion of subnational officials. 20 The Political Economy of Decentralization Reforms in Developing Countries The question of sequencing may hold important clues as to who domi- nates the decentralization process—national or subnational politicians—and how external actors can help move the process along. It would be a mistake, however, to overstate the importance of sequencing or to conclude prema- turely that sequences preferred by national governments will succeed in keeping subnational officials under their control.22 For example, even where powerful national governments are able to devolve administrative responsi- bilities before money or political independence, mayors and governors can use administrative decentralization to lobby for additional changes that are more beneficial to subnational governments.23 Institutional incentive structures are also critical in understanding the choices that national government actors make as they consider whether to decentralize to intermediate or local-level governments.24 Recent political science research has suggested that, in reality, institutional incentives rather than normative concerns tend to dominate these decisions. As representa- tives of the central state, national politicians are influenced by the nature of threats to state formation and state survival, particularly in developing countries where these threats are significant.25 In Indonesia, after the fall of Suharto, fears among the military and other national institutions that decentralization could promote secessionist movements in resource-rich and ethnically or religiously identified provinces resulted in legislation that favored local governments over the more threatening provinces (Malley 2009, 139). Likewise, decentralization in Bolivia favored municipalities over regional governments due to concerns that regions would use de- centralization to challenge the state, already one of the weakest in Latin America (Grindle 2000). In Uganda, provincialism was likely avoided because of the perceived link with traditional kingdoms, which favored a more federal form of government that was not acceptable to the National Resistance Movement. Decentralization can also be designed to favor intermediate-level govern- ments over local-level governments, again for political reasons. In Mexico, that states benefited from decentralization more than municipalities reflects the leading role that state governors played as protagonists in the decentral- ization process (Grindle 2008). In Ethiopia, the loss of Eritrea and the fear of losing other ethnically identified states led to the development of what has been called the “ethnic federalism” constitution (Cohen 1995; Turton 2006), which focused on decentralization to states and mostly neglected local governments. In India, attempts to accommodate linguistic diversity after independence led to the creation of new states in a model of “linguistic federalism” that identified the state rather than the locality as the principal unit of subnational government (Mawdsley 2002). Understanding Political Incentives and Behavior 21 The relative treatment of subnational levels is a critical piece of in- formation for development partners who are seeking to support decen- tralization. This point is particularly germane for activities that seek to improve local outcomes and interface chiefly with local authorities. Where the center has greater cause to fear—and thus control—intermediate-level governments, local officials are often in a relatively privileged position. Where the center has favored intermediate governments, on the other hand, local officials may be much more engaged with and controlled by intermediate-level officials, who have their own political interests and agendas. In both cases, the center’s treatment of different subnational levels is best understood in the light of institutional incentives and not as the result of normative considerations. Presidents, Legislators, and Judges in Democratic Regimes By virtue of their identity as actors who inhabit the national government, national officials in all types of political regimes share incentives to defend the center. But in democratic regimes where power is divided among differ- ently constituted institutions, formal institutional arrangements can provide additional openings for decentralization. Presidentialism, for instance, creates specific incentives for the adoption of decentralization that may be absent from parliamentary systems (Carey and Shugart 1992; Linz 1994). Whereas the legislative and executive branches are elected separately in presidential systems, under parliamentarism the executive emerges out of a majority in the legislature—a majority that must sustain confidence in the executive in order for the government to survive. Under presidentialism, different parties or political groupings can control the legislative and executive branches, a scenario that is unlikely under parliamentarism, where the cabinet depends on the confidence of the assembly. Under presidentialism, opposition parties with a majority in the legis- lature can push for decentralization in an attempt to check the discretion of the governing party. In Argentina, the opposition Peronist party passed an automatic revenue-sharing law after winning legislative elections in 1987, thus limiting the ability of Radical President Raúl Alfonsín to use revenue transfers to bolster support for his reform agenda in Congress. The president was forced to sign the fiscal decentralization bill in exchange for opposition party support on critical stabilization measures. Just four years later, the shift to unified Peronist control of both the legislative and executive branches triggered movement in a more centralizing direction (Dickovick and Eaton 2004). Development partners working in presidential systems should thus be aware that decentralization dynamics may well correspond to sudden 22 The Political Economy of Decentralization Reforms in Developing Countries shifts between unified and divided partisan control over the executive and legislative branches. In both parliamentary and presidential systems, national legislators may have cause to find decentralization threatening. First, presidents and prime ministers can more credibly claim credit for the collective benefits generated by decentralization than can individual legislators. Second, and more impor- tantly, the decrease in national-level discretion implied by decentralization weighs especially heavily on legislators, who often depend on bringing public resources and government-sponsored projects home to their local constitu- ents.26 Transferring resources and responsibilities to subnational officials, who subsequently have a say in how these transfers are used in their jurisdictions, can directly challenge the time-honored role that legislators play as personal brokers of goods and services from the center. Fiscal decentralization in the form of automatic revenue sharing can be particularly troubling to national legislators by boosting the fiscal authority of subnational officials, who may be political opponents, while limiting the fiscal revenues that the national legislature can appropriate (and take credit for). As a result, legislators often operate as a brake on decentralization. In the Philippines, where legislators are deeply involved in pork-barrel politics, Congress successfully vetoed a number of decentralizing measures in the post-independence, pre-Marcos period (1946–1972). Only after President Corazon Aquino identified decentralization as her chief legislative goal, together with the imposition of term limits that led many legislators to anticipate subnational careers, did Congress agree to decentralization. After endorsing the automatic revenue transfers at the center of Aquino’s Local Government Code in 1991, legislators moved to appropriate additional monies that essentially duplicated these transfers, but that they could claim personal credit for delivering (Eaton 2002). In Kenya, using the excuse of poor local government performance, the Parliament adopted a Constituency Development Fund (CDF). The CDF represents a larger volume of funds than the intergovernmental transfer system that was instituted only a few years earlier, vastly complicating lines of accountability and undermining rational public budgeting. Development partners would benefit from understanding the institution- al practices that drive legislators’ decentralization preferences. In bicameral systems in which one chamber is dedicated to the representation of territo- rial interests, legislators in this chamber will likely have a special role to play in decentralization. The size of electoral districts also matters, as legislators in single-member districts are often able to personalize the delivery of gov- ernment resources to such a degree that decentralization can be extremely threatening. Whereas legislators are usually elected in subnational districts of Understanding Political Incentives and Behavior 23 varying sizes, in some countries legislators are elected in a single nationwide district (for example, Peru’s unicameral system until 2001, Colombia’s Senate after 1991, the Philippine Senate after 1986). These districting choices create different institutional incentives. In the Philippines, single-member districts in the House of Representatives gave legislators an interest in slowing down the decentralization process. In contrast, the Senate’s single nationwide dis- trict led senators to champion the cause of decentralization, as it alleviated concerns that mayors and governors might use decentralization to displace senators in subnational districts. In some countries, legislators are elected in districts that overlap with the electoral districts used to elect mayors and governors, whereas in other cases different electoral districts are used. The former rule is likely to increase legislators’ fears that subnational officials will use decentralization to challenge their institutional interests. Judicial institutions also have potentially significant, though as yet mostly unstudied, effects on decentralization. Whether decentralizing measures take the form of presidential or ministerial decrees, or new statutes endorsed by the assembly, judges are sometimes in a position to rule on their constitu- tionality. Where decentralizing reforms occur via constitutional revision, the institution of judicial review gives judges the authority to determine whether or not subsequent implementing legislation conforms to the new constitution. In countries where the judiciary enjoys sufficient independence from elected politicians, judges can be an important institutional resource for subnational officials who strive to hold the national government to the letter of the (decentralization) law. Given that poor implementation is often the weakest link in the decentralization chain, robust judicial institutions can prevent the executive and legislative branches from reneging on legislation that they formally endorsed but may not have fully intended to implement. In scores of developing countries and new democracies, the judiciary does not enforce decentralization provisions in response to legal complaints from mayors and governors because judges are simply not independent enough from the national politicians who can remove them if displeased. To develop a fuller sense of the political context in which decentraliza- tion is unfolding, development partners might consider whether judges are serving as referees in the conflicts that arise between national and subnational governments. In Mexico, growing judicial independence since the democratic transition in 2000 has reinforced the gains that decentralization has brought to subnational governments (Staton 2004). In the Philippines in 2004, the Supreme Court found in favor of local governments in a case charging the national government with a failure to honor decentralizing legislation by delaying and manipulating what were supposed to be automatic revenue transfers. 24 The Political Economy of Decentralization Reforms in Developing Countries Informal Institutions A complete institutional analysis must consider informal social norms that govern individual behavior and structure interaction between social actors (North 1990). Informal institutional constraints permeate politics in all types of regimes, but especially in nondemocratic settings, where formal institu- tional constraints are often ignored. Informal institutions come in myriad forms and are harder to identify and categorize than formal institutions, but their significance suggests that they should not be ignored. One particularly salient example of an informal institution that has shaped decentralization is clientelism, a concept that political scientists use to describe the informal but pervasive exchange of material benefits for political voice (Fox 1994). Though clientelism takes a variety of forms, it refers everywhere to a personal relationship between unequal individuals: a hierarchically superior patron with access to goods and services and a hi- erarchically inferior or politically subordinate client who surrenders his or her status as an independent political actor in order to receive benefits from the patron.27 Over the last decade, the resilience of clientelism has generated renewed interest among political scientists, who have debated its impact on development outcomes (Keefer 2007; Doner and Ramsay 1997) and the quality of democracy (Stokes 2007; Vicente and Wantchekon 2009).28 As an informal but often highly institutionalized practice, clientelism can influence how politicians approach decentralization. Many governing officials in both democratic and nondemocratic systems are in their positions precisely because they have mastered the art of clientelist exchange, regularly securing political support from local brokers in exchange for appointments to government jobs (patronage) and other governmental largesse, which these local brokers can in turn exchange for political support at still more local levels. Clientelism can thus be just as important in understanding a politician’s ascent to power as the internal party dynamics discussed above. By relocating decisions about revenue distribution and public expenditures, decentralization can directly challenge established clientelist patterns. In more centralized systems, national politicians have cause to worry that they will lose the ability to claim credit for the benefits they deliver. Decentralization can also lead to the reconstruction of clientelism at the local level, giving subnational political officials greater resources with which to build their own clientelistic networks and thereby displace their national patrons. For international development partners, collecting information about informal institutions and how they shape decentralization can be much more difficult than acquiring knowledge about formal institutions. But it is not impossible. For example, development partners may be able to gauge the Understanding Political Incentives and Behavior 25 importance of clientelism and estimate the related probability that national legislators will find decentralization threatening. One key indicator is the nature and extent of legislators’ participation in the policy-making process. In political systems in which legislators rely informally on the construction and defense of clientelistic networks in order to pursue their careers, they will typically devote little time to acquiring policy expertise or seeking to influence substantive policy debates in parliament. Development partners can similarly analyze the logic of political campaigns to generate informa- tion about the salience of clientelism. Where electoral outcomes turn largely on the ability of political machines to exchange votes for material benefits, campaigns will not center on the articulation of distinct policy or program- matic stances between rival parties. Finally, development partners should anticipate that different parties are not likely to be equally concerned by the threats that decentralization poses to clientelism because parties in a given country are not equally dependent on clientelistic exchanges (Calvo and Murillo 2004). Parties with poorer and more vulnerable constituents are more likely to deploy clientelism because small-scale material benefits are more likely to sway the voting behavior of these constituents. Analysts can therefore compare the strategies of different parties to develop a more sophisticated sense of the incidence of clientelism. Coalitional Incentives In evaluating decentralization reforms, politicians consider how proposed changes are likely to affect the coalitions that support them and/or the opposi- tion coalitions with which they are embattled. All political incumbents—re- gardless of whether they are operating in democratic or authoritarian contexts or in relatively stable or highly unstable environments—are engaged in the construction and maintenance of support coalitions. Whereas electoral, par- tisan, and institutional incentives force us to look within the state, coalitional incentives involve the analysis of societal actors more directly. Much of the political science literature suggests that, while the initiative to decentralize rarely comes from civil society, societal groups do shape decentralization by responding to the proposals of national and subnational politicians (Grindle 2000; Manor 1999).29 Other scholars place civil society at the center and politicians in the reactive role, attributing the adoption of decentralizing measures to the articulation of bottom-up demands from a range of social groups (Goldfrank 2007; Van Cott 2008; Wampler 2007; Yashar 2005). Development partners may have a more difficult time uncovering and assessing the coalitional incentives that politicians confront in the area of decentralization. In part because they are less visible than elections, party 26 The Political Economy of Decentralization Reforms in Developing Countries organizations, and institutions, coalitions do not lend themselves as easily to empirical analysis.30 As governments seek to project an image of unity and support for the common interest, it can be difficult to identify the coalitional imperatives and struggles that may occupy much of the government’s at- tention “behind the scenes,” particularly in single-party systems. Coalitional incentives are often more complex than the other incentives reviewed in this volume because coalitions reflect the intricacy of underlying social relations. Making sense of coalitional incentives in a given country typically requires attention to historical dynamics, geographic and topographical factors, and frequently deep-seated cultural and religious divisions. Despite these inher- ent challenges, coalitional analysis is worthwhile and important. To illustrate the potential gains, we examine below the coalitional dynamics that can be generated by widespread class and ethnic divisions. Labor Unions and Business Associations With respect to class-based coalitions, politicians must often consider the preferences of labor unions and business associations vis-à-vis decentraliza- tion. Drawing from their own experience with collective wage bargaining, labor leaders tend to take a dim view of decentralization.31 Particularly in countries with corporatist patterns of interest articulation, union organi- zations themselves typically reveal a high degree of centralization.32 Even in countries without a formal history of corporatism, but where decision making is centralized, attempts to maximize labor’s political leverage may lead unions to develop centralized organizational practices of their own. As a result, unions and union leaders may resist decentralizing changes that reduce the authority of the national government, where their collective leverage tends to be greatest. Unions have shaped politicians’ design choices in a number of important cases of decentralization. In Thailand, teachers’ unions resisted the trans- fer of teachers from the center to local governments when education was decentralized. In the Philippines, the steadfast opposition of public sector unions helps explain why the devolution of education was not included in the landmark 1991 Local Government Code—in contrast to health care, social welfare, and much greater shares of internal revenue. The national labor confederation in Bolivia similarly opposed the national government’s 1994 decentralization law, ultimately forcing it to relax the rules governing who could sit on the newly created civil society oversight committees so that unions could participate (Grindle 2000).33 In other cases, attempts to unmake and/or remake coalitions have shaped politicians’ attitudes toward decentralization. In Chile, decentralization Understanding Political Incentives and Behavior 27 appealed to dictator Augusto Pinochet because the transfer of schools and hospitals to municipal governments promised to permanently weaken the negotiating strength and relevance of national public sector unions.34 Union dynamics were also significant in Argentina and Mexico in the 1990s, where traditionally labor-based parties introduced market reforms and moved to restructure and reduce the significance of labor in governing coalitions.35 Business associations may have stronger reasons to support decentraliza- tion.36 If the collective power of labor can be brought to bear more powerfully at the national level, attempts to outmaneuver unions may lead the private sector to prefer more decentralized arrangements. Similarly, business as- sociations often expect that the decentralization of taxation and regulation will create subnational government competition, which businesses can use to their advantage in negotiating regionally specific investment deals with mayors and governors.37 Pro-market business associations may also support decentralization in an attempt to constrain national governments that have endorsed statist or socialist economic models.38 Decentralization can for other reasons fail to elicit the support of the business community. Where they are effective, umbrella associations that are designed to focus business pressure on the national government stand to lose authority when fewer critical decisions are made at the center.39 In countries that are contemplating the decentralization of fiscal authority via automatic revenue sharing, business associations are also likely to be concerned about problems of moral hazard; where subnational governments spend revenue they do not pay the political or administrative costs of collecting themselves, fiscally profligate behavior often results, along with bailouts, budget deficits, and greater inflationary pressures (Rodden 2006; Wibbels 2005). Finally, business associations worry that local governments will not have sufficient capacity to provide the decentralized services they need, including the con- struction and maintenance of roads, highways, and irrigation. Development partners may therefore wish to assess how powerful labor unions and leading business organizations view decentralization. One useful exercise would be to document the degree of decentralization in the structure of a country’s labor union apparatus. Labor unions are subject to a great deal of cross-national variation in the degree of discretion they allow union leaders below the apex of the organization. The extent of union centraliza- tion is significant, as highly centralized unions will likely work to oppose or reverse reforms that decentralize authority within the polity. Business associations also come in a variety of shapes and sizes. Where nationwide umbrella organizations exist, development partners could seek to establish whether these associations have a stated position on decentralization and how they are lobbying the national government on this issue. Where na- 28 The Political Economy of Decentralization Reforms in Developing Countries tionwide associations are absent, development partners could benefit from determining if regional or sectoral business organizations are active on the topic of decentralization and whether they focus their lobbying efforts on national or subnational authorities. Ethnic Cleavages and Customary Authorities In ethnically divided societies, national politicians tend to privilege concerns about the likely impact of decentralization on the monoethnic or multiethnic