92914 Republic of Burundi Fiscal Decentralization and Local Governance Managing Trade-Offs to Promote Sustainable Reforms Burundi Public Expenditure Review OctobER 2014 B Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms OctobER 2014 WORLD BANK Cover Design and Text Layout:Duina Reyes Managing Bakovic Republic of Burundi Fiscal Decentralization and Local Governance: Trade-Offs to Promote Sustainable Reforms i Standard Disclaimer: 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 paper 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 judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Copyright Statement: 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. If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. 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, http://www.copyright.com/. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA, fax 202-522-2422, e-mail pubrights@worldbank.org. Cover photo: A colline meeting in Rutegama Commune, Burundi. Photograph by the World Bank. Cover design and text layout: Duina Reyes-Bakovic Republic of Burundi Public Expenditure Review Fiscal Decentralization and Local Governance ii Acknowledgments T he Burundi Fiscal Decentralization and dian Revenue Authority and Mr. Kieran Holmes Local Governance report is part of a series (former commissaire général, Burundian of programmatic policy studies developed Revenue Authority). by the World Bank to assist the government in its efforts to promote accelerated and inclusive The team would also like to acknowledge the growth, improve access to basic services, and considerable time and information provided by strengthen transparency and accountability in members of communal administrations across the use of public resources. While the Government the country—from administrators to members of Burundi is the primary beneficiary of the of communal staff—who worked side by side analysis, the studies are intended for a wider with the team to produce data for this report. audience, including members of the donor This report would not have been possible community, business leaders, representatives without the generous contributions of the of civil society groups, and other stakeholders Swiss Agency for Development and Cooperation engaged in Burundi’s economy. (SDC), and the German Agency for International Cooperation, (GIZ). Finally, the donor working The report has been prepared jointly with the group comprised of SDC, GIZ, the European Government of Burundi, and the World Bank Union, and the Belgian Technical Cooperation Task Team is most grateful to its government (BTC) provided valuable support and technical counterparts for this fruitful collaboration. In inputs during all the stages of the analysis. particular, the team would like to thank His Excellency Mr. Dieudonné Giteruzi (minister The Bank task team was led by Marco Larizza of communal development), His Excellency Mr. (TTL, senior governance and public sector Jean-Claude Ndihokubwayo (former minister specialist) and included Leanne Michelle Bayer of communal development), His Excellency Mr. (co-TTL, senior social development specialist), Edouard Nduwimana (minister of interior), Professor François Vaillancourt (international Honorable Innocent Nkurunziza (senator), as consultant, fiscal decentralization and public well as Honorable Aimé Nkurunziza (deputy at finance), Professor Julien Nimubona (national the national assembly), Mr. Hilaire Ntakuwundi consultant, governance and institutional deve- (permanent secretary, Ministry of Communal lopment), Thea Hilhorst (senior land governance Development), Ms. Immaculée Bigirimana specialist), Sadia Afolabi (consultant), Lydie (permanent secretary, Ministry of Finance and Ahodehou (program assistant) and Pacifique Economic Development Planning), Mr. Jean Ndoricimpa (team assistant). Marie Ntihirageza (director general, Ministry of Communal Development), Mr. Thérence Chapter 2 on fiscal decentralization was Mbonabuca (director general, Ministry of discussed with senior experts of the Interior), Mr. Firmin Sindayihebura (director Decentralization and Subnational Finance of communal finances, Ministry of Interior), Thematic Group, and valuable comments Madame Spes-Caritas Njebarikanuye (director were received from Jonas Frank (senior public general of FONIC), Mr. Nicolas Ndayishimiye sector specialist), Vivek Srivastava (lead (director general of ISTEEBU), Mr. Domitien public sector specialist), Jamie Boex (senior Ndihokubwayo (commissaire general, Burun- fellow, Urban Institute), Serdar Yilmax (senior Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms iii economist), Anne-Lucie Levebfre (senior public Overall technical oversight and guidance was sector specialist), and Balakrishna Menon provided by Albert Zeufack (sector manager), Parameswaran (lead social development Jacques Morisset (lead economist and sector specialist). Chapter 3, on decentralized land leader), and Chiara Bronchi (lead public sector management, was discussed with senior and governance specialist and cluster leader). sector experts, who provided useful insights The team is also grateful to Philippe Dongier and comments to improve the quality of the (country director) and Rachidi Radji (country analysis, including: Andre Teyssier (senior land manager), whose advice and warm support administration specialist), Caroline Plancon were extremely useful in moving the preparation (senior land policy specialist), and Enrique process forward and engaging in effective Pantoja (senior land administration specialist). dialogue with government authorities. Credit is also due to a distinguished group of The task team remains responsible for any peer reviewers who provided comments on errors or omissions. the draft report: Robert Ebel (senior fellow, Urban Institute and Urban-Brookings Tax Policy Center), Enrique Pantoja (senior land adminis- tration specialist), and Helene Grandvoinnet Washington, DC (lead social development specialist). October 23, 2014 iv Table OF CONTENTS ........................................................................................................ iii Acknowledgments. Abbreviations and Acronyms ........................................................................................ 1 Glossary of Key Terms ..................................................................................................3 Executive Summary....................................................................................................... 4 Key Policy Issues under Investigation . ............................................................................................................................... 6 Main Messages: Cross-Cutting Issues................................................................................................................................ 11 .....................14 Chapter 1. Understanding the Context of Decentralization in Burundi. Introduction: The Political and Macroeconomic Context........................................................................................ 14 1.1 The Decentralization Process since the End of the Conflict: An Overview ........................................... 17 1.2 The Decision to Decentralize: Understanding Key Incentives ................................................................... 19 1.3 Institutional Design: Mapping Key Actors ............................................................................................................ 20 1.4 Institutional Design: Promoting Citizens’ Engagement to Strengthen State-Society Relationships . ...................................................................................................................................................................... 25 1.5 Obstacles to Moving the Reform Agenda Forward: Understanding Institutional Constraints and Political Incentives . ...................................................................................................................... 29 1.5.1 Institutional Constraints .................................................................................................................................. 30 1.5.2 Political Incentives............................................................................................................................................... 