coalitions that brought them to power and/or sustain their rule. Particularly where ethnic cleavages overlap closely with territorial divides (Amoretti and Bermeo 2004), the decision to decentralize is likely to have far more to do with the negotiation of interethnic coalitions than with service delivery, poverty reduction, or administrative efficiency. Many recent episodes of decentralization can be understood as at- tempts by politicians to end long-standing ethnic and religious conflicts. Decentralization may aid in the construction of new ethnic and religious coalitions in the hope of sustaining peace, not unlike “holding-together” forms of federalism that have enabled the state in diverse societies to fore- stall secessionist movements—as in India and Indonesia (Stepan 2001). In the Democratic Republic of Congo, following three decades of highly cen- tralized rule (1965–1997), ethnic tensions from the neighboring Rwandan war spilled over into the country and provoked a brutal five-year civil war. As part of negotiations to end the war, leaders of different armed factions agreed to several decentralizing changes that were inserted into the country’s 2006 constitution.40 Ethnic conflict has encouraged politicians to introduce decentralizing measures in such disparate cases of post-conflict negotiation and constitution writing as Afghanistan, Angola, Bosnia, Cyprus, Ethiopia, Iraq, Mozambique, Nigeria, South Africa, Sudan, and Sri Lanka. In other ethnically divided societies, politicians have resisted decen- tralization as potentially threatening and destabilizing. Where ethnic ac- commodation occurs largely in the form of power-sharing arrangements between rival groups that control various ministries in the national cabinet, decentralization may limit coalition-building options at the center and undermine “national unity” governments.41 Based in large part on his study of Nigeria, Horowitz (1985) argues that where substate units are ethnically homogenous, decentralization may help replace troubling interethnic conflict at the national level with less problematic intraethnic conflicts that emerge in competition for decentralized resources. However, where substate units are ethnically heterogeneous, decentralization can multiply the sites of in- terethnic conflict and their intensity. Finally, and perhaps most commonly, Understanding Political Incentives and Behavior 29 national officials in ethnically divided countries often worry that, far from limiting conflict, decentralization will provoke secessionist movements, especially when it takes the form of fiscal transfers that could be used to finance these movements.42 Where ethnic cleavages are politically salient, development partners will want to determine if national politicians see decentralization as a means to balance ethnic divisions and/or accommodate ethnic diversity, and whether they think it will bolster or threaten governing coalitions. In addition, de- velopment partners will want to be attentive to the likelihood that ethnic cleavages will generate asymmetric approaches to decentralization. According to Boone’s research in West Africa (2003), the pursuit of distinct coalitional strategies vis-à-vis local elites in different parts of the same country led to a regionally differentiated pattern of decentralization. In Senegal, for example, the material and spiritual authority of rural notables in the Woloff groundnut basin led the national government to rule this region “indirectly” via these notables, who received devolved powers. The absence of similar local elites in the region of Lower Casamance produced a more centralized, top-down pattern of rule. As Boone’s analysis demonstrates, formal decisions about decentralization are still made by high-level national officials (with whom development partners often interact quite closely), but these decisions are shaped largely by underlying societal structures and coalitional alliances. Examining the role of customary authorities also helps to understand how politicians approach decentralization and coalition building in the con- text of ethnic diversity. In many developing countries, customary authorities (such as tribal chiefs, local headmen, and village elders) wield substantial decision-making authority over the most critical and divisive local issues, including the use of communal lands and forests, access to wells and other local infrastructure, and mediation of local disputes.43 Customary authorities may also function as important coalition partners for national incumbents, particularly where political parties and other mass organizations do not play significant aggregating roles. As a result, national governments (and inter- national development partners) often decide to “recognize” the authority of customary actors to the detriment of elected local governments (Ribot 2004; Ribot, Chhatre, and Lankina 2008).44 Focusing on politicians’ coalitional in- centives can offer a useful window onto how interactions between customary authorities and subnational governments affect the decentralization process. Notes 1. There are still good reasons to be skeptical when national politicians announce that they will decentralize or have decentralized. Likewise, there are good reasons to sus- 30 The Political Economy of Decentralization Reforms in Developing Countries pect that politicians will approach decentralization in ways that reduce the potential loss of political control. Approaching decentralization as a paradox thus holds clues as to why decentralizing reforms are often partial and incomplete. 2. In some cases, national governments have endorsed the shift from appointment to election of subnational officials, but reserve the right under certain conditions to dismiss the individuals who are consequently elected, a prerogative that dramatically constrains the independence of subnational officials (Bland 2010). 3. In addition to their impact on national governments, elected local governments can have profound implications for customary authorities at the local level, who may fear displacement by local elected bodies that have been strengthened through the transfer of resources and responsibilities from the center (Ribot 2004). 4. Nothing guarantees that nonconcurrent subnational races will actually turn on local issues, however, and subnational races may well operate as a sort of referendum on the performance of the national government when they are held non-concurrently. Likewise, dominant or hegemonic parties can try to use nonconcurrent subnational elections as an opportunity to build and strengthen local electoral networks that reinforce their performance in national elections, as in Cambodia, where the first commune council elections were held in advance of a new round of national elec- tions to lay the groundwork for them. 5. Profound uncertainty regarding the likely performance of different actors in elections after the transition may also, as Smith (2008) argues for the Indonesian case, reduce the relevance of electoral incentives as a cause of decentralization. 6. The regional electoral strength of the National Party in Western Cape and of the Inkatha Freedom Party in Kwa Zulu Natal led them to demand the inclusion of guaranteed revenue transfers to provincial governments in the new Constitution. Agreeing to decentralization thus constituted an important component of South Africa’s negotiated transition away from apartheid and toward democratic rule (Dickovick 2007; Heller 2001). 7. In the end, the decision to (re)introduce elections at both levels reflected a historic compromise by politicians across the ideological spectrum who sought to bolster their electoral prospects (Bland 2004; Eaton 2004b). 8. Once subnational elections were reintroduced in the course of redemocratization in the 1980s and 1990s, military-era fiscal and administrative changes that had enhanced the stature of subnational governments took on even greater importance because they were now controlled by separately elected politicians (Eaton 2006). 9. In addition to intra- and inter-party dynamics, the decentralization literature has emphasized the importance of parties that are regional rather than national in scope. Brancati (2006), for example, argues that decentralization can encourage ethnic conflict and secessionism by bolstering regional (as opposed to national) parties. 10. The evolving balance of power within parties may be particularly critical in Latin America, where parties are on average older than in other developing regions. 11. Brazil is a case in point, where the re-emergence of governors as powerful actors in the course of the democratic transition in the 1980s enabled them to leverage highly advantageous changes in fiscal transfers without taking on additional administrative responsibilities (Samuels 2004). Understanding Political Incentives and Behavior 31 12. Diaz Cayeros (2006) shows that, before the creation of a hegemonic governing party, Mexican governors in the 1920s and 1930s refused to delegate their fiscal author- ity to the national government. Fiscal centralization occurred only in the 1940s, by which time Mexico’s governing party had become sufficiently institutionalized and powerful to protect governors from the adverse consequences that supporting fiscal centralization might have had for their political careers. Eaton (2004a) finds additional empirical support for partisan theories of decentralization. According to Eaton, waves of democratization in the 19th and 20th centuries in Latin America produced demands for decentralization only in those cases where subnational politi- cians dominated political parties. 13. Currently, decentralization is being expanded from the communes to higher-level governments through indirect elections of district and provincial councils in ways that seem likely to further strengthen the dominance of the CPP. 14. The Uganda case is discussed in Francis and James (2003) and Smoke (2008b). 15. See also Sabatini (2003) on the cause and effect of decentralization within parties. 16. Indonesia currently imposes restrictions concerning cross-national representation for political parties. A political party must maintain branches in 60 percent of the country’s provinces and offices in at least 50 percent of districts and municipalities within those provinces. With an accompanying parliamentary threshold of 2.5 percent, the law is designed to restrict the entry of smaller parties into parliament. Candidates in direct elections for head of region (districts and provinces) must be nominated by parties with at least 15 percent of national representation. An exception made in the context of Aceh’s Special Autonomy, which allowed for local parties, was changed in 2007. Notably the Aceh Party allowed the Free Aceh Movement (Gerakan Aceh Merdeka) to run candidates in provincial and local elections in the absence of national representation. However, local parties have not been granted national representation. See Ismail, et al. (2009). 17. Devoting analytical energy to partisan incentives can also be worthwhile in that they are likely to be less volatile than electoral incentives, which may repeatedly shift in response to short electoral cycles. 18. Whereas the simultaneous transfer of significant levels of administrative, fiscal, and political authority could dramatically reduce the power of national politicians relative to subnational politicians, nonsimultaneous transfers of those powers pose fewer threats to politicians at the center (as do, of course, simultaneous transfers of only limited amounts of administrative, fiscal, and political authority). 19. For example, in Italy politicians at the center agreed to introduce regional elections in the 1970s only because they were confident they could successfully veto other forms of decentralization that would make regional officials threatening. They were able to keep this up for nearly two decades (Putnam 1993). 20. For example, under Deng Xiaoping, China’s Communist Party leadership transferred to provincial governments a significant measure of fiscal and administrative author- ity, without allowing a transition to subnational democracy (Weingast, Qian, and Montinola 1995). 21. Note that, while the chief executive may prefer administrative to fiscal decentraliza- tion, when administrative decentralization involves the devolution of personnel, it may provoke the strong opposition of sectoral ministries within the national execu- tive, who may not prefer it to fiscal decentralization. 32 The Political Economy of Decentralization Reforms in Developing Countries 22. Sequencing may be less critical, from the standpoint of national politicians, where there is a substantial gap between de jure and de facto levels of decentralization. In these cases, national politicians can endorse legislative changes that increase the de jure autonomy of subnational governments in all three dimensions, safe in the confidence that incomplete implementation will preserve their de facto dependence on the center. 23. Although sequencing has occurred in many countries, there are important cases in which decentralization has combined simultaneous changes in political, adminis- trative, and fiscal authority. In Uganda, these three facets of decentralization took place at the same time in the context of post-conflict transformation. In Indonesia, a politically motivated “big bang” framework involved considerable changes in all three dimensions. 24. In addition to the world’s federations, which typically include robust intermediate- level governments, some unitary countries have introduced and are in the process of strengthening levels of government that operate between the national and municipal levels (e.g., states, provinces, departments, regions). 25. The dominance of political calculation is particularly common where officials in national institutions face possible secessionist threats, or where incomplete processes of state formation effectively limit the state’s territorial control in subnational regions. 26. The pressure to deliver these benefits can be particularly pronounced in marginal- ized districts that are not economically vibrant or that do not have dynamic private sectors and thus depend heavily on public funds. 27. The concept can also extend to an interconnecting chain of dyads through which the most marginalized individuals in peripheral regions may be linked with governing officials at the very apex of the political system. 28. Most of this literature suggests that clientelism diverts government expenditures away from better uses of public revenue, and that it simultaneously limits democracy by preventing citizens from forming the type of collective organizations that can hold politicians accountable. 29. Societal dynamics matter even in these politician-centered accounts; in the decision to decentralize and in the design choices that decentralization requires, politicians often take into consideration the likely impact on their supporters and opponents. 30. It is also harder to generalize about the dimensions along which coalitions vary. For example, whereas it is possible to compare all political parties according to how much authority they give national as opposed to subnational party leaders, and whereas party systems can be categorized as multi-party, two-party, or single-party dominant systems, coalitions are subject to much greater heterogeneity. 31. Decentralized collective bargaining, as in the influential Japanese example of en- terprise unionism, tends to limit the ability of unions to strike the best deals for rank-and-file workers. 32. As it is usually practiced in developing countries, corporatism involves attempts by national governments to incorporate labor into the political system as a formal coali- tion partner. According to the bargain at the heart of corporatism, labor unions lose some of their independence but gain regularized access to decision-making channels at the center (Collier and Collier 1991). 33. Union dynamics may also enter into the calculations of subnational politicians. For instance, in Ecuador, where the national government decentralizes services on Understanding Political Incentives and Behavior 33 a case-by-case basis in response to petitions from subnational governments, several mayors have refrained from requesting the transfer of schools due to concerns about possible political conflicts with teachers’ unions (Cameron 2009). 34. These unions were an important component of the coalition that in 1970 brought Socialist President Salvador Allende to power; in the wake of his overthrow, decen- tralization can be understood in part as a political strategy to fragment this coalition. 35. Teachers’ unions in Argentina and Mexico opposed the decentralization of educa- tion and extracted important concessions in the latter case, including pay raises and establishment of a federal teacher statute (Murillo 1999, 43). 36. Political scientists have tended to study labor as a political actor in greater detail, but a growing literature has emerged on the associational life of business (Handley 2008; Maxfield and Schneider 1997; Schneider 2004; Durand and Silva 1999). 37. In Brazil, the decentralization of the value-added tax has operated as a key enabling condition for the so-called fiscal wars (guerras fiscais) between the states, through which automotive and other industries have secured advantageous investment terms (Tendler 2000). 38. In Bolivia, chambers of industry, finance, and agriculture have spearheaded the au- tonomy movement in Santa Cruz, which seeks to decentralize resources and policy authority away from the national government that is now controlled by President Evo Morales’ Movement Toward Socialism (Eaton 2007). 39. Furthermore, these cross-sectoral associations typically focus not on sectoral policy but on broad indicators of macroeconomic health, including fiscal balance and inflation rates. 40. The new constitution increased the number of provinces from 11 to 26, reversing President Mobutu’s earlier attempt to centralize power by reducing the number of provinces. The new constitution also seeks to balance pressures from differently situated provinces by stipulating that provincial governments can keep 40 percent of the tax revenues they collect, which benefits wealthier provinces, and by establishing a solidarity fund for investment in less developed provinces. 41. Politicians may also fear that decentralization will destabilize fragile ethnic coalitions if relevant groups perceive that reform disproportionately benefits some coalition members relative to others. 42. In the Philippines, the Christian-dominated national government’s concern about how armed actors might use expanded powers in the Autonomous Region of Mus- lim Mindanao has held back the transfer of greater resources and responsibilities to this region. 43. As such, customary authorities often exert greater influence than the formal subna- tional governments that officially rule at the local level and that ostensibly have the final say on these issues. 44. Ribot’s research shows that decisions to privilege customary authorities relative to incipient subnational governments can have negative consequences for the quality of representation, the meaning of citizenship, and the scope of the public domain. According to Ribot and his coauthors, customary and indigenous authorities may be far less representative than they at first appear, they may operate in ways that reinforce authoritarian and exclusionary practices at the local level, and their continued rule may depend on the repression of intra-group differences. 4 Understanding Bureaucratic Incentives and Behavior W hile the recent political science literature on decentralization reviewed above has focused on politicians, it is important to extend the analysis of incentive structures to include nonelected individuals in the bureaucracy. Bureaucrats do not merely imple- ment the decisions that politicians make—they are political actors in their own right, who pursue their own career and institutional interests (Aberbach, Putnam, and Rockman 1981; Cleaves 1974). Decentralization more often than not takes place in complex and (sometimes purposefully) poorly coordinated institutional environments. The incentives of those who manage and operate government agencies can be of considerable importance, particu- larly for how decentralization is implemented. There has not been much attention in the academic literature to how institutional and individual incentives affect decentralization, although there has been some work on how incentives affect the performance of those who undertake local development activities that may involve subnational governments.1 Even if there is (or appears to be) broad national political con- sensus to decentralize, we have seen above that such agreement is commonly based on goals other than the stated normative targets of mainstream international decentralization policy. Even in one-party or dominant-party systems, the various central agencies involved may have different opinions on how far decentralization should go and what their role in the decentralization process should be. In some cases, central government agencies may overtly or covertly 35 36 The Political Economy of Decentralization Reforms in Developing Countries obstruct the decentralization of major responsibilities when this reduces their political power and undermines their control over substantial resources. This behavior may be motivated mostly by institutional or individual concerns, but these can also be linked to the electoral or partisan incentives elaborated above. Perhaps the most central issue with respect to the bureaucratic dynamics of decentralization is the fact that national government agencies may have little or no desire to assume or divest responsibilities as envisioned under decentralization laws if they are not pressured to do so. In addition, they may have few incentives to work cooperatively with other agencies, even though coordination is normally seen as crucial for the “success” and sustainability of decentralization in terms of improved processes and outcomes. In some cases, powerful central agencies—and the departments and/or divisions within—may engage in direct competition for control of the decentraliza- tion agenda and the substantial internal and external resources that may be involved in its implementation. They may, if they have the power to do so, pursue policies and take actions that are inconsistent with the formal decentralization framework. As discussed above, bureaucratic politics can have an impact on the decision to adopt decentralization and foundational principles for how it should operate. It is in the detailed design and implementation of de- centralization, however, where bureaucratic politics can really take on a life of its own, including (intentionally or unintentionally) undermining the legal framework and/or moving decentralization in new directions. Without understanding such behavior and why it occurs—including how it may be linked to larger national political forces, legislative bodies, political parties, other central government agencies, or international do- nors—those international development partners preparing decentraliza- tion support are likely to make significant mistakes and design programs that have limited and/or unsustainable impacts or that may even cause more harm than good. In this section we review the range of central government actors typically involved in decentralization, characterize the positions they are most likely to take regarding decentralization, and consider how these may generate cooperation or competition. We also look briefly at the incentives of devel- opment partners to support decentralization and how they interact with the positions and incentives of country actors. Finally, we review the range of mechanisms that have been used to try to mediate the often-substantial problem of interagency conflict and to promote coordination around de- centralization, including with development partners. Understanding Bureaucratic Incentives and Behavior 37 Key Actors and Basic Roles A first step in understanding the bureaucratic politics of decentralization is to understand which actors are involved in the process and what role(s) they play, formally and informally. Both historical trajectories and current institutional configurations and dynamics are critical for understanding the orientation of particular agencies and their positions and interrelationships with respect to decentralization. Several major types of national agencies are typically pertinent for de- centralization reform. There is often an agency dedicated to local government or local administration, commonly a Ministry of Local Government, Home Affairs, or Interior. A Ministry of Local Government is normally found in developing countries where there has been some (often colonial) system of elected local government that may once have functioned with some degree of genuine autonomy from the central government, as in former British colonies in East and Southern Africa. A Ministry of Interior or Home Affairs is more typically found in countries with centralist systems of territorial and institutional control, as in former French colonies in some parts of West and Central Africa or Southeast Asia. A Ministry of Local Government is generally more likely to understand and support mainstream ideas about decentralization than a Ministry of Interior or Home Affairs. There are cases, however, in which the former have been more interested in increasing their own functions and resources relative to other central agencies than they have been in empowering local governments (such as Kenya), and there are cases in which the latter have to some extent genuinely supported the development of stronger and more autonomous local governments (such as Indonesia). There are invariably key cross-cutting agencies that manage overall national development policies and resources, such as a Ministry of Finance, Planning, or Civil Service. Given the overarching nature of their functions, these agencies usually have some degree of legal and administrative jurisdic- tion over most other government agencies, including those at the subnational level. In many countries, these agencies have inherent missions or tendencies toward standardization and central control, as in carefully managing public finances, ensuring the effective use of limited public investment resources, or optimizing the deployment and performance of public sector employees. It is not fair to say, however, that all central agencies with cross-ministerial functions always obstruct decentralization reform. There are cases in which these types of agencies (or departments or individuals within them) have led or supported efforts to increase local government functions and autonomy. Sectoral (line) ministries are also important actors in the bureaucratic arena of decentralization. These agencies have responsibility for traditional 38 The Political Economy of Decentralization Reforms in Developing Countries sectoral service delivery functions such as education, health, infrastructure, social services, transport, and water, in various configurations. In some countries, portions of these sectoral portfolios may be coordinated under territorially oriented national bodies, such as a Ministry of Regional, Rural, or Urban Development. Typically, sectoral agencies are heavily oriented toward service delivery, so they are often much more concerned, for example, about health, education, and infrastructure services and outcomes than about build- ing subnational institutions and governance processes. In many countries, sectoral ministries exert a powerful centralizing tendency, either through heavy management and regulation of resources officially devolved to local governments for services related to their sectoral area of competence (as in Uganda) or through continued delivery of services even when those func- tions have been assigned to local governments (as in the Philippines). Again, there are cases where sectoral ministries have promoted decentralization and local government capacity building for service delivery in their sector, so they need not automatically be presumed to be hostile or indifferent to local government. Sectoral ministries, however, are on balance more inclined to support deconcentration to subnational offices of national ministries or delegation to service delivery management units (such as school boards) than to support devolution to subnational governments. As public functions are decentralized, all of these agencies experience a number of changes. Even in a modest decentralization, there is by definition some de jure loss of control over their respective jurisdictions—for example, local governments are expected to assume certain revenue-raising powers formerly controlled by the Ministry of Finance, requirements for Ministry of Interior or Ministry of Local Government review and approval of lo- cal decisions are officially relaxed, or decisions on the use of resources for some health services are legally made by local governments instead of the Ministry of Health. There may be some transfer of responsibilities across national ministries; for instance, a Ministry of Local Government may take over responsibility for monitoring some local government activities formerly managed by the Ministry of Finance or the Civil Service Commission. In most cases, decentralization fundamentally alters the roles that many central agencies play, with a movement away from decision making and control to facilitation, support, and monitoring. Although the loss of central agency power and interagency competi- tion for a prominent role in decentralization provide powerful incentives for bureaucrats to engage in overtly or covertly obstructive behavior or careful institutional positioning, less obvious dynamics can also be at work. Especially in least developed countries with weak institutions, informal mechanisms for raising and managing resources may trump formal systems. Understanding Bureaucratic Incentives and Behavior 39 These informal mechanisms are often clientelistic or patrimonial in nature, and it is difficult for reformers to change these established rules of behavior.2 This is particularly true where there is limited transparency, where informal rules are somewhat derivative of or validated by traditional cultural norms, or where civil service salaries are so low that government staff feel compelled to extract whatever resources they can out of the system by whatever means feasible. Even where it is possible to pursue reform under such circumstances, it is likely to take a long time. Within specific institutions, individual employees will naturally be concerned about their own career trajectories. Often this will require behavior consistent with the position of the agency they work in and/or the interests of their immediate supervisor. It may be the case that (even well-placed) civil servants who are supportive of decentralization may be unable or unwilling to take actions that reflect this support if their views are inconsistent with the official or dominant stance of their agency or su- pervisor. Thus, in some cases, the people who by personal disposition and official position may seem best placed to promote decentralization reform consistent with the values and priorities of development partners (and in some cases official government policy documents) may in fact not be able to advance even basic reforms. Government Agency Interrelationships It is important to understand how the various national agencies involved in decentralization relate to each other in terms of de jure and de facto roles, the relative power enjoyed by each, and the sources of that power (finan- cial, legal, and/or political). There are some fairly predictable differences of opinion and conflict. Friction is relatively common, for example, between Ministries of Local Government and Ministries of Finance. As noted above, the Ministry of Local Government is often (if not always) trying to expand the role of local governments and to enhance its own relationship to local governments, while the Ministry of Finance is not uncommonly reluctant to relinquish both meaningful fiscal powers to local governments and oversight of local governments to the Ministry of Local Government. Since Ministries of Finance are typically stronger than Ministries of Local Government (be- cause of more influential political connections and the power inherent in official control over the national budget), they often have the upper hand and can sometimes preclude the effective assumption of Ministry of Local Government roles and policies in ways large and small. In South Africa, ef- forts by the Department of Provincial and Local Government to develop the local government system have often been at odds with the Department of 40 The Political Economy of Decentralization Reforms in Developing Countries Finance, which usually prevails. This has also been true to a great extent in Uganda, where the Ministry of Finance, Planning, and Economic Develop- ment almost always prevails over the Ministry of Local Government, even in areas where the latter has formal legal jurisdiction. Similar dynamics can also emerge with other broad-based central agencies, such as the Ministry of Planning and the ministry or commission that oversees the civil service. In other cases, the Ministry of Interior has had the upper hand over the Ministry of Finance in the development of decentralization reform policy. In Cambodia, the Ministry of Interior is closely aligned with the ruling party, and the Minister serves as Deputy Prime Minister. This Ministry has historically been a powerful agency in charge of territorial control and public administration, while the Ministry of Economy and Finance had little ef- fective control over how powerful provincial governors (appointed by the Ministry of Interior) raised and spent resources until public finances were recentralized in the post-conflict period. The net effect of these political relationships and historical roles in Cambodia has been to somewhat offset the “natural” advantage often enjoyed by finance ministries through their role in controlling the public purse. In still other cases, there are multiple cross-cutting national agencies with sufficient legal, financial, or political power to create a more complex bureaucratic dynamic. Indonesia is an interesting case in point. A number of national ministries and agencies have been heavily involved with local and regional development since well before decentralization occurred in the post-Suharto era. The Ministry of Finance, the Ministry of Home Affairs, Bappenas (the national planning agency), and the Ministry of Public Works have all been able to maintain substantial influence in the way decentral- ization plays out on the ground. This situation has resulted in conflicting policies and inattention to the implementation or harmonization of some critical aspects of decentralization reform. It would be a mistake, however, to assume that an individual central government agency necessarily has a unified position on decentralization. There are cases in which dueling departments within a cross-cutting ministry can complicate efforts to enhance decentralization. The department in the Indonesian Ministry of Finance in charge of intergovernmental fiscal matters, for example, has been trying for years to devolve the centralized property tax to local governments and to expand opportunities for local governments to borrow for infrastructure development and for donors to on-lend to local governments. This department has been consistently undermined by the department that has long managed the property tax for the central govern- ment and by the department that currently oversees mechanisms for donor on-lending and central government lending to local governments. To some Understanding Bureaucratic Incentives and Behavior 41 extent this is just an intra-agency power struggle over policies that involve substantial resources, but it may also be seen as part of the bureaucratic culture—nobody in a sufficiently senior position seems willing to take a stand that would cause major departments and senior officials within the ministry to lose face. It is possible, of course, that the individual bureaucrats involved have political links that protect their positions and preferences. A similar intraorganizational dynamic has been important within Cambo- dia’s Ministry of Economy and Finance, such that the Treasury director’s strong political connections have allowed him to neutralize a number of decentralization-enhancing reforms proposed by the budget department and apparently supported by the senior minister. The centralist tendencies of sectoral agencies, discussed above, often coincide conveniently with those of finance and planning agencies, entirely overpowering relatively weak Ministries of Local Government or Interior. But this is not always the case. Sectoral ministries sometimes adopt their own systems and procedures for managing services—such as mechanisms for channeling resources or special sector-specific districts unrelated to subnational administrative jurisdictions or governments—that are at odds with official decentralization policy and formal intergovernmental finance, planning, or human resource management systems promoted by cross- cutting ministries. Even though finance or civil service ministries should de jure be able to discipline such behavior, they may not be in a position to do so because of the powerful political connections of other ministers or the preferences and actions of international development partners who bring considerable sectoral resources to the table. The Role of External Development Partners Donor support is often framed around technical and normative issues, with little attention to political and institutional realities. As previously explained, individual donors sometimes must (or feel they must) support their own formal institutional or departmental policies even when these lead to pro- grams that do not fully support official government policy or decentraliza- tion laws in the country they are supporting or when they are inconsistent with programs supported by other donors. In order to facilitate compliance with their own management and accountability requirements, for example, development partners may insist that the government adopt their preferred procedures and modalities, and in many cases they continue to work through separate implementation units or other parallel or semi-parallel mechanisms. Inconsistent procedures and separate mechanisms can inhibit the develop- ment of a unified public sector system and place a significant burden on 42 The Political Economy of Decentralization Reforms in Developing Countries government counterparts at the national and subnational level. Despite considerable rhetoric about institution building, the immediate interests of development partners are not always well served by it. Such efforts can be time consuming and difficult, and may therefore delay the disbursement of resources needed to justify funding requests for the following year. The pressure to follow disbursement schedules, however, may keep funds flowing even when it is or should be clear that the program is not really supporting and may even be undermining decentralization. When a particular donor has specific priorities and mandates, its per- spectives on decentralization may not be compatible with those of other donors. If decentralization is viewed through the lens of public financial management or civil service reform, there may be a tendency to focus on central standards and control. A sectoral lens often frames decentralization in terms of deconcentration or facility-based autonomy rather than local government roles. A local government perspective is likely to emphasize local accountability and governance and the granting of fiscal autonomy. When working to empower citizens and communities through the promotion of local democracy or community-driven development, sometimes owing to mistrust of formal government institutions, donor initiatives may ignore or bypass (and potentially undermine) elected and legally empowered local government councils. In any of these cases, development partners may focus on specific aspects of reform and target only certain subnational levels of government or individual jurisdictions and ignore others, or in some cases sidestep all levels of government. These concerns about donor behavior do not end with interdonor behav- ior. As suggested earlier, individual development partners are also internally diverse—individual departments within the same agency may be fragmented along dimensions similar to the central government bureaucracies of de- veloping countries, and they may support or undermine decentralization accordingly, often on the basis of their philosophy and mission rather than on the basis of empirical evidence. These perspectives and derivative actions—both between and among development partners—can be highly consequential for decentralization. Different agencies or different units of a large organization like the World Bank may work primarily with specific agencies in client countries that hold compatible views on decentralization, local development, and service delivery. But these agencies almost invariably constitute only a portion of the diverse set of country actors and incentives involved in decentralization and local development. More than likely, they concentrate only on particular, limited aspects of reform, and their views and approaches may clash substantially with those of other national agencies. Understanding Bureaucratic Incentives and Behavior 43 In other words, development partners, like national agencies in the countries they work in, are driven by a set of heterogeneous and sometimes incompatible objectives. In pursuing these objectives, they and/or their con- stituent departments may reinforce tensions among these national agencies by defining activities that are mutually beneficial to the two parties directly involved but not necessarily compatible with official government decentral- ization policy. Thus, one donor or donor department may provide support to public financial management or civil service reform programs with a Ministry of Finance or Civil Service, but develop systems that differ from those being defined under a local government reform program supported by a different donor, or a different department of the same donor, working with the Ministry of Local Government. In Uganda, multiple development partners and one World Bank project were working on local government financial management procedures with the Ministry of Local Government, while other donors and a different World Bank project were supporting public financial management reforms that involved developing an entirely different local financial management system. At the same time as other reforms are underway, a sectoral department of one or more donors may be promoting service delivery reform for a par- ticular sectoral ministry using systems and procedures that may or may not coincide with those being promoted under reform efforts in public financial management, civil service, or local government. The use of sectorwide ap- proaches, which coordinate different donors working on the same sector but rarely coordinate donors or departments working on different aspects of public sector or subnational reform, has become widespread. The tendency of sectorwide approaches to centralize initiatives under a sectoral ministry may sometimes coincide with the centralizing tendency of public financial management reforms under a finance ministry, and, more rarely, with the more decentralization-oriented efforts of a local government ministry. But is it not uncommon for these various reforms being undertaken at various ministries with the support of different development partners to push the development of institutions and procedures—and their outcomes—in in- consistent directions. The Challenge of Coordination: Mechanisms and Realities3 Given the complex landscape of institutional actors and their divergent objectives and levels of power, it has long been recognized that policies and programs need to be coordinated and harmonized. Some type of informal or semiformal cross-departmental mechanism to develop, coordinate, and implement decentralization policies and to mediate interdepartmental differ- 44 The Political Economy of Decentralization Reforms in Developing Countries ences is common, but typically not very effective, either because policies were not intended to be effective (due to the political and bureaucratic realities reviewed above) or because they were poorly designed. Informal or Ad Hoc Mechanisms In some countries, there is not really anything that could be considered explicit coordination of decentralization; instead, there is at best a reliance on ad hoc coordination of existing national agencies that interact with subnational governments. In China, where significant decentralization is underway without being framed as a formal policy, regular central govern- ment ministries “manage” decentralization by interacting with subnational governments in the course of their routine operations; there is limited formal coordination. Similarly, in Vietnam, decentralization has been implemented and managed through existing institutions. Various central government ministries, particularly Finance, Planning, and Investment and their pro- vincial counterparts, have played a key role in managing decentralization, but without formal coordination. At the same time, Vietnam has a relatively well-articulated and hierarchical system that is reinforced by the single-party structure and is changing only gradually, so the lack of formal coordination is not as problematic as it has been in some other countries.4 In other cases, there are somewhat more formal attempts to coordi- nate, at least for certain purposes. The Philippines, for example, established an interagency oversight committee to take responsibility for monitoring implementation of the Local Government Code, but most of the real power lies in individual agencies that are expected to follow its provisions but do not always see eye to eye. In South Africa, the national government established a number of mechanisms not to coordinate decentralization per se, but to promote in- tergovernmental relations. These include sector-specific Intergovernmental Relations Committees of Ministers and Members of Provincial Councils (MinMecs) and the President’s Coordinating Council, which aims to improve overall intergovernmental and intersectoral coordination. The MinMecs are fairly inclusive, but are primarily consultative and have limited authority. The President’s Coordinating Council, on the other hand, is a high-level body that focuses on the provinces; local governments are a secondary concern. A proposal to more broadly improve integration at the central level between the Department of Finance and the Department of Provincial and Local Government through a local government policy coordination mechanism under the Office of the Vice President (which is charged with government policy coordination) never materialized, most likely because the powerful Understanding Bureaucratic Incentives and Behavior 45 Ministry of Finance did not support it and probably would have continued to have the upper hand even if the mechanism had been created. More Formal Mechanisms A number of countries have attempted to facilitate interdepartmental and donor coordination by empowering some relatively broad-based body with responsibility for coordination and to give this body the means to enforce its authority. There are at least three broad approaches. One involves con- stitutional, legislative, or administrative provisions that give a single central government department the responsibility and power to coordinate all decentralization and subnational government reforms. This is particularly common in former British territories where relatively independent local governments were set up in the colonial era. Kenya’s Local Government Act, for example, gives this power to the Ministry of Local Government, and Uganda empowered a Decentralization Secretariat under its Ministry of Local Government when it decentralized in the 1990s.5 Single departments rarely coordinate effectively, however, because they are seen as competitors or inferiors in the hierarchy by other ministries whose cooperation is required for decentralization. Other ministries may not cooperate with mandates that they feel reduce their own power or pres- tige. Single-department coordination can only work if the responsibilities of each agency are clear, and if the