31 1.5.3 Bureaucratic Incentives ................................................................................................................................... 33 1.6 Recommendations for Policy Dialogue: “Unlocking” the Reform Agenda. ............................................ 33 A Summary of Main Arguments and Key Findings ................................................................................................... 35 Chapter 2. Fiscal Decentralization: Context, Policy Issues, and Possible Changes..........................................................................................................35 2.1 Fiscal Decentralization Today...................................................................................................................................... 37 2.2 Managing Fiscal Decentralization under Fiscal Constraints: Key Policy Issues.................................. 44 2.2.1 Policy Question 1: Is the Number of Communes in Burundi Optimal?. ..................................... 44 2.2.2 Policy Question #2: Which Functional Responsibilities Should Be Assigned to Communes? ....................................................................................................................................................... 48 2.2.3 Policy Question #3: How Can Communes Increase their Own Revenues? .............................. 49 2.2.4 Policy Question #4: How Can a Pro-Poor Funding Transfer Scheme for Current Spending Be Established to Complement FONIC?.............................................................. 51 2.2.5 What Are the Potential Implications of Mining Revenues for Communes Finances? ................................................................................................................................................................ 54 2.3 Future Fiscal Decentralization Reforms in Burundi: Possible Paths and Recommendations............................................................................................................................................................. 60 Chapter 3. Decentralized Land Administration: Challenges and Opportunities . ............................................................................................................ 64 Summary of the Main Arguments and Key Findings . ............................................................................................... 64 3.1 Land Scarcity and Tenure Insecurity as Drivers of Conflict in Burundi. ................................................. 65 Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms v 3.2 Background: Types of Tenure and the Institutional Setting . ....................................................................... 69 3.3 Promoting Land Governance Reforms: Rationale and Policy Objectives . ............................................. 71 3.4 Piloting Decentralized Land Administration: Results and Challenges . ................................................. 77 3.5 The Way Forward: Actions Needed to Scale Up Land Governance Reforms in Burundi................. 85 3.5.1 Short-Term Priorities for Reform (1–3 Years) ....................................................................................... 86 3.5.2 Medium-Term Priorities for Reform (3–5 Years).................................................................................. 88 Chapter 4. Promoting Social Accountability through Decentralization: ................................................................................................... 90 The Voice of Citizens. Summary of Main Arguments and Key Findings ........................................................................................................ 90 4.1 Social Accountability and Local Governance: Their Connection and Importance. ............................ 92 4.2 Analytical Framework, Data, and Methodology.................................................................................................. 93 4.3 Empirical Analysis: Key Findings............................................................................................................................... 96 4.3.1 Obstacles to Participation at the Local Level........................................................................................... 97 4.3.2 Preparation of Communal Development Plans ..................................................................................... 98 4.3.3 Budget Preparation and Execution ............................................................................................................. 99 4.3.4 Central Transfers: The Case of FONIC.......................................................................................................100 4.3.5 Tax Collection. ......................................................................................................................................................101 4.3.6 Monitoring Project Execution: A Case Study from the Communes of Makebuko and Rutegama—The Experience of the World Banks’ CDD Operation in Burundi......................105 4.4 Recommendations . .......................................................................................................................................................110 4.4.1 Support a Comprehensive Communication Strategy........................................................................110 4.4.2 Increase Awareness-Raising Efforts Among Both Local Citizens and Elected Officials....................................................................................................................................................................110 4.4.3 Empower Civil Society Actors to Bridge the Gap between Local Citizens and Public Authorities.............................................................................................................................................................110 4.4.4 Harmonize Tax Collection across Communes . ....................................................................................111 4.4.5 Support Citizen Engagement and Participation within the Current Decentralization Framework .......................................................................................................................112 References ...................................................................................................... 113 ANNEXES .................................................................................................................. 119 Annex 1. Communal Merger: Supplementary Information ........................................ 121 Annex 2. List of People Consulted during the Preparation of the Study...................124 Annex 3. Validation Workshop Summary Report........................................................128 List of figures Figure 1.1 Burundi Governance at a Glance, 2012 Data.................................................................................................... 15 Figure 1.2 Chronologie du processus de décentralisation au Burundi, 2000–2015. .......................................... 19 Figure 1.3 Institutional architecture of decentralization in Burundi ......................................................................... 23 Figure 1.4 Opportunities for Engagement at Different Levels of Burundi’s Decentralization Process. .................................................................................................................................................................. 28 Figure 2.1 La composition du secteur public local dans 15 pays, 2010–12 (%)................................................... 40 Figure 2.2 Communal Revenue as a Percentage of Total Government Revenue, Burundi: 2001, 2005, and 2010........................................................................................................................................................................ 41 vi Figure 2.3 Local Revenue as a Share of Total Public Revenue in 22 African Countries: 2006–08. ............... 42 Figure 2.4 Fiscal Revenue of Communes, by Quintile, 2012............................................................................................ 42 Figure 2.5 Fiscal Expenditures of Communes, by Quintile, 2012. ................................................................................ 