coordinator is well respected, sufficiently resourced, and adequately empowered to require compliance. If functions are unclear, if the coordinating ministry is seen as biased or self-interested, or if it has insufficient legal authority or political power to enforce coordi- nation, it is unlikely to perform effectively. In Kenya, the Ministry of Local Government has long been seen as weak and biased, so other ministries tend to ignore its decisions, and it has no real control over the finance and sectoral ministries—in fact, it has become subject to decisions by these other ministries. In Uganda, the now-defunct Decentralization Secretariat was originally seen as legitimate because, although it reported to the Ministry of Local Government, it was entirely staffed by non-civil-service hires largely detached from the government bureaucracy. In cases where the Secretariat clashed with the Ministry of Finance, Planning, and Economic Development, however, it almost invariably lost, and over time was marginalized. It was dismantled with the recommendation that a higher-level coordination body (under the Office of the President) be established, but this never happened. A second approach to facilitating coordination is to establish a formal interdepartmental body to oversee the development and implementation of intergovernmental fiscal reforms. Many interdepartmental mecha- 46 The Political Economy of Decentralization Reforms in Developing Countries nisms have not fared well, however, degenerating into forums in which representatives of different departments argue with each other, and even if they come to an agreement, often fail to implement it. In many cases, these bodies are chaired by ministries that suffer from the same legitimacy problems outlined above with respect to single-agency coordination mechanisms. To be effective, such bodies probably need to be chaired by a neutral party or have rotating leadership responsibility, and they must have strong, formal enforcement powers. This can be difficult to realize, however, since political or bureaucratic incentives may give the lead to a ministry irrespective of its suitability or legitimacy in the eyes of other actors. But without the right type of structure and a willing and able leadership, such mechanisms are likely to atrophy, a common fate with ad hoc interministerial coordinating committees that rely primarily on the voluntary cooperation of members. One example is Cambodia’s National Committee to Support the Com- munes, which was founded after the creation of elected commune councils in 2001 to manage the further design and implementation the commune system. Although it was interministerial, it was chaired by the Ministry of Interior (MOI) and perceived by other members to be largely under MOI control, and it gradually lost credibility with other key agencies. In addition, responsibility for deconcentration to the provincial administration was ini- tially placed under a separate interministerial body, the Council for Admin- istrative Reform, whose director was, like the Minister of Interior, a Deputy Prime Minister seeking to establish influence within the ruling party. As a result, coordination between the two bodies was weak. The responsibilities of the Council were transferred in 2006 to a coordinating mechanism chaired by the MOI, but they were not integrated with efforts to further develop the commune system. The current mechanism for managing all subnational reform, the National Committee to Support Democratic Development, was created in 2008, but it is still seen by other ministries as dominated by MOI.6 Indonesia has used various interdepartmental coordination mechanisms for purposes related to regional and local development dating back to the late 1980s. Although such bodies have been productive in some ways, they have suffered either from credibility problems associated with single-agency leadership or from too great a detachment from practical issues. Most of these bodies have been unwilling or unable to coordinate beyond general policy development, leaving the door open for individual ministries to behave as they wish during implementation. Current arrangements for coordination under the Directorate General for Regional Autonomy in the Ministry of Home Affairs are structured in a way that seems unlikely to improve performance. Understanding Bureaucratic Incentives and Behavior 47 A third approach involves the appointment of a substantially indepen- dent body to coordinate local government reform activities. Such an arrange- ment reinforces the neutrality of the coordination mechanism and helps raise it to some extent above interdepartmental politics. This type of mechanism is rare, however. India’s State Finance Commissions, which are appointed by each state government to oversee implementation of the constitutional amendment on municipal decentralization, approximate this arrangement. An independent body with more limited scope is Uganda’s Local Government Finance Commission, which has influence over local government revenues and transfers. As a legally mandated independent body that reports directly to the president, it can sometimes resolve fiscal decentralization disputes between the local government and finance ministries, but its role is funda- mentally advisory (Smoke 2008b). Another example of a more independent, higher-level body is the Na- tional Decentralization Committee in Thailand, which is the strategic policy development unit for decentralization. Its broad-based membership includes central and local authorities, as well as nongovernmental representatives. In addition to playing a key role in the design of decentralization, this body is charged with monitoring and implementing reform and, as appropriate, making further reform recommendations to the national cabinet. The Com- mittee developed a detailed implementation strategy, but reform momentum has been undermined by recent political events so that implementation has been less successful (World Bank 2005). Development Partner Coordination As discussed above, the failure of development partners to work coopera- tively on decentralization support and the tendency of their behavior to further complicate government agency coordination have posed major problems in developing countries. In response to extended criticism of the way they operate, donors have recently made some attempts to coordinate among themselves—both generally and with respect to decentralization. Various models have been used, including coordination committees, joint government-donor decentralization working groups, and, less commonly, separately managed mechanisms designed to provide incentives for donors to work more cooperatively (such as the Decentralization Support Facility in Indonesia). The extent to which these mechanisms have been effective depends, of course, on how they are structured, why they were implemented (whether a genuine desire to coordinate or a superficial attempt to respond to prevailing aid harmonization rhetoric), and the extent to which common ground can be 48 The Political Economy of Decentralization Reforms in Developing Countries negotiated in spite of the conflicting institutional incentives discussed above. In the best-case scenario, a reasonable degree of coordination is achieved, and it is even possible to promote better coordination among government agencies. At the other extreme, these mechanisms are highly ineffective. In some cases, the potential value of coordination on decentralization matters is undermined by other mechanisms for coordinating donor support, such as sectorwide approaches or broader public financial management reform initiatives. While it is not possible to provide a more detailed review of do- nor coordination on decentralization here, it is important for development partners to be aware of opportunities for and constraints to promoting such coordination, and to the extent possible, to encourage the creation of mechanisms that do a better job of accounting for the government and donor incentives and dynamics discussed above. Notes 1. See, for example, Tendler (1993; 1997). 2. See, for example, discussions of public financial management reform in Malawi (Rakner, et. al. 2004) and public sector revenue reform in Uganda (Fjeldstad 2005). 3. This material builds on Smoke (2010). 4. The experiences of China and Vietnam (this paragraph) and the Philippines (next paragraph) are reviewed in World Bank (2005). 5. Kenya’s system is discussed in Smoke (2003) and Kenya and Uganda are compared in Smoke (2008b). 6. Smoke and Morrison (2009) discuss the evolution of decentralization coordination mechanisms in Cambodia. 5 Understanding the Dynamism of Context and Incentives O ur goal in the previous two sections was not only to review the range of incentives that politicians and bureaucrats face when they consider decentralizing, but also to show how these political and bureaucratic incentives lead to different choices and behaviors vis-à-vis decentralization. In this section, we seek to incorporate greater dynamism into the analysis, beyond recognizing that poli- ticians tend to dominate the decision to initiate decentralization and that bureaucrats substantially influence its implementation. The generally highly dynamic nature of decentralization across its successive stages—initiation, design, implementation, and modifica- tion—poses both special challenges and potential opportunities for development partners who seek to support the process. In many cases, decentralization dynamics shift substantially after the initiation phase ends. Some initially supportive or non- obstructive agencies may reconsider their position when they come to understand the implications for their own power and resources. In some cases, agencies that did not really see themselves as hav- ing a stake in decentralization may enter the field when they see opportunities to benefit from involvement or look for ways to preserve their own roles with respect to local functions. As threats emerge from agencies that are seen as competitors, the original lead agencies (or subagencies) on decentralization may develop new strategies that are more focused on maintaining their role than on effective decentralization policy. Changes may also oc- cur in the dominant factors that initially spurred official national 49 50 The Political Economy of Decentralization Reforms in Developing Countries policy on decentralization, and new knowledge constantly emerges about how decentralization is unfolding. Given the shifting incentives that politicians and bureaucrats face as decentralization unfolds, development partners working in this area must remain flexible. We examine the dynamic quality of decentralization below, highlighting three key points. First, although the political science litera- ture has emphasized the rational, purposive behavior of politicians in the design phase of decentralization, like all actors operating with limited and imperfect information, politicians make mistakes. Despite their best efforts, decentralization may set into motion unintended consequences that donors must take seriously. Second, as politicians and bureaucrats pursue their careers, the institutional positions they occupy will naturally change over time. Individual officials who previously supported decentralization may withdraw their earlier support in line with their new incentives, and initially unsupportive officials may come to favor decentralization if they come to believe that it may provide benefits. With decentralization, as in many other areas of public policy and public sector reform, “where you sit is where you stand,” and where you wish to go matters, too. Third, as decentralization has aged—particularly in the developing countries that were “early adopters”— and conditions change or new information becomes available, struggles have unfolded not only over the implementation of initial frameworks, but also over attempts to redesign these frameworks. In some cases, there have been attempts to recentralize. Unintended Consequences According to the choice-theoretic approach that has dominated the recent political science literature on decentralization, national politicians choose to decentralize because they believe that it can, if designed appropriately, further their own interests. Decentralization can be risky, however, setting in motion changes that often differ from or go substantially beyond what national politicians had contemplated—and in ways that cannot simply be reversed. In Chile, decentralization in the 1980s and 1990s reflected the preferences of a national government that was committed to fiscal austerity and that consequently sought to decentralize administrative responsibili- ties without additional revenues. Subnational officials were virtually absent from the design process, which was dominated instead by national party leaders and by the Ministry of Finance, whose views about the importance of maintaining Chile’s hard-won fiscal stability were paramount. Over time, however, the decentralization of schools and hospitals increased the vis- ibility of mayors and triggered the creation of a powerful new association Understanding the Dynamism of Context and Incentives 51 of municipal governments. The national government now must regularly consult with this association, which has used its place at the table to lobby successfully for additional decentralizing measures (Eaton 2004b). As the Chilean case shows, even where design decisions appear to keep subnational officials on a tight leash, opportunities may emerge down the line for donors to support these officials as they mobilize for further changes. Politicians are rarely able to anticipate fully how other political actors will respond to decentralization. A powerful case in point is Bolivia, whose 1994 Law of Popular Participation initiated one of the world’s most in- novative decentralization experiences, including the creation of oversight committees for civil society actors in over 300 new municipalities. Scholars have advanced a number of explanations for why the governing (MNR) party supported decentralization, interpreting it as an effort to shore up the legitimacy of the state (Grindle 2000), to strengthen the municipal level of government where the party’s electoral prospects were brightest (O’Neill 2005), or to forestall more threatening demands for decentralization to the regional level of government (Roca 2008). More important than any of these intended consequences, however, is the fact that decentralization unwittingly generated a new and dire challenge not just to the MNR itself, but to all of Bolivia’s established parties through the emergence of a new indigenous party, the Movement Toward Socialism (MAS), whose victories in the municipalities of Cochabamba fueled the party’s meteoric rise to the highest levels of government, culminating in the 2005 presidential victory of Evo Morales (Van Cott 2008).1 The success of this party has in turn generated new pressures in Bolivia’s eastern half—the only part of the country where the MNR is still dominant—for decentralization to the regional level instead. The interaction of political calculation with structural and/or cir- cumstantial changes in a country’s political economy can also generate unanticipated consequences. In Peru in the 1990s, the national government established a system that shares royalties from natural resources with the regions that produce these resources (revenue canons). The system was de- signed largely to address grievances around environmental degradation in a few mining towns. However, following liberalizing changes in Peru’s foreign investment code and a dramatic spike in mineral prices after 2000, royalty payments to select subnational jurisdictions took on great fiscal significance. Because the national government never passed a fiscal decentralization law requiring that revenues be shared with all subnational governments regardless of their subsoil resources, the royalty system generated new forms of conflict within and between subnational governments. In Bosnia and Herzegovina, where some donors pushed for rules-based sharing of the value-added tax with municipalities, unexpectedly buoyant revenues have provided mu- 52 The Political Economy of Decentralization Reforms in Developing Countries nicipalities with a degree of autonomy beyond that implied by an initially technical reform. This suggests that development partners’ choices to support efforts on one front can have more far-reaching systematic consequences. Another unintended consequence of decentralization can be seen in the proliferation of subnational governments in several countries. Particularly where it takes the form of guaranteeing some minimal level of revenue trans- fers to all subnational governments, decentralization may create incentives to divide subnational jurisdictions into two or more units, each of which would be eligible for revenue transfers. In Indonesia, in the decade since the transition to democracy, the number of provinces has increased from 26 to 33 and the number of districts from 290 to 450, largely in response to vertical coalitions composed of politicians at the national, regional, and local levels (Kimura 2007). As Fitrani, Hofman, and Kaiser (2005, p. 66) show, the lump-sum nature of Indonesia’s general allocation grant means that “two new districts get effectively twice as much as the larger older district” from which they were formed. Several countries in Africa, including Kenya and Uganda (Smoke 2003; Smoke 2008b), have also substantially increased the number of local governments after decentralization. The phenomenon of proliferating subnational governments has been particularly prevalent in Latin America. In Brazil, generous constitutional re- forms in 1988 that mandated federal-municipal and state-municipal revenue sharing help account for a sharp increase in the number of municipalities over the last two decades. Across Latin American countries with sizable indig- enous populations, fiscal decentralization to municipalities has encouraged the organization of historically marginalized rural and indigenous com- munities that are now demanding their fair share of municipal resources relative to the more urbanized town centers that historically dominated municipal spending decisions (Cameron 2009; Fox 2007; Van Cott 2008). In some cases, indigenous communities are demanding that submunicipal units be separated out of existing municipalities as the only way to ensure that nonindigenous authorities do not continue to benefit disproportionately from public revenues. Thus, whereas scholars in the public administration literature have argued for the amalgamation of existing subnational governments, which are frequently too small and fiscally weak to provide important services ef- fectively, political incentives often point in the opposite direction. Peru is a case in which even stated intentions to amalgamate could not be realized. Donors provided technical assistance for a 2002 law that sought to merge the country’s 25 regional governments into a more viable set of larger units. Voters defeated the formation of these “macro-regions” in a 2005 plebiscite, which was characterized by heavy lobbying against amalgamation by elected Understanding the Dynamism of Context and Incentives 53 officials in existing regional governments. Since 2002, additional provinces and districts have been created despite the fact that many of them do not meet the minimum population requirements established in the 2002 law (PRODES 2007). Incentives to increase the number of governments—at both intermediate and local levels—appear to be widespread, and donors should anticipate how continued proliferation could affect their activities at the subnational level.2 More generally, decentralization leads to unintended consequences because the new institutional rules and procedures that it creates frequently overlay long-existing institutions and behaviors, which do not disappear as a result of decentralizing changes.3 Decentralization undeniably generates new institutional rules governing the distribution of resources and responsi- bilities between levels of government, but these new institutions must often coexist with earlier institutional practices and incentives. In the Philippines, short-term electoral pressures and presidential politics forced legislators to decentralize fiscal authority in 1991, despite the risks that fiscal transfers would pose to their time-honored roles as brokers of governmental spending. Legislators have not reversed decentralization, but the institutional incentives they face have not changed; if anything, decentralization has reinforced their dependence on constituency funds allocated by the national government, for which they can individually claim credit and which enable them to compete politically with newly empowered subnational governments. As a result, the institutional incentives that produce pork-barrel dynamics now coexist with institutional rules that give subnational governments a much greater share of tax revenues. Reformers did not perhaps intend decentralization to produce such irrational public budgeting practices and the complicated accountability relationships they foster, but these unintended consequences do reflect a rational political logic. Where You Sit Is Where You Stand Earlier, we argued that development partners should focus on the nature of the incentives facing politicians and bureaucrats, rather than seeking to gauge the extent of their political will to decentralize. One way to demon- strate the importance of political incentives relative to political will is to examine cases in which political actors’ attitudes toward decentralization change in line with the new institutional positions they assume. While it is possible to make some generalizations about the different incentives fac- ing presidents versus legislators, Ministries of Finance versus Ministries of Local Government, and national versus subnational party leaders—as we did in Sections 3 and 4—individual politicians and bureaucrats do not 54 The Political Economy of Decentralization Reforms in Developing Countries stand still. As these actors occupy different positions within the political and bureaucratic system, their stances on decentralization are likely to change in ways that reflect the incentives they face in their new position. Reform trajectories can be illuminated by the simple insight that where actors stand on decentralization is a function of where they sit, institutionally speaking. And it is reasonable to assume that the positions taken by individual actors on decentralization reflect the positions for which they have future ambi- tions. It is important to note that these insights have potentially significant implications for donor activities. Once a political party succeeds in taking power, for example, party leaders can renege on the commitment to decentralize (overtly or tacitly) if such a policy no longer serves their interests and they think they can get away with it. The emergence of a multiparty environment or changes in the balance of power among existing parties can lead actors to reposition their stances on decentralization (either pre- or post-election) and trigger a reconfiguration of the institutional landscape. This has clearly been the case in Nicaragua, where the Sandinistas worked in partnership with the Association of Nica- raguan Municipalities to push decentralization legislation when they were in the opposition; after winning the 2006 presidential election, local govern- ment advocates charged the new government with trying to undermine the municipalities. In Uruguay, the electoral and institutional dominance of the Colorado party encouraged the Blanco party to incorporate decentralization into its party platform—a position it failed to act on when it controlled the presidency between 1990 and 1995. In South Africa, the stance of the African National Congress with respect to decentralization has been similarly inconsistent through time. As noted above, at the moment of institutional design, the ANC agreed to decentralizing changes that would enhance the governing role of provinces in order to shore up the then-governing National Party’s support for a ne- gotiated transition to democracy. Since then, as the ANC has consolidated its control of the national government, the party has reneged in a variety of ways on its earlier support for decentralization. Undermining provincial governments has appealed to ANC party leaders as a means to undercut the provincial bases of opposition parties, as well as to weaken internal party rivals (Dickovick 2007). The ANC has dramatically reduced the number of local governments, from nearly 900 to 238, in a process of amalgamation that stands in stark contrast to the proliferation of subnational governments in so many other cases of decentralization. Although the stated reason for amalgamation followed the standard logic of creating more fiscally viable local governments, some analysts have attributed it to the ANC’s desire to deal with fewer local governments.4 Understanding the Dynamism of Context and Incentives 55 Prominent individual politicians and bureaucrats in key positions of influence can also change their minds about decentralization. In Peru, Ale- jandro Toledo announced an ambitious program of decentralization on the campaign trail for the presidency in 2000. When opposition candidates won the majority of regional governments in the elections that he introduced, Toledo’s enthusiasm for further fiscal and administrative decentralization quickly dissipated. Individual career interests and trajectories also play an important role within the bureaucracy, particularly with respect to high-level bureaucrats. In Uganda, the permanent secretary of the Ministry of Local Government, one of the architects of the national decentralization policy in the 1990s, has begun to promote greater centralization since becoming a permanent secretary in the Ministry of Education. The effect of dual roles on individual behavior can be important as well. Some of the key decentralization actors in Cambodia’s Ministry of Interior are also heavily involved in the ruling party and in the elections process, and thus face simultaneous pressures to defend subnational and national government prerogatives in struggles over decentralization. Movement in the opposite direction is also possible. Political actors who were initially hesitant about decentralization convert to the cause as political and electoral dynamics shift and as older political concerns are displaced by new challenges. In Bolivia in the 1990s, fear that decentraliza- tion to the regional level would generate centrifugal pressures and separatist movements led the governing MNR party to instead endorse municipal decentralization. When the emergence of a new party (MAS) relegated the MNR to the sidelines as a party that retains regionally concentrated support in the eastern lowlands, the MNR quickly became a vociferous advocate of political, fiscal, and administrative autonomy for the regions. In Chile, the transition to democracy in 1990 led parties to the right of center to fear that they would no longer be able to win control of the national government— fears that for two decades were well founded. These parties thus shifted from opposing political decentralization during the Pinochet dictatorship to sup- porting the introduction of elections in the thirteen regions that Pinochet created, some of which have been safe districts for the right ever since. In countries that opted to devolve significant resources and responsibilities, such as the Philippines, some legislators who were initially wary of losing authority became more aggressive advocates of decentralization after taking subnational office. In order to make sense of where individual politicians or bureaucrats “stand” on decentralization, it is important to pay attention not just to where they currently sit, but also where they aspire to sit in the future and 56 The Political Economy of Decentralization Reforms in Developing Countries where they sat in the past. With respect to the future, theories of progres- sive ambition encourage us to take common career paths seriously, along with cross-national variation in the strategies that politicians use to build their political careers. In Brazil, progressive ambition tends to lead national legislators back to their home states after their legislative terms end, rather than into elected or appointed positions in the federal government (Samuels 2003). As a result, governors often have significant influence in the design and defense of decentralization, and national legislators commonly view decentralizing measures through the lens of how it might affect their future perches in states and municipalities. In other cases, the return to one’s home district after a stint in the national legislature would be understood as a clear step backward; progressive ambition in these environments is likely to discourage legislators from offering strategic support for decentralization. With respect to past experiences, politicians and bureaucrats bring knowledge acquired in prior roles with them as they move through the politi- cal system. For example, national legislators and appointed bureaucrats who previously served as mayors and municipal councilors or city managers and municipal employees can bring valuable insights and pragmatic positions into the debate over decentralization. One significant example is Philippine Senator Aquilino Pimentel, whose authorship of the 1991 Local Government Code drew heavily on his previous experience as mayor of Cagayan de Oro, where he confronted the frustrations of overweening central government control. Pressures to Modify Initial Frameworks In recognizing that decentralization is a multistage and often lengthy and uneven process permeated by politics, development partners should not expect initial frameworks to be implemented just as politicians intended. As discussed in Section 4, bureaucrats may have significant scope for discretion and at times clear incentives to thwart the designs that politicians have em- braced. The initial frameworks themselves are often revisited and modified, sometimes culminating in recentralization and sometimes in changes that further expand the authority of subnational governments. For example, if a crisis stimulated the initial decision to decentralize, the waning urgency of the crisis can stall or undermine the decentralization agenda. This has been a major theme in Southeast Asia, including to some extent in Indonesia, the Philippines, and Thailand (World Bank 2005). The emergence of a new crisis can invigorate weak attempts to decentralize or create an impetus to recentralize. Following financial crises in Argentina and Brazil in the 1990s, both countries imposed new borrowing and fiscal accountability frameworks on subnational governments, partially undoing Understanding the Dynamism of Context and Incentives 57 decentralizing changes legislated in the 1980s (Dillinger and Webb 1999). In the late 1990s, Colombia’s worsening internal security crisis and challeng- ing economic environment dampened support for decentralization, lead- ing to cuts in revenue transfers and a return to appointed mayors in select municipalities. Finally, where decentralization has emerged as an essential part of the bargaining process around newly democratic regimes, the end of the transition can lead to modifications in transition-era frameworks. This has been the case in Brazil, the Philippines, South Africa, and Uganda, to name a few. Another source of pressure to alter initial frameworks can come from political decentralization itself. Where subnational elections have been intro- duced, the new independently elected subnational governments—which did not exist when earlier decisions on fiscal and administrative decentralization were made—may demand a say in the distribution of resources and respon- sibilities. Political decentralization brings into the game new actors who no longer serve at the pleasure of national officials, who may not be happy with the initial framework that national officials adopted, and who can use their new political stature to lobby for change. Acting individually, mayors and governors seldom have the political clout to bring about substantial modi- fications in the overarching decentralization framework, and the complaints of politically troublesome mayors and governors can be accommodated by special treatment from the center on an individual basis. In some cases, however, leagues and associations of subnational governments, which enable mayors and governors to act collectively (often with donor support), have been able to enhance the leverage of disparate subnational officials. Emerging evidence on the performance of decentralization can also generate (productive or problematic) efforts to strengthen or weaken the initial framework. Decentralization often triggers a series of framing struggles between advocates who trumpet the local champions that have made great use of decentralizing reforms, and detractors who point to the local governments that have squandered their opportunities. Which side wins depends on the evidence that emerges and the way in which those with power process this evidence. In Uganda, hard data on problematic resource use by local govern- ments and on weak service delivery outcomes, although limited primarily to a few sectors, provided plenty of ammunition for central government agencies that wanted to constrain local autonomy. These agencies received substantial backing from development partners who supported centralizing public financial management reforms in the Ministry of Finance, Planning, and Economic Development and sectorwide approaches in a number of the key sectoral ministries. The government’s revised fiscal decentralization strategy, issued in 2002, considerably restricted local government budget- 58 The Political Economy of Decentralization Reforms in Developing Countries ing autonomy, which had been significant in the original decentralization legislation and reform program (Smoke 2008b). In Indonesia, similar concerns about substandard local financial man- agement performance and the creation by local governments of problematic revenue sources led the authorities to pass some partially recentralizing leg- islation in 2005.5 In the Philippines, the apparent decline in service quality in several government-run hospitals following decentralization prompted their transfer back to national government control. In Latin America, grow- ing evidence of irresponsible fiscal behavior in subnational governments drove recentralizing changes in Argentina, Brazil, and Colombia. On the other hand, the early success (in certain respects) of the nascent commune decentralization in Cambodia created a political window to push for broader decentralization to district and provincial levels. 6 Attention to the same set of incentives that politicians face when they de- cide to adopt decentralization can help explain the outcome of later struggles over recentralization. For example, when electoral incentives dominate and incumbent politicians endorse decentralization simply because they fear their party will lose power in the next election—or because a different party happens to control the presidency—subsequent governments may have little trouble reversing decentralization. When politicians’ short-term calculations dominate and decentralization follows the electoral cycle, decentralizing changes may not be fully implemented or may not be in force for long enough to produce local stakeholders who can defend these changes in the period before pressure builds for recentralization. If decentralization occurs as the result of intra- party dynamics, as where subnational party leaders wield significant influence within parties and force decentralization from below, advocates of recentral- ization are likely to encounter more challenging obstacles. Where multiple institutional actors—such as separately elected legislatures and independent courts—participate in decentralization design and/or the enforcement of rules affecting subnational governments, supporters of recentralization may likewise be checked. The absence of such institutional veto players, on the other hand, is likely to clear the way for recentralizing proposals should these emerge. Finally, where the initial move toward decentralization is the result of deep coalitional changes and societal pressures from below, politicians at the center may have a difficult time moving the balance of power back toward the center. Recentralization may be particularly challenging in post-conflict settings where decentralization was adopted to cement new coalitions among former combatants. In all of these hypothetical cases, appreciating electoral, partisan, institutional, and coalitional incentives should help development partners understand not only the initial move toward decentralization but also subsequent efforts to change or reverse course. Understanding the Dynamism of Context and Incentives 59 Notes 1. In a nearly opposite but equally unintended dynamic, reformers in Colombia adopted decentralization in the hopes that it would facilitate the entry of new political ac- tors who could challenge the hegemony of the Conservative and Liberal parties. In practice, members from these two traditional parties have been able to win and hold control over most of these newly empowered subnational governments. 2. Note that, while the proliferation of subnational governments is often the unin- tended result of decentralization, in some cases it corresponds to a more purposive logic. In the post-conflict Democratic Republic of Congo, a new constitution in 2006 increased the number of provincial governments to 26 in a move that reversed Mobutu’s earlier decision to decrease the number of provinces from 23 to 11 in a bid to centralize power. 3. We are grateful to Lily Tsai for raising this point. 4. Moreover, there was a push in the Parliament to bolster the power of the districts, each of which contain a number of municipal governments, perhaps as a way of further reducing the number of more highly empowered local governments. 5. An additional factor here was the emerging realization that the original local-centric decentralization unnecessarily marginalized the once-effective provinces. 6. Nevertheless, decentralization advocates have questioned the degree to which the anticipated expansion would represent genuine decentralization or instead further strengthen the ruling party and undermine the modest but reasonably autonomous commune-level decentralization already in place. 6 Applied Political Economy of Decentralization Diagnostics T he premise of this volume is that a deeper understanding of political economy factors that shape decentralization in a given country would enhance development partners’ ability to provide support that contributes to desired development results, such as improved resource management, better service delivery, or stronger state-society relations. Development partners may differ in their perspectives on what priorities and objectives are most important. For example, strengthening local democratic accountability may be deemed by certain partners or in some situations to be more critical than immediate improvements in service delivery. However, politi- cal economy of decentralization (PED) analysis is relevant for any priority and objective. It also applies to the behavior of development partners themselves—for example, how their support relates to a candid analysis of country contexts and goals and not just their own institutional priorities and resource allocation time horizons. In recent years, development partners—including the World Bank—have increasingly conducted governance assessments de- signed to identify the challenges and opportunities associated with attaining normatively desirable decentralization reforms. These assessments complement a range of typically technical analyses around, for example, fiscal and institutional aspects of reform. However, they rarely go into sufficient depth in evaluating: (a) whether the actors whose cooperation is needed to define and implement decentralization face appropriate incentives and have the capacity to support reform; (b) whether these often-diverse ac- 61 62 The Political Economy of Decentralization Reforms in Developing Countries tors—including other development partners—are likely to work in harmony or at cross-purposes; and (c) the implications of (a) and (b) for whether and how a development partner might provide productive support with sustainable results. These gaps and available evidence about reform and program perfor- mance suggest that development partners have provided significant support for decentralization and local government reform efforts that they have not fully understood. In its Independent Evaluation Group review of decentral- ization interventions, the World Bank (2008) emphasizes the need to establish clear political ownership prior to engagement. Unfortunately, the ownership criterion is often judged to be satisfied if a national decentralization policy or strategy has been adopted. Our central point is that ownership and “po- litical will” are considerably more complex, and that development partners need to look more carefully at realities on the ground before engaging, and in particular before deciding how best to engage. As discussed in Section 1, the results of decentralization reforms have been limited relative to the substantial amount of resources devoted to supporting them. Throughout this volume we have tried to show how in- sufficient understanding of political and institutional dynamics may have contributed to disappointing or unexpected performance. Thus, it seems evident that those development partners who wish to make better use of resources to support decentralization reforms may benefit from devoting more attention to these issues. The recommendation for more systematic political economy assessment is not made lightly, as we recognize that analytical work typically undertaken in programming development partner support for decentralization already imposes a heavy burden. Certain political economy issues do need universal attention, and in some cases the additional analysis will involve considerable effort. It would often be possible, however, to approach the analysis in ways that are not unduly onerous yet provide useful information. We must also emphasize again that judgment is central in terms of both how to approach the PED analysis and how to use it. Just as there are no nor- mative universal rules for the analysis, there are no universal prescriptions for how the insights it provides should determine the design and implementation of programs. The point of the PED analysis is to provide more complete and better information about issues that are often only partially considered, not to yield definitive advice on how to proceed based on a certain set of findings. PED analysis is both science and art, and those who undertake it must still be prepared to take difficult, if better informed, decisions. It is important to re-emphasize here that donors are themselves political actors in the arena of decentralization. Particularly in aid-dependent settings, Applied Political Economy of Decentralization Diagnostics 63 their behavior may have major consequences, both productive and problem- atic. Even if their objectives are explicitly developmental, their actions will be political in some respects—for example, the choice of counterparts with whom they work at the national and subnational levels of government can generate or reinforce positive or negative dynamics and outcomes. The remainder of this section considers how better PED analysis can be realized by development partners in practice, and what this might imply for actual strategies and interventions on the ground. First, we highlight a number of considerations that need to be taken into account when applying insights from PED analysis for operational purposes. Second, we present two types of stylized PED analysis to give readers (and prospective task manag- ers) a concrete sense of the typical steps and resources this analysis would require in practice. Third, we underscore the importance of validating and following up on PED inputs. Finally, we highlight the implications that PED analysis can have for country engagement; while the analysis may suggest that certain desired actions are not feasible, it should also generally reveal potential entry points for creative and realistic engagement. Expectations for Political Economy Assessments As mentioned above, the proposed PED approach does not presume that decentralization is inherently desirable or undesirable, either in general or in some particular form. Rather, this broad analytical framework is intended to be relevant to the analysis of any type of decentralization reform. In practice, decentralization is multifaceted and complex, and care must be taken to avoid seeing it as a unidimensional process or an all-or-nothing phenomenon. In any given setting, decentralization reforms must in the first instance be seen as the means to an end, whether improved service delivery or enhanced political participation. The specific policies chosen will be contingent on historical setting, development objectives, and subnational capacity, among others, but political perspectives matter as well. What one actor sees as decentral- ization (for example, central transfers to communities or improved central monitoring of subnational finances) may to others seem centralizing. The degree and nature of decentralization are also likely to change over time as circumstances and perspectives evolve. Ideally, debates concerning preferred reform strategies will be based on empirical evidence about whether decentralization is good or bad for service delivery, governance, and poverty reduction. From the perspective of operational engagement, however, this may not be the most productive entry point and is rarely by itself a sufficient basis for sound decisions. There remain significant empirical challenges in conducting impact evaluations 64 The Political Economy of Decentralization Reforms in Developing Countries of decentralization (Kaiser 2006; Ahmad 2009). Studies have yielded highly inconsistent findings, and they must be interpreted in the context of nu- merous factors and overlapping policy measures that are rarely considered. Decentralization is only one type of policy reform available to achieve a particular set of objectives, and it often requires major institutional and procedural changes that cannot all be implemented at once. Systems and capacity typically need to be developed over time, and other factors may evolve during this process. Because decentralization (and centralization) are pursued for a range of mixed and often-changing motives, the challenge for development partners is to assess not only what initiatives would be politically feasible at some point in time, but also whether they have the potential within the political context to bring about sustainable results over the medium to long term. One potential litmus test for assessing whether a government is truly concerned with outcomes is the extent to which it has adopted mechanisms to monitor results regularly and whether it uses this information to modify the decentralization framework accordingly. At the same time, the motivations for and use of information about decentralization need to be appreciated in a political context; different government agencies and donors generate and leverage information about decentralization in ways that support their own interests. Understanding why specific types of evidence on decentralization are or are not generated, when this evidence is gathered, and whether these are one-time or ongoing efforts is important for understanding the decen- tralization trajectory in any given setting. From an operational standpoint, the overarching concern of PED analy- sis is to inform decisions about whether and how to support decentralization efforts in a particular country. If a development partner decides to engage on decentralization, decisions must be made about the most appropriate scope and scale of interventions, the advantages and disadvantages of work- ing with alternative partners, and how these efforts relate to other efforts and programs.1 Political economy analysis might suggest, for example, a more modest effort than originally foreseen or new opportunities to link decentralization productively to sectoral or community-driven develop- ment activities. PED analysis could also suggest that engagement with the Ministry of Finance might be more productive