42 Figure 2.6 Poverty Rates (%) Burundi Communes, by Quintile, year 2008............................................................. 43 Figure 2.7 Projected Per Capita Transfers to Rural Communes, Using Hypothetical Allocation Formulation ......................................................................................................................................................................................... 52 Figure 2.8 FONIC Investment Funds, 2010–12...................................................................................................................... 54 Figure 2.9 Active Mines. .................................................................................................................................................................... 56 Figure 2.10 Mining Exploration Activities, as Indicated by Claim Area .................................................................... 57 Figure 3.1 The Population Density of Communes, by Quintile....................................................................................... 66 Figure 3.2 SFC Revenues over Time (2011–13).................................................................................................................... 82 Figure 4.1 The Key Elements of Social Accountability: A Conceptual Framework............................................... 94 Figure 4.2 Public Posting of Communal Budget..................................................................................................................101 Figure 4.3 Rutegama Commune Revenues, Actual and Projected (2010–13)......................................................102 Figure 4.4 Makebuko Commune Revenues, Actual and Projected (2010–13).....................................................103 Figure 4.5 Comparison of Revenue Collection Efforts in Makubeko and Rutegama (2010–13). ................103 Figure A1.1 Distance Walked, 16 Communes, Burundi 2013 (Communes Ranked by Area). ......................122 Figure A1.2 Time Walked, 16 Communes, Burundi, 2013 (Communes Ranked by Area)..............................122 List of tables ES.1 Key Policy Recommendations.............................................................................................................................................. 13 Table 1.1 Information on the Size of Communes, Rural Provinces, Burundi, 2008.............................................. 21 Table 2.1 FONIC Budgetary Summary, 2013 and 2014. .................................................................................................... 45 Table 2.2 Number of Local Governments (Lowest Tier) and Average Populations in 17 African Countries, 2006–08. ............................................................................................................................................... 46 Table 2.3 Boroughs Around the World: Strengths and Weaknesses in Three Areas........................................... 47 Table 2.4 An Example of a First-Generation Transfer Formula: Kenya...................................................................... 52 Table 2.5 A Simulation of Revenue Sharing with Producing Areas Only .................................................................. 60 Table 2.6 Simulating Revenue Sharing with Producing and Nonproducing Areas............................................... 60 Table 3.1 The Frequency of Land Disputes, Ranked by Household Type and Method of Acquisition ....... 67 Table 3.2 Perceived Risks to Land Rights, Ranked by Household Type and Source. ........................................... 71 Table 3.3 The Acquisition of Plots in Rural Areas, Ranked by Method and Type of Household . .................. 75 Table 3.4 Estimating the Timing of On-Demand Registration ...................................................................................... 81 Table 3.5 Plots Registered and Certificates Issued by the SFC in Ngozi Province (SDC Pilot) ....................... 83 Table 3.6 Costs of First-Time Registration (On-Demand) for 8 SFCs . ........................................................................ 83 Table 4.1 List of Communes Where Focus Group Discussions Were Convened. ................................................... 95 Table 4.2 List of Communes Where Additional Data Was Collected Beyond the Focus Groups. ................... 96 Table A1.1 Characteristics of Communes Surveyed in August 2013 ........................................................................121 Table A2.1 Government of Burundi...........................................................................................................................................124 Table A2.3 World Bank Staff..........................................................................................................................................................127 Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms vii List of Boxes Box 1.1 Key Legislation ..................................................................................................................................................................... 18 Box 1.2 Legal Provisions Regulating the “Tutelle” System............................................................................................... 24 Box 1.3 Current Decentralization Procedures in Burundi................................................................................................ 27 Box 1.4 Sectoral Ministries Involved in the Decentralization Process........................................................................ 31 Box 2.1 Assessing the Costs and Benefits of Reducing the Number of Communes: An Empirical Analysis with Unchanged Services.................................................................................................................. 47 Box 2.2 Establishing an Equitable and Transparent Transfer System for Communes: A Proposal with Simulations........................................................................................................................................................... 53 Box 2.3 Scenario 1: Mining Revenues Shared Across Producing Areas.................................................................... 58 Box 2.4 Scenario 2: Mining Revenues Shared Across Both Producing and Nonproducing Areas ............... 59 Box 3.1 Gender Equity and Land Ownership: Learning from Ethiopia and Rwanda......................................... 68 Box 3.2 Land Governance: A Definition.................................................................................................................................... 72 Box 3.3 Decentralizing Land Management: The Experience of Madagascar.......................................................... 74 Box 3.4 Securing Land Rights in Burundi: Why Does It Matter?.................................................................................. 76 Box 3.5 Principles for First-Time Registration Systems. .................................................................................................. 78 Box 4.1 GoB Performance Evaluation of Communes ......................................................................................................100 Box 4.2 Participatory Budgeting (PB) in South Kivu, DRC: Supporting Decentralization and Empowering Citizens to Participate in the Budgetary Process through ICT................................................104 Box 4.3 Balancing Short-Term Needs with Long-Term Goals: Findings on Community Driven Development (CDD) Projects from the 2013 IEG Report. ..............................................................................................107 viii Abbreviations and Acronyms ABELO Association Burundaise des Elus Locaux (Association of Burundi’ s Locally Elected Officials) ADLP Association pour le Développement et la Lutte contre la Pauvreté (Association for Development and the fight against Poverty) APDH/ICCO Association pour la Paix et les Droits de l ‘ Homme (Association for Peace and Human Rights)/International Coalition CAS Country Assistance Strategy CCDC Comité Communal du Développement Communal (Committees on Communal Development) CDC Conseil de Developpement Collinaire (Colline Development Council) CDD Community Driven Development CDPs Communal Development Plans CEM Country Economic Memorandum CENI Commission Electorale Nationale Indépendante (National