than engagement with only the Ministry of Local Government, or that focusing on subnational entities may be more useful in making progress on the ground than engaging in national policy dialogue. Assessments of the value of various aid modalities and time horizons—including decisions about whether to provide technical assistance, support investment, or pursue development policy lending—will flow from these larger decisions about the nature of interventions. Applied Political Economy of Decentralization Diagnostics 65 In cases where development partners conducting PED analysis find their own objectives at odds with those of a particular country, we are not sug- gesting that their independent objectives should never be pursued. External support that is inconsistent with formal government decentralization policy may under some circumstances be able to stimulate experimentation and dialogue, and perhaps help to promote better systems and outcomes over time. To undertake such efforts effectively, however, development partners still need to understand the basic political and bureaucratic dynamics un- derlying decentralization in that particular country context. Much is often made of the question of how much leverage development partners have in any given setting, and to what use they put this individual and collective leverage. Arguably, leverage will be higher in aid-dependent settings. But even where development partners are able to more effectively exert pressure on policy and program design, the political economy dynamics that evolve during implementation can trump the decentralization initiative’s original official intentions. In settings where development assistance is less important, the manner in which donors contribute ideas to prevailing debates may still have a significant impact. For example, large-scale conditional cash transfer programs have gained significant traction in a number of middle- income countries. Arguably, these programs have been driven not by financial concerns, but by the diffusion of international ideas and experiences in the context of receptive domestic environments. In promoting PED analysis, we do not in any way denigrate the critical role of good empirical and technical analysis, which provides critical under- standing of what does and does not work, and which structural and proce- dural problems in the current system need attention. Diagnostic work should ensure that development partners are able to calibrate not only the overall level of financial support, but also—and most importantly—the design and implementation of their support, including the choice of aid modalities and partners and how these might evolve over time. Those tasked with commis- sioning PED analysis will need to be confident about its prospective value added relative to the informal political economy analysis often undertaken, as well as the feasibility of mounting a timely and cost-effective effort. Commissioning Country-Specific PED Analysis While the academic literature reviewed in this volume can provide a useful orientation for practitioners seeking to embark on decentralization support, it cannot substitute for applied, contextualized analysis. Here we present some general guidance for commissioning pragmatic country-level political economy analysis and applying it for operational engagement. The approach 66 The Political Economy of Decentralization Reforms in Developing Countries set out here draws on the authors’ own experiences in engaging with different aspects of decentralization reform, and on a limited number of case studies to which the approach was informally applied.2 Although preliminary, these case studies have sought to demonstrate both the tractability of preparing a PED report and the types of insights that development partners can hope to gain. Task managers can use the guidelines provided here in simple or comprehensive ways, depending on the analytical needs in a given case. Based on observations and pilot work across a range of settings, we iden- tify two main types of PED analysis: (a) a broad-based Political Economy of Decentralization Country Assessment; and (b) a “problem-driven” Political Economy of Development Issue Analysis that drills down into specific top- ics.3 The first type of analysis illuminates the overarching political economy context for pursuing decentralization in a given country. The second type focuses on narrower issues that are priorities for development partners and their government counterparts. These issues are likely to be motivated by some consensus about desirable decentralization-related reforms, such as improved service delivery and poverty reduction, but framed in the context of what is politically and institutionally feasible. Issue analysis may focus, for example, on the specific steps needed to decentralize service responsibilities in particular sectors (such as health or water), to assign own-source revenues (such as property taxes or user charges), to create rule-based intergovern- mental transfers, to determine the balance between subnational government autonomy and central control, and to develop systems for subnational gov- ernment monitoring and financial control. It is important to reiterate that we have not focused in this volume on local-level political and bureaucratic dynamics beyond ad hoc consideration of how certain elements interact with the national level in shaping the over- all intergovernmental system, but we recognize that they can be extremely important. Understanding actors’ positions, attitudes, and opportunities at both national and local levels may suggest that development partners could be more influential by focusing on subnational or nongovernmental interven- tions that help make local actors more effective on the ground where the scope for meaningful central-level interventions is limited. A better-developed framework for helping development partners to more accurately diagnose the subnational and nongovernmental situation and opportunities would be a useful addition to their analytical tool kit in the future. Political Economy of Decentralization Country Assessment The Political Economy of Decentralization Country Assessment (PEDCA) is intended to provide an overview of the political and institutional dynamics Applied Political Economy of Decentralization Diagnostics 67 covered in this volume and an assessment of their broad relevance for de- centralization reform in a particular country. This analysis should aim to lay out in an accessible manner the basic context of a country’s decentralization trajectory, its political and bureaucratic drivers, and the shifting context and incentives that are relevant to decentralization reforms. Depending on the circumstances of decentralization, the types and quality of analysis already conducted, and available resources, development partners can consider an appropriate methodology and expectations for the PEDCA analysis. The PEDCA will generally conclude with possible entry points for supporting aspects of reform with reasonable potential to both align with normative goals of decentralization, such as service delivery and poverty reduction, and account for on-the-ground political and institutional realities that govern the feasible scope and pace of reform. In countries where decentralization and local government issues are expected to be central to public sector reform or are being contemplated by government counterparts, it may be worthwhile to attach the PEDCA exercise to the larger process of preparing a development partner country approach, such as the World Bank’s Country Assistance Strategy. Doing this would facilitate the preparation of better decentralization support, including through an explicit analysis of what other development partners are doing in this area. In addition, a properly prepared PEDCA would explicitly illuminate the linkages between decentralization and other overarching public sector reform efforts, such as public financial management and civil service reform. It could also highlight potential complementarities and conflicts with work being undertaken by sectoral departments or through community-driven development initiatives that aim to improve service delivery and account- ability. In formulating decentralization programs, these linkages would be considered not only in technical terms, but also with respect to the political and institutional actors and dynamics associated with each of the relevant activities. A PEDCA should be able to review the most important aspects of the political and institutional context and dynamics in eight to ten thousand words (15–20 single-spaced pages). The study would ideally cover the main issues raised in the framework elaborated in this volume. A typical outline would include: (i) an introduction; (ii) a background section that places decentralization in historical perspective; (iii) a discussion of the design and adoption of decentralization in the country being analyzed, including a thorough review of political and bureaucratic incentives and behavior, and the role of development partners; (iv) a discussion of shifting conditions and incentives and how these have affected or could affect the reform trajectory; and (v) a conclusion that summarizes key points and their relevance for 68 The Political Economy of Decentralization Reforms in Developing Countries whether and how to support decentralization in the country under consid- eration. More detailed information can be provided in annexes, as necessary. (A more complete scope and guiding questions for conducting a PEDCA are outlined in Annex 1.) Depending on available background information and resource persons, this analysis can be undertaken by qualified consultants with perhaps two to three weeks of effort. Given the time needed for preparation of Terms of Reference, identification and contracting of consultants, and review and quality assurance processes, a period of closer to two to three months will typically be needed to complete these types of reports. Depending on the country setting, the work could draw on international experts with significant specialization in the country, domestic experts with proven international writing skills, and/or mixed teams of international and domestic experts. Political Economy of Decentralization Issue Analysis Although it would be possible to conduct a political economy assessment of a specific aspect of decentralization without having done a PEDCA, the danger of doing so is that critically important broader dynamics and larger connections may become lost in the more targeted analysis, possibly leading to narrow programs that could run into problems generated by unconsidered factors. Thus, we would generally recommend that task teams wishing to engage in more targeted political economy analysis of particular decentralization issues first build on knowledge generated from the overall decentralization context analysis provided in the PEDCA. The more targeted Political Economy of Decentralization Issue Analysis (PEDIA) may be particularly relevant if development partner programs are seeking to support a specific initiative as set out by government priori- ties or as suggested by development partner priorities. For example, while government laws or policies may formally decentralize specific functions or revenues, this is often done, as documented throughout this volume, with- out sufficient analysis of how well a particular reform is likely to align with the political economy incentives of actors that determine the feasibility of implementation and the sustainability of reforms. A critical challenge for a robust PEDIA is to ensure that intended reform objectives are clearly framed and to concretely identify measures that are anticipated to achieve them. The PEDIA cannot substitute for good techni- cal analysis, and it must ideally be linked to such analytical work. It should serve primarily to inform whether normatively desirable reform measures are politically and bureaucratically feasible and/or to highlight the conditions under which they would be. Even if it is not reasonable to expect “first-best” Applied Political Economy of Decentralization Diagnostics 69 reforms, the PEDIA could stimulate dialogue among development partners, government counterparts, and other civil society stakeholders as to which alternative measures are likely to be more politically feasible and suggest how to begin to move the system toward more desirable behaviors and outcomes. The analysis could also identify the most promising entry points for engage- ment, which would take advantage of reform-minded actors, supportive coalitions, or existing capacities. As with the PEDCA, a PEDIA would ideally be succinct (no more than 20 single-spaced pages or ten thousand words). The scope of a PEDIA will vary, so the length will depend on what needs to be covered for a particular issue in a specific case. The recommended components would include: (i) an introductory statement of the issue of concern; (ii) articulation of how this issue relates to a particular development objective; (iii) a description of the methodological approach; (iv) a summary of the main stakeholders and institutional arenas pertinent to this issue, regardless of methodological approach; (iv) a discussion of interdependencies with other issues, includ- ing the wider decentralization trajectory (ideally drawing on a previously prepared PEDCA); and, finally, (v) a summary of implications for develop- ment partners. The PEDIA could be supported by expertise similar to that required for the more general PEDCA analysis, although specific technical skills related to the type of reform being contemplated may be required. Review and Dissemination A report prepared by consultants cannot constitute the final word on how the political economy of decentralization might affect future development partner support in a country. Although the experts preparing such an analysis would be expected to consult with key actors in government agencies, development partner organizations, and civil society, the resulting report would be need to be vetted to ensure broad acceptability of the content and options presented— for example, through a workshop or workshops held in the country under consideration, if it is likely that the stakeholders to be consulted will work productively in a joint meeting. In some cases, smaller workshops with more focused groups of participants might be appropriate. Certainly in the initial phases of this work, development partners should emphasize that the analysis is simply one of several contributions to enrich the debate. It would be useful to underscore that the assessment reflects the views of a particular analyst, often from local or international academic circles, and that the purpose of the workshop is to constructively critique and build on this analysis. The vetting process chosen would need to incorporate two objectives: (i) review and validation of findings; and (ii) assessment of the operational 70 The Political Economy of Decentralization Reforms in Developing Countries implications for development partners. Depending on the context and the quality of initial reports, this process may need to be conducted in phases, beginning, for example, with internal working meetings followed by external, multistakeholder events. Sufficient time would therefore need to be allocated to organize the process and, in tandem, to allow consultant deliverables to be finalized on the basis of feedback from the country team and other stakeholders. While the authors of the PEDCA and PEDIA reports can make recommendations about what development partners could and should do given the prevailing political economy dynamics, these views cannot substi- tute for genuine internalization of the implications of the PED analysis on the part of development partners (and relevant counterparts). Intellectually, it would be important to segment the political economy analysis from the (normative) assessment of what development partners should ideally do, and/or to clearly illustrate how and why one follows from the other. It would also be important to encourage the political economy analysis to put options on the table, even under circumstances that, at least in the short run, appear to largely preclude the most desirable reforms. The first objective of the follow-up to the initial drafting of the report would be to validate the consultants’ interpretations of the prevailing political economy dynamics around the decentralization trajectory under discussion. If there are major points of dissension after an initial round of comments, these should be noted for discussion. The preferred inputs on these initial drafts would depend on the country context. In some settings, contributions would be limited to academic and think tank counterparts. Ideally, however, it would be desirable to elicit comments from a broader group, including gov- ernment counterparts (executive and legislative, national and subnational). The ability to achieve this would depend on the potential sensitivity of the topics under discussion given the prevailing country context (for example, major political milestones or a timeframe prior to elections). Those with significant country familiarity would be in the best position to determine the optimal validation process for any given setting. Depending on the country context, the sponsoring agency would then need to decide how best to present and disseminate the findings of the analysis after all revisions have been incorporated. This would include the translation of key parts of the analysis into local languages and outreach beyond the “usual suspects” in the capital. Given the frequent turnover of development partner representatives in country, broader dissemination would be critical to ensuring some degree of institutional memory around these issues. Go- ing the extra distance to ensure the publication of findings (even if only as a web-based resource) would also help enhance the impact of this analysis over time. Typically these will not need to be formal reports published by Applied Political Economy of Decentralization Diagnostics 71 the development partners; they can instead be put in the public domain as academic or independent policy think tank working papers. The final objective of the vetting process would be to draw out implica- tions for donor engagement. The final sections of the PEDCA and PEDIA should provide a significant basis for this discussion (Annex 1), but actual implications would need to be determined by development partners them- selves. Are partners working harmoniously or at cross-purposes with gov- ernment counterparts and each other? Are they pursuing feasible goals and/ or promoting viable mechanisms to achieve these goals? Are the motivating incentives of the government consistent with the development objectives of donors? Is there genuine political motivation (whatever the ultimate drivers) for advancing proposed reforms? If so, what aspects of reform seem most strategic and urgent? For example, is there a need for framework develop- ment, creation of new systems and procedures or modification of existing ones, or support for more effective connections between subnational gov- ernments and citizens? If there does not seem to be a genuine motivation for reform with which development partners feel comfortable, does it make sense to disengage, to think about supporting alternative reforms, or to work behind the scenes in a limited way on activities that might influence policy development later on? The answers to these operational questions will not always be imme- diately obvious or undisputed following the completion of initial political economy assessments. Development partners will need to iterate and become more comfortable with this type of analysis over time. Assessments will also need to be revisited periodically, particularly if there is evidence of relevant existing or impending political shifts. A more disciplined and systematic approach to asking some of the key questions, including about traditional ways of approaching issues and modes of engagement with particular coun- terparts, would increase the likelihood of making more informed choices, which would in turn help generate decentralization investments with more significant development impacts. From Analysis to Action At its essence, the value added by PED analysis will be determined by the extent to which it informs development partners’ prioritization and sequenc- ing of their decentralization engagement. Good analysis must be able to generate tangible operational insights for task teams, country management units, headquarters, and, ideally, government counterparts—and to effectively communicate these insights. PED analysis should generate concrete options, rather than focusing exclusively on identifying obstacles. Good PED analysis 72 The Political Economy of Decentralization Reforms in Developing Countries must also show that options and recommendations flow from context-specific analysis, in ways that “standard” technical analysis (or normative precepts) would not generate. While we reiterate that the purpose of PED analysis is to generate a more complete and nuanced understanding of the environ- ment in which a development partner wishes to intervene, it will rarely lead to unambiguous clarity about exactly what to do, and with which actors or over what time frame to engage. Such analysis will, however, almost invari- ably uncover insights that are better to gain in advance than in hindsight. Failing to conduct at least a minimum of political economy analysis on decentralization raises considerable risks. Evaluation reports of donor engagement around decentralization, such as those prepared by the World Bank’s Independent Evaluation Group (World Bank 2008), highlight the need to pay greater attention to the political context in order to enhance the effectiveness of decentralization reform. Although the PEDCA and PEDIA are new instruments and documentation of their specific value is not available, there are examples of how attention (or lack thereof) to political economy factors in the formulation of development partner decentralization support activities likely influenced the effectiveness of these activities. Uganda provides a powerful example of a case where a better balance of political and technical analysis could potentially have averted some of the current concerns about the sustainability of the decentralization process. There were initially strong political drivers to proceed with decentralization in Uganda, and development partners took the ensuing robust decentralization framework seriously as an indicator of political will. As a result, the develop- ment partner programs that were created to support the implementation of this framework, including the World Bank’s multitranche Local Government Development Program, largely reinforced overly ambitious reform targets that overwhelmed governance and technical capacity at the local level and threatened the powers and prerogatives of powerful central actors. Most of the development partners supporting decentralization interacted primarily with the relatively weak Ministry of Local Government and virtually ignored more powerful players. Many of these players eventually took actions that many analysts see as having weakened local governments considerably, often with support from development partners engaged in other public sector reforms, such as public financial management and service delivery restructuring. Had more effort been put earlier into better understanding the political drivers behind decentralization, and into working cooperatively across and within a broader range of government agencies and with a more diverse set of development partners, there might have been less decentralization early on, but perhaps more gradual, effectively coordinated, and sustainable decen- tralization over time. Even if such cooperation could not have been secured, Applied Political Economy of Decentralization Diagnostics 73 development partners working specifically on decentralization might have taken a different approach to supporting reform. Even efforts explicitly designed to enhance and better coordinate devel- opment partner support for decentralization can be weakened or undermined by inattention to political economy issues. A particularly powerful example is the Indonesian Decentralization Support Facility (DSF), a multidonor part- nership that drew inspiration from the Paris Declaration on Aid Effectiveness with the goal of creating more consistent and better-coordinated decentral- ization activities across government ministries and development partners. The structure of the DSF was initially highly donor-driven, however, and in the eyes of some critics dominated (although not primarily funded) by the World Bank. As a result, the DSF ended up reinforcing many long-standing competitive tensions among development partners and government agencies involved in decentralization. Although Indonesia constitutes a particularly challenging environment for decentralization, it seems clear that a more care- ful mapping of the interests and incentives of various external and internal actors would have yielded a different approach to beginning the process of improving long-fragmented and inconsistent government policies and development assistance for decentralization and local governance. There are also cases where increased development partner attention to political economy issues has led to more influential and effective decentral- ization support. When Cambodia’s first decentralization law created elected commune councils in 2001, the World Bank was planning to continue its successful support to communes and their constituent villages through the Ministry of Rural Development. After more careful consideration, however, the Bank realized that the Ministry of Rural Development was one of the few ministries under the control of an increasingly weakened opposition party and that it would make more sense to channel support through the formal transfer mechanisms set up for the newly designed intergovernmental system, despite some concerns about how this rather hastily initiated system would operate. The Bank developed a new support program that bridged the Ministry of Economy and Finance and the Ministry of Interior—the two key ministries (tied to the ruling party) involved in developing the commune government system. This program created much stricter provisions for organizing and controlling resource allocation and financial management mechanisms than would likely have been the case had the Bank not been involved. These procedures have continued to evolve, and some observers have noted that they are better aligned with international standards than the procedures used by the national government. Similar forces can come into play when dealing with more targeted de- centralization issues. In Kenya, for example, development partner attempts 74 The Political Economy of Decentralization Reforms in Developing Countries over the years to improve the use of property taxation by local authorities were largely ineffective because they were primarily channeled through the relatively weak Ministry of Local Government and narrowly targeted on technical tax administration reforms, development of local government systems, and capacity building. The powerful Ministry of Lands, which is in charge of land titling and property valuation, was not adequately engaged, and without an explicit attempt to link property tax increases to better ser- vice delivery (which in some cases would have required the involvement of sectoral ministries), it was difficult to convince citizens to pay the increased taxes that resulted from tax system improvements. Later development partner efforts to improve local government revenues in Kenya were handled more carefully. A group of development partners sup- porting the Kenya Local Government Development Program, including the World Bank, took care to better link the Ministry of Local Government to the Ministry of Finance when supporting the efforts of the former to abolish the much-criticized Local Authority Service Charge—in many areas the most or second-most important stream of own-source revenue for local governments. The service charge was replaced with an intergovernmental transfer system based on rules for determining and allocating resources, which required that a portion of the funds be used for local development projects linked to participatory planning and transparent monitoring processes. Without identifying the key players in the Ministry of Finance who were willing to work with the Ministry of Local Government on these initiatives, the reforms would never have proceeded or become institutionalized. None of these examples involved exclusively successful or unsuccessful efforts—each had some positive and negative aspects. Yet the point here is not to exhaustively analyze specific countries, but to emphasize that failing to undertake political economy analysis for decentralization in general, and for specific aspects of the intergovernmental system, can reduce the effectiveness of reforms, whereas taking the time to conduct such analysis can help to improve the effectiveness of reforms. And while it is understood that few development partners completely ignore such matters, we argue that it is worth greater effort to approach PED analysis more deliberately and systematically than is generally the case. It is also important to recognize that PED analysis may in some cases point to structural or macro aspects of the political system and how they affect policies, laws, and structures—and sometimes decentralization operations. These factors are typically beyond the scope of significant development partner influence, and thus must be seen primarily as contextual variables to consider in deciding whether and how to go about decentralization support and with which government ac- tors to collaborate. Applied Political Economy of Decentralization Diagnostics 75 A key theme of PED analysis is identifying tangible, feasible actions— emphasizing reform space and opportunities rather than mechanically assessing the viability of one intervention over another. Fritz, Kaiser, and Levy (2009:16) present a spectrum of possible generic entry points that are informed by political economy insights, ranging from actions that are fea- sible in the present period to those that would require changes in underlying conditions. The focus in the latter case is not on automatically abandoning desirable reforms, but on identifying ways to nurture them—for example, through coalition building. To better understand how to transition from PED analysis to concrete operational interventions, we adopt a simple typology of the “actionable” implications that PED analysis could be expected to yield in the area of decentralization, both now and in the future (Table 1). Through this typology, we hope to provide a structure for systematically harvesting future examples to further strengthen the PED framework. A starting point would be to assess the feasibility of any given techni- cal proposal. It is important to stress that the intent of PED analysis is not to be discouraged by discovering that some desired reforms are not pos- sible, but to encourage development partners to innovate—by designing measures that are feasible or by helping to change the factors underlying feasibility over time. At the same time, it is important to appreciate that PED analysis may conclude that some reform options are non-starters in a given political environment. For example, at the time of writing, various desir- able decentralization reforms in Peru—notably around revenue sharing— are deadlocked. At the same time, there is a broader consensus that the country’s basic intergovernmental framework is in need of reform. Under such circumstances, rather than pushing what might seem like an obviously important reform that cannot move forward, emphasis could be placed on gaining consensus around a package of reforms that could gain traction or on preparing for a future political moment when effectively pursuing the desired reform becomes possible. In the meantime, development partners TABLE 1 PED Analysis-to-Action Typology Feasible Creating Reform Space Present “Non-threatening” Gaining Consensus Alternative/“Good Enough” Supporting Domestic Constituencies/ Coalitions Future Good Idea Whose Time Has Come Appreciating New Players/Dynamics Window of Opportunity 76 The Political Economy of Decentralization Reforms in Developing Countries could try to support more feasible measures that help resource-rich regions to save or spend resources well. Another case in which there might be temptation to pursue infeasible reforms is the Philippines, where technical reports have consistently high- lighted important limitations of the intergovernmental framework estab- lished through the 1991 Local Government Code with respect to service delivery. Challenges include the fragmentation of roles and responsibilities across levels of government, and the continued role of national legislators in using constituency funds directly allocated to them at the national level. These funds are used to provide local services, which some analysts have argued interferes with local government accountability. A PED analysis of the Philippines suggests that abolishing the constituency funds would meet insurmountable resistance. Thus, development partners would likely have a more productive impact by focusing on promoting feasible “good enough” measures (Table 1, top-left quadrant) that encourage greater development coordination between national legislators and local spending. These could be carried out while broader discussions and coalition building take place around more fundamental, systematic reforms (Table 1, top-right quadrant). PED analysis may identify some desirable reforms or interventions that are clearly feasible (Table 1, top-left quadrant). However, some of these ac- tions may be simple technical or supply-driven capacity-building measures that are modest in scope and effect. These can be productive, but there is a risk is that development partners would stop here, restricting themselves to narrow, easy interventions, even if they may not be sustainable or have significant development impact. This could be the case, for example, with local government budgeting reforms that would not on their own have a major impact in low-resource or limited local accountability environments, and may not be sustained if they are not put to good use. Appropriately conducted and interpreted PED analysis can help develop- ment partners to think more creatively about engaging on “good enough” and feasible actions rather than pushing unrealistic “first-best” reforms.4 To prevent such reforms from stalling, however, some attempt to stretch beyond typical approaches would be in order. For example, technical budgeting reforms could be made more attractive to local governments if they were tied to the provision of local services that are important to citizens, creat- ing an incentive for local governments to continue to use these procedural improvements. “Good enough” reforms could be useful in Uganda, for example. Given weak local government capacity, the political environment following the recently instituted multiparty elections, and the reluctance of central ministries to cede power, there is little chance that major devolution of Applied Political Economy of Decentralization Diagnostics 77 expenditure decisions will happen any time soon. It could be possible, however, for receptive ministries to give more limited discretion to lo- cal governments on certain types of decisions in order to improve their incentives and capacity to plan, and to enhance accountability to their constituents.5 In such scenarios, PED analysis could help determine which ministries might be most open to such an approach (and for what reasons). The second row of Table 1 highlights the opportunities—both techni- cal and political—that might emerge in the future. For example, sustained efforts to strengthen local capacity (even in deconcentrated entities) could lay a foundation for future devolution, especially if political circumstances change enough to open windows of opportunity (Table 1, bottom-left quadrant). A case that illustrates the potential value of thinking ahead is Indonesia, where a significant shift in political conditions created incen- tives to undertake a major decentralization (devolution) initiative in 1999 and into the following decade. Subnational government structures had long existed and had been supported by several development partners, but they were dominated by deconcentrated agencies that were heavily managed by the central government. Devolution was able to leverage the existing local institutional structures and substantial deconcentrated capacity. Despite having been held in check for decades, the long-standing local institutions and human resource capacity provided a critical building block for stronger political, fiscal, and administrative decentralization characterized by greater local powers and autonomy. Development partners can also benefit from understanding and an- ticipating the policy implications of trends and shifts in national and bu- reaucratic politics. A thorough PED analysis may help to identify emerging actors and/or dynamics that have the potential to influence decentralization trajectories. As discussed earlier, even widely supported political decen- tralization reforms may trigger countervailing forces around design and implementation, and it is almost certain that some aspects of reform will not proceed as planned. Care must be taken to support reform trajectories that build toward promising outcomes and to maintain focus on possible gaps in decentralization reforms. For example, both the Philippines and Indonesia passed major local government legislation in the wake of significant politi- cal moments, generating reforms that were in some respects passed quickly. At the same time, both systems remain partially implemented in practice and service delivery has not improved to the extent expected. Development partners need to understand why and how such situations evolved and how the interest and influence of relevant stakeholders may change, including through the emergence of potentially important constituencies (Table 1, lower-right quadrant) that may open up new opportunities. For example, 78 The Political Economy of Decentralization Reforms in Developing Countries increasingly powerful urban governments or a stronger middle class may become more prominent actors as development progresses, and this in turn could create new dynamics in the decentralization process (such as increas- ing claims over local resources to finance growing infrastructure needs). Development partners working in fragile and conflict-affected countries need to pay particular attention to how the politics of decentralization is play- ing out in overall efforts to build a stable state while beginning to enhance the delivery of basic public services (including security) in the context of severe capacity constraints. As argued above, decentralization initiatives are often at the heart of a post-conflict political settlement. Donors may prefer to focus on strengthening the central state rather than on decentralization, but attempts to consolidate power at the center can defy political needs and conflict with the expanded autonomy and resource transfers required for front-line ser- vice delivery. At the same time, central governments themselves may end up clinging to powers even if they formally adopt reforms that seem to support decentralization. For example, while the Democratic Republic of Congo’s post- conflict constitution mandates fiscal transfers; political realities have limited the implementation of transfers and reinforced the role of the center, despite capacity and accountability constraints at all levels. Development partners may be able to assist in various ways, for example by supporting the long- term priority of building capacity in local governments, while giving them space to function in an environment where key intergovernmental reforms (such as predictable transfers) may not be fully implemented by the center. Especially in fragile and conflict-affected countries where decentraliza- tion many be integral to developing and sustaining a stable state, development partners need to carefully craft a longer-term strategy that is sensitive to the overall context, and they must be forthright about their own institutional incentives. Given that counterpart government agencies may have incon- sistent views about decentralization, PED analysis could point to the value of supporting legally mandated platforms for intergovernmental political dialogue in these contexts. For example, in the Democratic Republic of Congo, the President is legally required to formally meet all Governors twice a year. In a setting where intergovernmental relations are often character- ized by limited transparency, informality, and bilateral bargaining, support for institutionalized negotiation mechanisms could have a more significant and lasting impact than more direct interventions in the decentralization process, and it could help to provide direction to development partners. In such environments, especially where donor engagement is substantial, there are no easy answers. However, understanding these situations allows development partners to better assess the benefits and risks associated with different types of decentralization-related initiatives and can help them to Applied Political Economy of Decentralization Diagnostics 79 avoid pushing reforms that may destabilize the fragile and often-evolving political equilibrium. Looking Ahead Pursuing political economy analysis will almost certainly force development partners to confront challenges that complicate how they approach reform, in some cases perhaps even going against what they see as their own organiza- tional, departmental, or individual objectives. Is decentralization support even worthwhile if the dynamics underlying a country’s reforms are not consistent with the attainment of development partner objectives? Should a donor think about working with a different government agency than it was intending to work with or than those that it has historically worked with on subnational government matters? Are there ways in which a development partner could help to promote better coordination—among other development partners, among fractured government agencies, between development partners and government agencies, or within their own organization—even if at the expense of individual glory? What might create incentives, support, and/or momentum for such an approach? Some development partners may prefer to maintain the status quo in how they approach decentralization support and collaboration, and they may see the approach advocated in this volume to be challenging and potentially onerous. On the other hand, the PED approach does build on governance assessments and other analytical work that development partners themselves have already seen the need to undertake. Perhaps most important, given the often-underwhelming performance of decentralization reform in develop- ing countries, broader-based and more careful analysis underlying decisions about whether and how to engage on decentralization reform in a particular country could be well worth the effort for development partners and the countries they support. At the same time, we see this volume only as an initial contribution to making progress in this underexplored and challenging area. We believe that there remains a need for significant learning-by-doing at the country level, which can hopefully feed back into an enhanced consolidation of lessons that emerge from different settings and can support the preparation of more robust guidelines for PED analysis. We hope that this volume provides a useful primer for development partners, country teams, and decentralization ana- lysts, offering a flexible framework to support deeper but pragmatic analysis of these demanding issues, to communicate the results of their analyses more effectively, and to act in a more informed way so as to strengthen the effectiveness of decentralization support. 80 The Political Economy of Decentralization Reforms in Developing Countries Notes 1. Development partners and client countries continuously debate which aid modali- ties and engagement strategies promise to deliver the greatest development results. Decentralization, community-driven development, and sectoral policies all share some common goals and are often undertaken simultaneously in the same coun- try. Development partners have grappled with how to best reconcile bottom-up community-driven development interventions with more top-down approaches to reforming intergovernmental fiscal systems and strengthening local government and subnational public sector institutions. Other debates focus on supporting particular sectors or programs versus taking a more territorially based approach. In all of these cases, technical analysis will need to assess the extent to which particular approaches promise to contribute to development objectives and the probability that they complement or undermine other initiatives. From a political economy perspective, these policy debates and shifts must be understood in terms of how they are rooted in the interests of salient actors. 2. Case studies on Cambodia, Democratic Republic of Congo, Peru, the Philippines, Uganda, and Vietnam, prepared as background to this volume, are available on request. 3. Within the World Bank, there is an active Political Economy Community of Practice, which has developed a “Menu of Products” (April 2010) for country, sector, and project teams, identifying three groups of political economy analysis geared toward different issues and resource constraints. The diagnostics identified in this volume are intended to be complementary and consistent with this comprehensive Menu. 4. The economics literature has also referred to “second-best” institutional arrange- ments to highlight this contrast. 5. Central governments and agencies often argue for their own bureaucratic interests that greater decentralization is not possible owing to limited capacity or accountability at the subnational level. However, if greater devolution is seen as a desirable objective, development partners could also work in a concerted fashion to strengthen subna- tional capacity as a precursor to this objective. For example, work on strengthening public financial management at the provincial level in the Democratic Republic of Congo are integral to efforts to transfer legally mandated roles, responsibilities, and financing to this level of government. Annex 1: Guidance for Political Economy of Decentralization Analysis T he Terms of Reference (ToR) for PEDCA (and PEDIA) stud- ies should aim for an output of eight to ten thousand words (about 15–20 single-spaced pages). Longer reports are likely to provide diminishing returns to informing operational engagement by development partners. The ToR for these reports should include opportunities for the authors to present their initial findings, and to revise them as needed based on feedback and subsequent investigations. The outline of the study should cover the main is- sues raised in this guidance framework, drawing in particular on Sections 3, 4, and 5 of the main PED volume. The elements of the study would include: I. Introduction II. Background: Decentralization in Historical Perspective III. The Design and Adoption of Decentralization in Country X A. Political Incentives in Country X B. Bureaucratic Incentives in Country X IV. Shifting Incentives and Contexts V. Conclusion The authors should begin the analysis by reviewing the initial country context and the most critical motivations for starting decentralization and local government reform, with a view to assessing their consistency with prevailing development partner objectives. 