Independent Electoral Commission) CIOPD Comité Interministeriel d’Orientation et de pilotage de la decentralization –CIOPD (Interministerial Steering and Orientation Committee on decentralization) CNDD-FDD Conseil National Pour la Défense de la Démocratie–Forces pour la Défense de la Démocratie (National Council for the Defense of Democracy–Forces for the Defense of Democracy) CNTB Commission Nationale des Terres et Autres Bien (National Commission on Land and Assets) CSLP Cadre Stratégique de Croissance et de la Lutte contre la Pauvreté (Strategic Framework on Growth and the fight against Poverty) CSO civil service organization CTB Coopération Technique Belge (Belgian Technical Cooperation) DRC Democratic Republic of Congo DTF Direction des Titres Fonciers (Land Titling Directorate) EAC East African Community ESW Economic Sector Work EU European Union FAO Food and Agriculture Organization FBu Franc Burundais (Burundian Franc) FNL Forces nationales de libération (National Forces of Liberation) FONIC Fonds National d’Investissement Communal (National Investment Funds for Communes) Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 1 GDP Gross domestic product GIS Geographic information system GIZ Gesellschaft für Internationale Zusammenarbeit (German Agency for International Cooperation) GoB Government of Burundi ICG International Crisis Group ICT Information and communication technology IDPs Internally displaced persons IE Impact Evaluation IGEBU Institut Géographique du Burundi (Burundi Geography Institute) ISTEEBU Institut de Statistiques et d’Etudes Economiques du Burundi (Burundi Institute of Statistics and Economics Studies) LC Land code LGAF Land Governance Assessment Framework MDGs Millennium Development Goals MEEATU Ministère de l’Eau, de l’Environnement, de l’Aménagement du Territoire et de l’Urbanisme (Ministry of Water, Environment, National Planning and Urbanism) NDS National Decentralization Strategy NGO Nongovernmental organization OBR Office Burundais des Recettes (Burundi Revenue Authority) ODA Official Development Assistance OGR Opération Groupée de Reconnaissance (Organized Grouped Titling) PAF Poverty Alleviation Fund PB Participatory budgeting PCDC Plan Communal pour le Développement Communautaire (Communal Development Plans) PNF Programme National Foncier (National Land Program) PRADECS Projet d’Appui au Développement Communautaire et Social (Social and Communal Development Support Project) PRODEMA Projet de Productivité et de Développement des Marchés Agricoles PCCD Provincial committee on communal development PS Permanent Secretary REGIDESO Régie de Production et de Distribution d’Eau et d’Electricité SA social accountability SDC Swiss Agency for Development and Cooperation SDV social development SEGOM World Bank Oil, Gas and Mining Unit SFC Service Foncier Communal (Communal Land Service) UNDP United Nations Development Programme UPRONA Union pour le Progrès national (Union for National Progress) VAT Value added tax WB World Bank WDI World Development Indicators 2 Glossary of Key Terms Administrateur communal Councillors (and/or a mayor) elected to lead at the commune level Autorités collinaires/ chef collinaire Person or group of persons in charge of administrating the colline Bashingantahe Customary institution, the “men of integrity”, responsible for settling conflicts at all levels in society and can be requested to mediate land disputes Commune Administrative unit of decentralization according to Burundian Constitution and other legal texts. Councillors and mayor (administrateur communal) are elected Colline “Hill,” an administrative subunit under the commune; can be translated as a “ward” Domaine de l’Etat Privé Public lands that can be allocated by authorities to private parties Domaine de l’Etat Public Public lands, such as protected forest lands and parks, that cannot be allocated Mairie “City hall”. The french term used only for Bujumbura City Marais/Bas-Fonds Marshland, parts of which are cultivable (bas fonds) Paysannats Annual leaseholds on public lands that are part of government schemes for cash crops (for example, cotton, palm oil, rice), located in the fertile plains Service foncier communal Commune-level department in charge of land registration in rural areas and mandated to issue land certificates Sous-colline The third level of administrative subunit under the commune Tutelle Supervisory relation between the commune and the central government Villagisation The development of villages that bring together dispersed rural populations and allow the resettlement of returning refugees Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 3 Executive Summary Rationale and Objectives international policy makers that the exclusion and deprivation of the rural masses was one 1. Despite the remarkable progress of the driving forces of war. In this context, achieved since the end of the conflict, decentralization (together with the inter-ethnic Burundi still faces significant development power sharing mechanisms enshrined in the challenges. Prior to the conflict, the percentage 2005 constitution) emerged as a key strategy of people living below the national poverty line for opening up space for citizen participation was 36.4 percent (1990). During the conflict, in local development, improving access to basic this increased to 68 percent (by 1998). As of services and—by doing so—strengthening the most recent household survey, the rate the social contract between the state and its was still high: 67 percent of the population citizens.2 As in other countries of the region, was living below the poverty line in 2006.1 decentralization was also seen as a way to Although education and health outcomes have improve the institutional capacity of weak improved significantly over the past few years postcolonial states and move toward more open (life expectancy rose from 43 years in 2000 to societies.3 50.4 years in 2011), access to basic services and infrastructure such as potable water and 3. Since 2005, the Government of electricity remain limited (for example, less Burundi has embarked on a potentially than 3 percent of the population has access transformative process of decentralization, to electricity). Meanwhile, high population with the aim of strengthening social density and growth rates—about 90 percent cohesion, improving local governance, and of the population lives in rural areas—increase promoting access to basic infrastructure competition for scarce resources, especially and service delivery. Since the promulgation land. As a result, Burundi still ranks low in the of a new constitution and the installation of a UNDP’s Human Development Index (178 out of democratically elected government in August 187 countries in 2013) and is unlikely to meet 2005, the government has made a number of most of the MDGs by 2015. important steps forward, including revising the electoral code prior to communal elections, and 2. Given the historical legacy of conflict, strengthening both national and subnational institutions has emerged as a national 2  The main goal of decentralization in Burundi is “to help priority, with the aim to improve stability communes to be the foundation of development and to decentralize and service delivery. In 2000, peace accords government institutions. The government will transfer resources to the communes, so they can plan development projects in were signed by opposing political parties. In the domains, such as education, public health, roads and water supply. context of the 2000 Arusha Peace Agreement, To transfer this development means, injustice should be avoided” a consensus emerged among national and (Arusha Accords, 2000: 171). 3  The content and direction of this policy debate is not unique to Burundi and resembles the historical experience of other postconflict and fragile states, including Sierra Leone (Larizza 1 The results from the latest household poverty survey will be and Glynn 2013) and the Democratic Republic of Congo (Keefer available end-2014 and provide updated poverty figures. 2011). 