81 82 The Political Economy of Decentralization Reforms in Developing Countries The initial country contextual circumstances and most critical motivations for starting decentralization and local government reform: What are the main motivations for the current or planned decen- tralization reforms? Are the motivating incentives of country actors consistent with the service delivery and development objectives of development partners? If so, what aspects of reform seem most strategic and pressing to support? For example, is there a need for framework development, for creating new systems and procedures or modifying existing ones, or for supporting more effective connections between subnational governments and citizens? (See assessment of actors below.) If there does not seem to be a genuine motivation for reform, is it recommendable to disengage, to think about supporting alternative reforms, or to work on small-scale activities that might influence policy development later on? Answers to these basic questions need to be complemented by more detailed assessments in order for development partners to be able to identify appro- priate partners (national and external) and determine the levels and types of support that are likely to be viable and sustainable. The range of country actors involved in moving decentralization forward: Who are the key government actors and organizations that are or ideally should be involved in decentralization? What incentives do they face with respect to decentralization? Are they likely to support or oppose reform? In what ways, under what conditions, and at what level of intensity? To what extent are some individual actors likely to work with others that have similar inter- ests and incentives? What is the level of influence of each main actor, both in terms of his or her official position via legal or administrative empowerment and through unofficial channels that may sometimes be more im- portant than formally defined roles? Given the range of actors, their incentives, and their relative levels of power, where are there productive opportunities to engage? What are the specific opportunities, potential benefits, and potential risks associated with partnering with particular agencies? The answers to these questions will shed light on whether particular agencies or individuals—perhaps those who are not the partner’s traditional coun- Annex 1: Guidance for Political Economy of Decentralization Analysis 83 terparts in the country—might be more disposed than others and better able to promote decentralization-related reforms in line with development partner priorities. In this regard, it is also important to take stock of current and planned decentralization efforts by all development partners. The range of development partners involved in or interested in supporting decentralization: What is the nature of the current and/or potential decentralization support of each development partner? Have past and existing support activities been evaluated and re- flected in proposals for prospective programs? What are the existing relationships of each development partner with particular government agencies and other development part- ners? Although donor coordination is always a challenge and may not always be desirable at early stages, when there could be benefits to experimentation, a greater effort to understand the overall development partner situation and how it relates to national political and institutional dynamics, especially in aid-dependent countries, can provide useful information about how to position support and the effect that it is likely to have. Finally, it is important to examine the trajectory of decentralization reforms, including whether any major shifts in key dynamics have occurred or may occur in the near future. The trajectory of decentralization reform: Where decentralization has been around in some form for a while, how has recent reform been evolving in general terms and differ- entially across aspects of decentralization (administrative, fiscal, political) and across sectors? What accounts for the observed evo- lution of reforms? What is known about the performance of these reforms and the factors underlying it? Have there been changes to the political and economic conditions (and incentives) that initially triggered reform or to the institutional landscape in which these changes have played out? How have these affected the path and effects of reform to date? Might they affect prospects for the future? The conclusion of the PEDCA (or PEDIA) should begin to reflect on how the answers to these questions may affect a development partner’s deci- sion on whether and how to support decentralization in the future, with 84 The Political Economy of Decentralization Reforms in Developing Countries specific attention to the advantages and disadvantages of concrete options for engagement. References Aberbach, Joel, Robert Putnam, and Bert Rockman. 1981. Bureau- crats and Politicians in Western Democracies. Harvard University Press. Ahmad, Ehtisham, and Giorgio Brosio. 2009. Does Decentralization Enhance Service Delivery and Poverty Reduction? Cheltenham, UK: Edward Elgar Publishing Ahmad, Ehtisham, and Vito Tanzi. 2002. Managing Fiscal Decen- tralization. Oxon, UK: Routeledge. Amoretti, Ugo, and Nancy Bermeo, eds. 2004. Federalism and Ter- ritorial Cleavages. Johns Hopkins University Press. Arghiros, Daniel. 2001. Democracy, Development and Decentraliza- tion in Provincial Thailand. Cambridge University Press. Bahl, Roy, and Johannes Linn. 1992. Urban Public Finance in Devel- oping Countries. Oxford University Press. Bahl, Roy, and Jorge Martinez-Vazquez. 2006. Sequencing Fiscal Decentralization. Working Paper No. 3914. Washington, DC: World Bank. Bahl, Roy, and Paul Smoke, eds. 2003. Restructuring Local Govern- ment Finance in Developing Countries: Lessons from South Africa. Cheltenham, UK: Edward Elgar. Bardhan, Pranab, and Dilip Mookherjee, eds. 2006. Decentralization and Local Governance in Developing Countries: A Comparative Perspective. Cambridge, MA: MIT Press. Bird, Richard M., and Francois Vaillancourt, eds. 1998. Fiscal De- centralization in Developing Countries. Cambridge: Cambridge University Press. 85 86 The Political Economy of Decentralization Reforms in Developing Countries Beer, Caroline. 2004. “Electoral Competition and Fiscal Decentralization,” in Al Montero and David Samuels, eds. Decentralization and Democracy in Latin America. University of Notre Dame Press. Bland, Gary. 2004. “Enclaves and Elections: The Decision to Decentralize in Chile,” in Al Montero and David Samuels, eds. Decentralization and Democracy in Latin America. University of Notre Dame Press. ———. 2010. “Elections and Local Democracy,” in Ed Connerley, Kent Eaton and Paul Smoke, eds. Making Decentralization Work: Democracy, Development and Security. Lynne Rienner Publishers. Blunt, Peter, and Mark Turner. 2005. “Decentralization, Democracy and De- velopment in a Post-conflict Society: Commune Councils in Cambodia.” Public Administration and Development 25 (1): 77–85. Boone, Catherine. 2003. “Decentralization as Political Strategy in West Af- rica.” Comparative Political Studies 36 (4): 355–380. Brancati, Dawn. 2006. “Decentralization: Fueling the Fire or Dampening the Flames of Ethnic Conflict and Secessionism.” International Organization 60 (3): 651–685. Calvo, Ernesto, and Maria Victoria Murillo. 2004. “Who Delivers? Partisan Clients in the Argentine Electoral Market.” American Journal of Political Science 48 (4): 742–57. Cameron, John. 2009. Struggles for Local Democracy in the Andes. Lynne Rienner Publishers. Carey, John, and Mathew Shugart. 1992. Presidents and Assemblies: Consti- tutional Design and Electoral Dynamics. Cambridge University Press. Cheema, G. Shabbir, and Dennis Rondinelli, eds. 2007. Decentralizing Gov- ernance: Emerging Concepts and Practices. Washington, DC: Brookings Institution Press. Chhibber, Pradeep, and Ken Kollman. 2004. The Formation of National Party Systems: Federalism and Party Competition in Canada, Great Britain, India and the United States. Princeton University Press. Cleaves, Peter. 1974. Bureaucratic Politics and Administration in Chile. Uni- versity of California Press. Cohen, John. 1995. “Ethnic Federalism in Ethiopia.” Northeast African Studies (New Series) 2 (2): 157–188. Connerley, Ed, Kent Eaton, and Paul Smoke, eds. 2010. Making Decentral- ization Work: Democracy, Development and Security. Lynne Rienner Publishers. Collier, David, and Ruth Collier. 1991. Shaping the Political Arena. Princeton University Press. Diaz Cayeros, Alberto. 2006. Federalism, Fiscal Authority and Centralization in Latin America. Cambridge University Press. References 87 Dickovick, Tyler. 2007. “Municipalization as Central Government Strategy: Center-Regional-Local Politics in Peru, Brazil and South Africa.” Publius 37 (1): 1–25. ———. 2010. When the Center Holds: Decentralization and Recentraliza- tion in the Developing World. University Park: Penn State University Press. Dickovick, Tyler, and Kent Eaton. 2004. “The Politics of Re-centralization in Argentina and Brazil.” Latin American Research Review 39 (1): 90–122. Dillinger, William, and Steven Webb. 1999. “Fiscal Management in Federal Democracies: Argentina and Brazil.” Policy Research Working Paper No. 111. Washington, DC: World Bank. Doner, Richard, and Ansil Ramsay. 1997. “Competitive Clientelism and Economic Governance: The Case of Thailand,” in Ben Ross Schneider and Sylvia Maxfield, eds. Business and the State in Developing Countries. Cornell University Press. Durand, Francisco, and Eduardo Silva, eds. 1999. Organized Business, Eco- nomic Change and Democracy. University of Miami. Eaton, Kent. 2002. Politicians and Economic Reform in New Democracies: Argentina and the Philippines in the 1990s. Penn State University Press. ———. 2004a. Politics Beyond the Capital: The Design of Subnational Institu- tions in South America. Stanford University Press. ———. 2004b. “Designing Subnational Institutions: Regional and Municipal Reforms in Post-authoritarian Chile.” Comparative Political Studies 37 (2): 218–244. ———. 2006. “Decentralization’s Non-democratic Roots: Authoritarianism and Subnational Reform in South America.” Latin American Politics and Society 48 (1): 1–26. ———. 2007. “Backlash in Bolivia: Regional Autonomy as a Reaction Against Indigenous Mobilization.” Politics & Society 35 (1): 71–102. Ebel, Robert, and Serdar Yilmaz. 2002. On the Measurement and Impact of Fiscal Decentralization. Policy Research Working Paper No. 2809. Washington, DC: World Bank. Falleti, Tulia. 2005. “A Sequential Theory of Decentralization: Latin American Cases in Comparative Perspective.” American Political Science Review 99 (3): 327–346. ———. 2010. Decentralization and Subnational Politics in Latin America. New York: Cambridge University Press. Fitrani, Fitria, Bert Hofman, and Kai Kaiser. 2005. “Unity in Diversity? The Creation of New Local Governments in a Decentralising Indonesia.” Bulletin of Indonesian Economic Studies 41 (1): 57–79. 88 The Political Economy of Decentralization Reforms in Developing Countries Fjeldstad, Odd-Helge. 2005. “Corruption in Tax Administration: Lessons from Institutional Reforms in Uganda.” CMI Working Paper No. 2005:10. Bergen, Norway: Chr. Michelsen Institute. Francis, Paul, and Robert James. 2003. “Balancing Rural Poverty Reduction and Citizen Participation: The Contradictions of Uganda’s Decentraliza- tion Program.” World Development 31 (2): 325–337. Fox, Jonathan. 1994. “The Difficult Transition from Clientelism to Citizen- ship: Lessons from Mexico.” World Politics 46 (2): 151–84. ———. 2007. “Rural Democratization and Decentralization at the State/ Society Interface: What Counts as ‘Local’ Government in the Mexican Countryside.” Journal of Peasant Studies 34 (3/4): 527–559. Fritz, Verena, Kai Kaiser, and Brian Levy. 2009. Problem-Driven Governance and Political-Economy Analysis. Washington, DC: World Bank. September Goldfrank, Benjamin. 2007. “The Politics of Deepening Democracy: Decen- tralization, Party Institutionalization, and Participation.” Comparative Politics 39: (2). Grindle, Merilee. 2000. Audacious Reforms: Institutional Invention and De- mocracy in Latin America. Johns Hopkins University Press. ———. 2008. Going Local: Decentralization, Democratization and the Promise of Good Governance. Princeton University Press. Handley, Antoinette. 2008. Business and the State in Africa: Economic Policy- making in the Neo-liberal Era. Cambridge University Press. Heller, Patrick. 2001. “Moving the State: The Politics of Democratic Decen- tralization in Kerala, South Africa and Porto Alegre.” Politics & Society 29 (1): 131–163. Hofman, Bert, and Kai Kaiser. 2004. “The Making of the Big Bang and its Af- termath: A Political Economy Perspective,” in James Alm, Jorge Martinez Vazquez, and Sri Mulyani Indrawati, eds. Reforming Intergovernmental Fiscal Relations and the Rebuilding of Indonesia. Cheltenham, UK: Edward Elgar, pp. 15–46. Horowitz, Donald. 1985. Ethnic Groups in Conflict. University of California Press. Huntington, Samuel P. 1991. The Third Wave: Democratization in the Late Twentieth Century. University of Oklahoma Press. Hutchcroft, Paul. 2001. “Centralization and Decentralization in Adminis- tration and Politics: Assessing Territorial Dimensions of Authority and Power.” Governance 41 (1): 23–53. International Monetary Fund. 2009. “Macro Policy Lessons for a Sound Design of Fiscal Decentralization.” Fiscal Affairs Department (July). Ismail, Mawardi, Saifuddin Bantasyam, Ben Hillman, and Benjamin Reilly. 2009. Local political parties in Indonesia: the Aceh test case. Policy Brief References 89 12. Canberra: Australia Indonesia Governance Research Partnership, Crawford School of Economics and Government, ANU College of Asia and the Pacific, and the Australian National University. Kaiser, Kai, 2006. “Decentralization Reforms,” in Aline Coudouel and Stefano Paternostro, eds. Analyzing the Distributional Impact of Reforms: A prac- titioner’s guide to pension, health, labor markets, public sector downsizing, taxation, decentralization, and macroeconomic modeling. Washington, DC: World Bank, pp. 313–353. Keefer, Philip. 2007. “Clientelism, Credibility and the Policy Choices of Young Democracies.” American Journal of Political Science 51 (4): 804–821. Kimura, Ehito. 2007. Provincial Proliferation: Territorial Politics in Post- authoritarian Indonesia. Ph.D Dissertation, University of Wisconsin- Madison. Kitschelt, Herbert, and Steven Wilkinson, eds. 2007. Patrons, Clients and Policies: Patterns of Democratic Accountability and Political Competition. Cambridge University Press. Linz, Juan. 1994. “Presidential or Parliamentary Democracy: Does It Make a Differnece?” in Juan Linz and Arturo Valenzuela, eds. The Failure of Presidential Democracy. Johns Hopkins University Press. Litvack, Jenny, Junaid Ahmad, and Richard Bird. 1998. Rethinking Decentral- ization in Developing Countries. Washington, DC: World Bank. Malley, Michael. 2009. “Decentralization and Democratic Transition in In- donesia,” in Gary Bland and Cynthia Arnson, eds. Democratic Deficits: Addressing Challenges to Sustainability and Consolidation around the World. Woodrow Wilson International Center for Scholars. Mamdani, Mahmood. 1996. Citizen and Subject: Contemporary Africa and the Legacy of Late Colonialism. Princeton University Press. Manor, James. 1998. The Political Economy of Democratic Decentralization. Washington, DC: World Bank. Mawdsley, Emma. 2002. “Redrawing the Body Politic: Federalism, Regional- ism and the Creation of New States in India.” Commonwealth & Com- parative Politics 40 (3): 34–54. Maxfield, Sylvia, and Ben Ross Schneider, eds. 1997. Business and the State in Developing Countries. Cornell University Press. Murillo, M. Victoria. 1999. “Recovering Political Dynamics: Teachers’ Unions and the Decentralization of Education in Argentina and Mexico.” Journal of Interamerican Studies and World Affairs 41 (1): 31–57. North, Douglass. 1990. Institutions, Institutional Change and Economic Performance. New York: Cambridge University Press. O’Neill, Kathleen. 2003. “Decentralization as an Electoral Strategy.” Compara- tive Political Studies 36 (9): 1068–91. 90 The Political Economy of Decentralization Reforms in Developing Countries ———. 2005. Decentralizing the State: Elections, Parties and Local Power in the Andes. Cambridge University Press. PRODES. 2007. “La agenda a cuatro años de iniciado el proceso.” United States Agency for International Development. Prud’homme, Remy. 1995. “The Dangers of Decentralization.” World Bank Research Observer 10 (2): 201–220. Putnam, Robert. 1993. Making Democracy Work: Civic Traditions in Modern Italy. Princeton University Press. Rakner, Lise, Luke Mukubvu, Naomi Ngwira, Kimberly Smiddy, and Aaron Schneider. 2004. The Budget as Theatre: The Formal and Informal Insti- tutional Makings of the Budget Process in Malawi. London: Department for International Development. Ribot, Jesse. 2004. Waiting for Democracy: The Politics of Choice in Natural Resource Decentralizations. Washington, D.C.: World Resources Institute. Ribot, Jesse, Ashwini Chhatre, and Tomila Lankina. 2008. “Introduction: In- stitutional Choice and Recognition in the Formation and Consolidation of Local Democracy.” Conservation and Society 6 (1): 1–11. Roca, José Luís. 2008. “Regionalism Revisited,” in John Crabtree and Laurence Whitehead, eds. Unresolved Tensions: Bolivia Past and Present. University of Pittsburgh Press. Rodden, Jonathan. 2006. Hamilton’s Paradox: The Promise and Peril of Fiscal Federalism. Cambridge University Press. Romeo, Leonardo. 2003. “The Role of External Assistance in Supporting Decen- tralization Reform.” Public Administration and Development 23: 1: 89–96. Sabatini, Christopher. 2003. “Decentralization and Political Parties.” Journal of Democracy 14 (2): 138–150. Samuels, David. 2003. Ambition, Federalism, and Legislative Politics in Brazil. New York: Cambridge University Press. ———. 2004. “From Socialism to Social Democracy: Party Organization and the Transformation of the Workers’ Party in Brazil.” Comparative Political Studies 37 (9): 999–1024. Shah, Anwar. 2004. “Fiscal Decentralization in Developing and Transition Economies: Progress, Problems, and the Promise.” Policy Research Work- ing Paper No. 3282. Washington, DC: World Bank. Schneider, Aaron. 2003. “Decentralization: Conceptualization and Measure- ment.” Studies in Comparative International Development 38 (3): 32–56. Schneider, Ben Ross. 2004. Business Politics and the State in Twentieth Century Latin America. Cambridge University Press. Smith, Benjamin. 2008. “The Origins of Regional Autonomy in Indonesia: Experts and the Marketing of Political Interests.” Journal of East Asian Studies 8 (2): 211–234. References 91 Smoke, Paul. 2001. Fiscal Decentralization in Developing Countries: A Review of Current Concepts and Practice. Geneva: United Nations Research In- stitute for Social Development. ———. 2003. “Erosion and Reform from the Center in Kenya,” in James Wunsch and Dele Olowu, eds. Local Governance in Africa: The Challenge of Decentralization. Boulder, CO: Lynne Reinner Publishers. ———. 2007. “Fiscal Decentralization and Intergovernmental Fiscal Rela- tions: Navigating a Viable Path to Reform,” in G. Shabbir Cheema and Dennis Rondinelli, eds. Decentralizing Governance: Emerging Concepts and Practices. Washington, DC: Brookings Institution Press. ———. 2008a. “Local Revenue Generation under Fiscal Decentralization in Developing Countries: Linking Policy Reform, Governance and Capac- ity,” in Gregrory K. Ingram and Yu-Hung Hong, eds. Fiscal Decentraliza- tion and Land Policies. Cambridge, MA: Lincoln Institute Press. ———. 2008b. “The Evolution of Subnational Development Planning un- der Decentralization Reforms in Kenya and Uganda,” in Victoria Beard, Faranak Miraftab and Christopher Silver, eds. Planning and Decentraliza- tion: Contested Spaces for Public Action in the Global South. Oxon, UK, and New York, NY: Routledge. ———. 2010. “Implementing Decentralization: Meeting Neglected Chal- lenges,” in Ed Connerley, Kent Eaton, and Paul Smoke, eds. Making Decentralization Work: Democracy, Development and Security. Lynne Rienner Publishers. Smoke, Paul, Eduardo Gomez, and George Peterson, eds. 2007. Decentraliza- tion in Asia and Latin America: Towards a Comparative Interdisciplinary Perspective. Cheltenham, UK: Edward Elgar. Smoke, Paul, and Joanne Morrison. 2009. “Decentralization in Cambodia: Consolidating Central Power or Building Accountability from Below?” Conference on Decentralization in Difficult Environments. Atlanta, GA: Andrew Young School of Policy Studies, Georgia State University. Staton, Jeffrey. 2004. “Judicial Policy Implementation in Mexico City and Mérida.” Comparative Politics 37 (October): 41–60. Stepan, Alfred. 2001. Arguing Comparative Politics. Oxford University Press. Stokes, Susan. 2007. “Is Vote Buying Democratic?” in Frederic Charles Schaf- fer, ed., Elections for Sale: The Causes and Consequences of Vote Buying. Lynne Rienner Publishers. Tanzi, Vito. 2001. Pitfalls on the Road to Fiscal Decentralization. Global Policy Program Working Paper No. 19. Washington, DC: Carnegie Endowment for International Peace. Tendler, Judith. 1993. New Lessons from Old Projects: The Workings of Rural Development in Rural Brazil. Washington, DC: The World Bank. 92 The Political Economy of Decentralization Reforms in Developing Countries ———. 1997. Good Government in the Tropics. Baltimore, MD: Johns Hop- kins University Press. ———. 2000. “The Economic Wars Between the States.” Paper prepared under the auspices of the MIT Department of Urban Studies/Bank of the Northeast Project. Ter-Minassian, Teresa, ed. 1997. Fiscal Federalism in Theory and Practice. Washington, DC: International Monetary Fund. Treisman, Daniel. 2007. The Architecture of Government: Rethinking Political Decentralization. Cambridge University Press. Turton, David, ed. 2006. Ethnic Federalism: The Ethiopian Experience in Comparative Perspective. Oxford: James Curry. Van Cott, Donna. 2008. Radical Democracy in the Andes. Cambridge Uni- versity Press. Vicente, Pedro, and Leonard Wantchekon. 2009. “Clientelism and Vote Buy- ing: Lessons from Field Experiments in African Elections.” Oxford Review of Economic Policy 25 (2): 292–305. Wampler, Brian. 2007. Participatory Budgeting in Brazil: Contestation, Coop- eration, and Accountability. Penn State University Press. Weingast, Barry R., Yingyi Qian, and Gabriella Montinola. 1995. “Federal- ism, Chinese Style: The Political Basis for Economic Success in China.” World Politics 48 (1): 50–81. Wibbels, Erik. 2005. Federalism and the Market: Intergovernmental Conflict and Economic Reform in the Developing World. Cambridge University Press. ———. 2006. “Madison in Baghdad? Decentralization and Federalism in Comparative Politics.” Annual Review of Political Science 9: 165–188. Willis, Eliza, Christopher Garman, and Stephan Haggard. 1999. “Decentral- ization in Latin America.” Latin American Research Review 34 (1): 7–56. World Bank. 2005. East Asia Decentralizes: Making Local Government Work. Washington, DC: World Bank. ———. 2008. Decentralization in Client Countries: An Evaluation of World Bank Support 1990–2007. Washington, DC: Independent Evaluation Group, World Bank. Yashar, Deborah. 2005. Contesting Citizenship in Latin America: The Rise of Indigenous Movements and the Postliberal Challenge. New York: Cam- bridge University Press. ECO-AUDIT Environmental Benefits Statement The World Bank is committed to preserving Saved: endangered forests and natural resources. • 3 trees The Office of the Publisher has chosen to • 1 million BTU’s of print The Political Economy of Decen- total energy tralization Reforms on recycled paper with • 246 lbs of CO2 equivalent 50 percent postconsumer fiber in accordance of greenhouse gases with the recommended standards for paper • 1,111 gallons of usage set by the Green Press Initiative, a waste water nonprofit program supporting publishers in • 70 lbs. of solid waste using fiber that is not sourced from endan- gered forests. For more information, visit www.greenpressinitiative.org. This volume presents a framework designed to help international development partners consider the relevance of political economy issues in programmatic support to decentral- ization and local government reform. Without advocating for decentralization in general or in any particular form, nor presuming or favoring any particular decentralization objective, the authors document the value of better understanding how—primarily national and intergovernmental—political and institutional dynamics may affect how decentralization reforms are aligned with other goals—such as service delivery, governance, and poverty reduction. Drawing on different country experiences, this volume illustrates how incentives can weaken, strengthen, or shift in response to changes in political and economic dynamics that arise after reform begins. Thus it helps policy makers to better identify feasible policy reforms and appropriate partners for specific country contexts. ISBN 978-0-8213-8840-2 SKU 18840