4 later adopting a new Communal Law governing and the second-largest share, at 5 percent, the main level of devolved entities at the from Gitega.4 The average annual communal municipal level, in the form of 129 communes revenue outside Bujumbura is FBu 103 million, (116 of which are outside the capital and thus equivalent to $68,000 or somewhat less than $1 mostly rural). These communes are broken per capita. Overall, communal revenues remain down further into zones (administrative, low and in decline, down to 2.7 percent of total deconcentrated units) and collines, which have internal revenues in 2010. Resources to finance an elected council but no autonomy or budget. the execution of Communal Development Plans The process of political decentralization was are therefore largely insufficient, with the result also launched with the first communal elections that available social infrastructure, as well as the in June 2005. The government commitment maintenance and operations of service delivery to decentralization reform was further points (health clinics, primary and secondary bolstered by the adoption of a “Policy Letter on schools, and water distribution points), remains Decentralization and Community Development limited to address citizen needs and improve (2007) and the publication of the new National access to social services. Decentralization Strategy (NDS) in 2009. 6. Given the fiscal fragility of the country 4. Since 2010, however, Burundi’s and its structural vulnerability to external decentralization experience has confronted shocks, the central government has little a number of challenges, and many issues fiscal space to improve the fiscal viability remain to be addressed before the reform of communes. As indicated in the recent PER agenda can move forward. Since 2010, when (2013), the government continues to struggle a few amendments to the 2005 communal law to create fiscal space and reduce debt from its were introduced, there has been little progress currently high level. Although the government is toward decentralization in Burundi. The law on expected to limit current expenditures growth in the new responsibilities of communes remains real terms, unless additional revenue measures to be adopted, as the parliament rejected the are introduced to offset the falling grants, it current proposal, given the lack of information will struggle to reduce its deficit. Under these on its financial implications (i.e. amount of fiscal structural constraints, the financial implications transfers to communes necessary for the proper of proposed legislation need to be carefully implementation of the proposed new devolved considered. Other options currently debated functions). A separate law on communal finances among policy makers—including the proposal is under discussion and has not been submitted to reduce the number of communes—are also to parliament yet, while the law on a territorial associated with potential risks. The government civil service is not much more advanced and the is aware of such trade-offs and has requested law revising the role of local administrators is on the assistance of the Bank and a group of donors hold. Finally, the law on revising the legal status to assist ongoing efforts to identify feasible of Bujumbura Mairie is also under discussion. options to manage the reform process and improve the medium-term sustainability of the 5. The weakness of the communal tax current system, under conditions of macro-fiscal system, coupled with low mobilization of fragility. local revenue and nonexistent (current) or negligible (capital) transfers from the na- 7. Addressing a specific government tional budget threaten the financial viability request, the present study aims to provide of communes, which struggle to support even basic operating costs. In 2012 total communal revenues reached FBu 16 billion, with almost 4  Data provided by the direction des finances communales, 25 percent from the city of Bujumbura (Mairie) Ministère de l’Intérieur. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 5 concrete policy recommendations to help the status of fiscal decentralization in Burundi, the Government of Burundi improve the and identifies key policy issues to be considered financial and institutional sustainability of to ensure the medium-term sustainability of the decentralization reform process, while the reform process while at the same time enabling communes to address popular addressing the short-term financial needs of demands and deliver better services. communes. Chapter 3 provides an in-depth Specifically, the study will focus on two main diagnostic of a key service delivery responsibility challenges: (i) how to put in place a sustainable, recently devolved to communes—the provision predictable, and transparent inter-governmental of land registration services—and discusses the fiscal framework able to meet the financial challenges and opportunities related to ongoing needs of the communes and improve access efforts to scale up access to these land services to basic services, taking into consideration across 116 rural communes and Bujumbura. macro-level fiscal constraints; and (ii) how to Chapter 4 shifts the focus to the nature of state- promote effective strategies to increase citizens’ citizen relations in an effort to better understand participation in local governance and other how citizen engagement in the decision-making measures of social accountability. Both topics process may be improved and local authorities are at the core of the national decentralization held accountable for the provision of basic strategy. The report will also look at the services. implications of these macro-level challenges at the sectoral level, through a case study of 9. The report is based on results the recent experiences of decentralized land from interviews, fieldwork research, administration services, whose responsibilities and qualitative focus group discussion, were recently transferred to communes (Land combined with existing administrative data Code 2011). The in-depth focus on land tenure and secondary sources on decentralization security reflects a service delivery area of high in Burundi. The fieldwork research was policy relevance for its potential contribution carried out in 16 communes (representing all to the prevention of conflict and promotion of provinces but Bujumbura). A multisector team local investment. prepared the report in close coordination with government officials and in joint collaboration 8. To avoid replication of previous with a group of key donors active in the sector. analysis,5 the present study is highly selective and organized into four thematic chapters. Key Policy Issues under Specifically, chapter 1 provides a snapshot of Investigation Burundi’s political and macroeconomic context, and reviews the evolution of the decentralization Understanding the Context of the process to better understand how institutional, Decentralization Process political, and bureaucratic dynamics have shaped 10. At the end of the conflict, the the historical trajectory of decentralization Arusha Peace Agreement and the 2005 and generated the outcomes observed today. Constitution laid the political foundations Chapter 2 provides a systematic investigation of for decentralization reforms. In 2005, the Communal Law and Electoral Code further confirmed the central government’s commitment 5  A wealth of analytical work and advisory services have recently been conducted by the government and other donors active in the to open the way for more direct participation decentralization sector, notably: (i) a review on “the Senate and the and accountability at the local level. The law Politics of Decentralization in Burundi”; (ii) a recent Swiss-funded recognizes communes as autonomous and study on the fiscal viability of municipalities (covering a small sample of communes) and options for effective decentralization decentralized entities managed by communal reform; (iii) a recent UNDP-funded study on FONIC. council members. It also recognizes the 6 communes’ deliberative and decision-making 2010. This can be explained by a combination powers by allowing them to adopt economic and of institutional constraints, as well as political financial measures, including the preparation and bureaucratic incentives: of the communal budget and the communal community development plan (Plan Communal • At the institutional level, the lack of a du Développement Communautaire, PCDC). strategic vision is reflected in a fragmented The 2005 Electoral Code, amended in 2009, legislative process. Several actors within established the number of communal council and outside the government (including members to be elected through direct universal members of parliament, local actors, and civil suffrage within each commune, and regulated society organizations) lament a piecemeal the election of the communal administrator approach to legislative initiatives, which among the councilors (via secret ballot) and in turn makes difficult for national policy nomination by a presidential decree. makers to influence the overall direction and coherence of the decentralization program. 11. To a large extent, the design of Moreover, ongoing competition between decentralization in Burundi addressed the central ministries and vice presidencies political concerns of the central government over the political leadership of the reform and ruling elite. The ruling elite perceived causes blockages and mutually reinforcing decentralization as a good opportunity to deadlocks during the processes of analyzing establish and maintain political control in the and adopting legal texts. countryside, thereby strengthening political • In terms of political incentives, ruling elites stability. Promoting decentralization was also continue to perceive decentralization as an seen as a conflict-mitigation strategy in the opportunity to establish power and control context of the broader peace-building agenda: a at the local level, thereby resisting efforts key source of popular discontent prior to the war to increase communes’ administrative was the marginalization of rural communities and financial autonomy from the center. and the centralization of economic and political In particular, uncertainties arising from power in the hand of urban elites. political crises (such as the crisis in early 2014) and the subsequent remapping of 12. Improving and encouraging state- political loyalties at the local level suggest citizens interactions also emerged as a clear that effective decentralization poses risks policy objective of the new decentralization the ruling government might not be ready process. As enumerated in the government’s to take in the short-term — likely not until new decentralization policy (2009), citizens are electoral outcomes in 2015 clarify which no longer considered “subjects” but “partners” party coalitions will be in control of the local in governance. According to the new framework, administrations. there is a clear division of responsibilities among the communal council, the administrator, and • In terms of bureaucratic incentives, among the hill council (conseil de colline), as well as a civil servants prospects of decentralization balance between opportunities for participation create fears about the security and location and greater communication. In practice, evidence of employment. In particular, civil servants suggests that various factors act as binding are reluctant to leave their civil service constraints on citizens’ effective participation positions to join communal administrations and the accountability of administrators. without assurance that they will retain full status. Furthermore, most civil servants 13. As previously noted, despite ambitious have no intention to be posted outside urban goals, little progress has been made toward centers and seem to be especially reluctant effective decentralization in Burundi since to be accountable to under- educated Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 7 local executives. This in turn undermines take into account communes’ taxable capacity the availability of (already scarce) human or needs. resources, thereby reinforcing the belief that communes are not prepared to fulfill their 16. In the absence of a predictable and new mandates. transparent intergovernmental transfer system, most communes struggle to remain 14. Going forward, the analysis suggests fiscally viable and to finance even basic that—at least in the short term and until operating costs. Communes’ socioeconomic the 2015 national and local elections—key diversity (in terms of size and economic political and bureaucratic incentives are potential) represents yet another technical unlikely to change significantly. Opportunities challenge to the equitable distribution of public exist, however, to unlock major institutional resources across the national territory, in a constraints. This implies the need for context where local revenues represent only a development partners to lower expectations and small fraction (2.7 percent) of total government set realistic targets for support. That said, it might revenues and remain largely concentrated in still be possible (and strategically meaningful) Bujumbura (generating alone 25 percent of total to focus efforts toward addressing institutional local revenues). constraints that may not necessarily reflect a well-defined political strategy and instead be 17. Given the existing assignment of the undesired effect of poor coordination across functional responsibilities, the current ministries. In particular: (i) efforts are needed territorial organization of Burundi is to reach a consensus on the strategic direction reasonable, and there is no need to reduce the of reform; (ii) a coherent vision needs to be number of communes. Moreover, regrouping developed as the foundation for a comprehensive communes as such does not necessarily alleviate and coordinated legal and institutional an overall lack of resources, and surveys framework; and (iii) a presidential decree suggest it may limit peoples’ access to services. should clarify the power and competencies of Compared with other African countries, the steering committee (CIOPD) to better signal Burundi—and its 116 rural communes, with an to the government’s political will and long-term average population of roughly 90,000—is not vision for decentralization reform. an outlier. Given that communes mainly provide Managing Fiscal Decentralization: local services with little technical content (high Context, Policy Issues, and Possible technical content is associated with services such Changes as water treatment or composting, for example) or fixed costs (similar services), the case for a 15. Despite its importance, fiscal national consolidation policy does not seem to decentralization represents the weakest link be justified. Meanwhile, the legal status of the in the overall decentralization framework. communes of Bujumbura should be clarified. An adequate transfer system for current expenditures is yet to be defined: to date, 18. General responsibilities assigned there is no mechanism in place to ensure that to communes should be clarified communes receive equitable and transparent before finalizing more robust financing transfers for current spending. This is unusual arrangements. The absence of clear and specific as territorial inequalities make this necessary in mandates reduces communes’ capacity to play most countries of the world and—in the case of a key role in improving local service delivery. Burundi—the uneven distribution of resources The draft law on communal competencies might contribute to subnational inequalities. under preparation represents a step forward For capital expenditures, FONIC provides some toward addressing uncertainty and improving support with an equal transfer that does not the institutional viability of communes. At the 8 same time, its financial and human resource communes for the environmental degradation implications have not been adequately that accompanies mining-related activities, (ii) assessed. sets aside some resource revenues for future generations, and (iii) shares some of the resource 19. On average, communal revenue revenues between producing and nonproducing collection practices are weak and there is communes. scope for significant improvement. The legal framework needs to be updated to ensure that Securing Land Rights through local revenues are adjusted by inflation. At the Decentralized Land Administration: same time, efforts are needed to strengthen Challenges and Opportunities local tax administration systems, improve tax 22. With the Arusha Peace and compliance, and strengthen the link between Reconciliation Agreement, signed in 2000, improved revenue collection and better services land restitution and land governance for citizens. Collaboration with OBR should be reform emerged as key priorities in efforts promoted to share information on taxpayers— to address the root causes of conflict. The in line with the recent experience of other Arusha agreement proposed a set of measures countries, such as Rwanda—and/or use OBR as connected to (i) the restitution of land to citizens tax collection agent for the communes. affected by violent conflict, and (ii) the reform 20. The introduction of a recurrent of land governance. Given the large number expenditure transfer scheme seems of refugees and internally displaced people in appropriate. The optimal degree of Burundi, it is not surprising that land issues are deconcentration and devolution in Burundi a key element in the Arusha accord and a focus is not clear. Burundi is a small country, of postconflict donor and nongovernmental characterized by a common language and is organization (NGO) programs, such as mediation not divided geographically into linguistic or and conflict resolution. With the establishment ethnic groups. Thus many of the typical reasons of the CNTB in 2006,6 a procedure for land for devolution do not hold. At the same time, restitution to refugees and displaced people was there is a need to address inequalities across developed, while the land governance reform communes and their need for more resources, and the revision of the 1986 Land Code took in particular given the provision of important longer to materialize and is only just beginning basic services now held at the communal level. (ICG 2013). Specifically, generalizing the communal land 23. A mountainous, landlocked country, service (Service Foncier Communal; see chapter Burundi has one of the highest rural 3) requires financial resources that fees alone population densities in Africa (276 people/ are unable to cover. Also, the expected increase square kilometers [km2] on average); 90 percent in FONIC capital transfers will create greater of the total population lives in rural areas. There demand for recurrent spending that will remain is significant variation in population density unfulfilled. across regions, especially when excluding Bujumbura Mairie.7 The amount of arable land 21. Given the expected increase in the is estimated at 35.8 percent, and cultivated land importance of large-scale mining activities and their geographic concentration, it may be important to consider a revenue- 6  With financial support from the UN Peace Building fund. sharing arrangement across provinces and communes. There is an opportunity to 7  Bujumbura, the only city of significant size, has fewer than 1 million inhabitants, making it relatively small for a country whose integrate both decentralization and mining population currently stands at nearly 10 million inhabitants. The through a mechanism that: (i) compensates country second city, Gitega, has fewer than 50,000 inhabitants. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 9 at 15.6 percent in 2011 (a marked increase coordination de program foncier national, an from 14.0 percent in 2000). Farming is the interministerial body located under the MEEATU main livelihood for most of Burundi’s rural (Ministère de l’Eau, de l’Environnement, de households, making land the main economic l’Aménagement du Territoire et de l’Urbanisme). asset for the majority of the poor. The aim of the pilot programs was to design and test land administration service-delivery 24. A combination of increasing systems for the registration of land rights that demand for land, growing land scarcity, and are trustworthy, transparent, accessible, and ambiguous land rights undermines stability legal. The first pilot programs started in 2007 and acts as a driver of conflict at the local level. mainly with support of the European Union Moreover, settlements carry high costs for those (EU) and the Swiss Agency for Development and engaged in land disputes. Communal officials Cooperation (SDC) in six communes. are frequently requested by the involved parties to act as mediators in land conflicts.8 Improved 27. The provision of communal land land administration can reduce the burden on registration services raises important scarce human resources by mitigating conflicts sustainability and financing issues, and speeding up dispute resolution. strengthening the case for a recurrent grant to cover operating costs. The 28 25. With the approval of the new Land communal land services currently operating Code (2011), the land certificate has been are primarily funded by donors. Unless fiscal- introduced as an alternative way to record administrative aspects of decentralized land rights in rural areas. Investments are being governance are clarified, the SFCs will continue made to improve capacity and strengthen the to depend on development partners to finance land administration system (through archives, their operations. Consequently, achieving the digitalization, training, and so on) which is goals of generating useful information and important for urban areas where informality reducing conflict through land registration may concern 60 percent of properties. But would be difficult and dependant on user fees overall capacity remains insufficient to serve in a demand-based system. Instead, a recurrent rural areas, where most land conflicts are found transfer to communes might be considered and where policy concerns are focused. To to cover recurrent costs of communal land address this challenge, the Land Policy Letter registration services, while set-up costs can be (LPL 2010) and Land Code (LC 2011) introduced addressed by donors (when available) and/or the establishment of decentralized, communal covered by the annual FONIC grant9. Increasing land services (the Service Foncier Communal, the financial resources of communes should SFC) for the registration of existing land rights, take into account their location and topography, land transactions, and other mutations. two factors that impact the costs of a communal land registration office. 26. The establishment of the SFC is coordinated by a special government unit 28. Plans for scaling up decentralized and has been largely funded by donors. This land services require the availability innovation in service delivery is the result of the of reliable and regular statistics and work of several donor-supported pilot programs administrative data to monitor results and working closely together with the Unité de processes. Currently, monitoring is weak with 9  To this ed, the FONIC grant would be targeted in the same way 8  The mayor (administrateur communal) in Ngozi estimated as past grants were targeted to primary schools – with national that she spent two days a week, on average, in the mediation of priorities overriding communal investment priorities for one conflicts often caused by land disputes (interview, June 2013). year. 10 respect to outputs, costs, quality, and so on, infrastructure. By facilitating open dialogue and land data are scattered and not aggregated and the inclusion of vulnerable groups, these at the central level. Supervision seems to interventions made a significant contribution be irregular while progress assessment and toward social cohesion and conflict mitigation. learning seems to depend on donor initiatives for self-assessment and external evaluations, 31. Despite this recent progress, which also seem insufficiently coordinated. For accountability relationships remain weak, example, the cost per certificate can probably be with limited opportunities for citizens and reduced even more by streamlining procedures communities to engage in the decision- and enhancing efficiency, but this requires making process and hold local authorities regular monitoring and testing, and comparing accountable. Several “entry points” for of alternatives. interventions can nevertheless be identified, particularly on the side of the state. Citizen Promoting Social Accountability through engagement in local decision-making and Decentralization participation in local development remains limited, undermining the potential benefits of 29. Although the nature of the state-society decentralization reform as a vehicle to improve relationship in Burundi has significantly accountability and strengthen local governance. improved since the government embarked Yet, given culturally embedded obstacles to on an ambitious decentralization process citizen participation, the accountability of state in 2005, political realities and the legacy of actors is recommended as an area of initial conflict persist. After decades of conflict, the focus. severity of violence observed across Burundi has been markedly reduced. However, the root 32. Another key message that emerged sources of the conflict (such as exclusion—social, from the focus group research was the economic, and political) remain largely intact. need for more transparent and systematic Even as social cohesion is being rebuilt across mechanisms to collect revenues and prepare/ local communities, the persistence of a neo- execute communal development budgets. patrimonial system of governance is blocking Specifically, there is a willingness and openness state reform, constraining improvements in from local officials to communicate more social service delivery at the local level. effectively with citizens to increase communal tax revenues. On the flip side, citizens are eager 30. Recent political reforms have created for greater transparency around local resources some opportunities for communes to play an and to better understand the use of tax revenue important role in building social cohesion by local officials. at the local level, in particular through the direct election of communal councils. Due to Main Messages: Cross-Cutting growing support from the Bank (the Community and Social Development Project, PRADECS) Issues and the donor community at large, communal A number of cross-cutting issues emerged administrations have become increasingly able from the study as critical to moving the to manage funds for small development projects. decentralization process forward: This included formulating the first Communal Development Plans, identifying priority First, there is a strong linkage between the subprojects through participatory community usefulness of a country-wide communal-level meetings, managing the procurement process, land titling process and the need for central monitoring subprojects, and guaranteeing government support of communal spending. the maintenance of newly built communal The importance of securing land rights to Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 11 preserve peace and improve security in Burundi and stronger social accountability. Large- (discussed in chapter 3) strengthens the “public scale mining is often disruptive of communal good” aspect of land ownership policies. This life as it leads to new use of land, demand on in turn makes funding these activities through communal services, and the perception that the existing fee-based model inappropriate and opportunities and risks are shared unequally unsustainable. As an alternative, the proposed across local communities. Securing land current expenditure transfer (discussed titling can make it easier to adapt to mining, in chapter 2) would allow communes the while better social accountability and citizen necessary resources to operate communal land engagement in the decision-making process services, while linking central spending with a will facilitate acceptance of mining-related key national priority. communal spending. At the same time, the Second, there is a need to improve provision of revenue-sharing arrangements for social accountability and local revenue natural resources at the communal level will generation—in step with each other. There reduce the perception that resources belonging is a clear need to improve communes’ own to the community are being shared unjustly. revenues through an updated communal tax Last, the need for coordination between code, but the communal population is currently ministries and agencies at the central not convinced that communal spending is level—as well as between central and local relevant to their welfare. Changing the tax code authorities—is clear from the three previous should thus be accompanied by concurrent points. To ensure that the various concerns changes in social accountability mechanisms to discussed above are put forward in an integrated, promote transparency and accountability in the coherent, and well-sequenced fashion requires use of public resources, including local taxes. a strong coordinating mechanism. This is a tall The recommendation to post information on order for a country where institutional capacity planned and actual budget expenditures in the is weak and few incentives exist to promote local language (Kirundi) illustrates this point. interministerial coordination. Targeted technical Third, the likely increase in large-scale assistance from the donor community will be mining calls for improved security in land critical to facilitate the process and move the ownership, strengthened communal finance, reform agenda forward. 12 ES.1 Key Policy Recommendations Objective Policy area Proposal for priority actions Timeline 1. Clarify the a. Territorial Settle the number of communes. Short term strategic direction organization Medium term of decentralization Settle the legal status of communes in reform b. Territorial Bujumbura. Short term organization Clarify the power and responsibilities of c. Central leadership the national Interministerial Steering and and organization Orientation Committee on decentralization (CIOPD). 2. Create a a. Communes’ Settle the functional responsibilities of Short term predictable and responsibilities communes. Short term transparent intergovernmental b. Communes’ own Adopt the new law on communal financing. Medium term financial system revenues Medium term Introduce an equitable and transparent, c. Transfer revenues formula-based transfer scheme for recurrent for communes expenses. d. Mining sector Integrate decentralization and mining sector development development through the design of revenue- sharing arrangements across communes. 3. Promote social a. Communes’ own Harmonize tax-collection methods and develop Short term accountability at revenues an information campaign for citizens and local Short term the local level CSOs. b. Budget Medium term transparency Harmonize and disseminate simplified budget information (planning and execution) in Kirundi. c. Communes’ responsibilities Reinforce citizen engagement mechanisms at the local level. 4. Scale up the a. Legal framework Adopt the draft decree on the certification of Short term decentralized for decentralized land rights, to codify the results emerging from Short term land management land administration the pilot programs and harmonize the approach system across communes. Short term b. Financial Medium term sustainability of land Introduce a transfer scheme for recurrent Medium term services spending that could be used to cover operating costs for land services (including SFC staff). b. Quality of land administration Develop standard procedures and guidelines for services provided by SFC guidance, supervision, and oversight. communes Establish institutionalized mechanisms for c. Human resource the capacity building and training of public management servants, councilors, and private parties engaged in decentralized land administration. d. Land information statistics Establish an integrated land information system to be used to track progress, inform policy dialogue, and improve accountability. In this regard, collaboration with central government agencies (ISTEEBU) should be promoted. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 13 Chapter 1 Understanding the Context of Decentralization in Burundi Introduction: The Political and ing the demobilization of former combatants, rehabilitation of security services, improved Macroeconomic Context public financial management, and establish- 1. Since the 2000 Arusha Peace Agree- ment of elected communal officials—Burun- ment, Burundi has made progress toward re- di remains one of the poorest and most frag- storing peace and political stability. The Peace ile countries in the world. Even though the Agreement initiated the end of hostilities and security and macroeconomic environment has helped launch democratic election processes. improved significantly in recent years, the situa- From 2000 to 2005, Burundi made a transition tion remains fraught with difficulties. In the past to a multiparty system of government as well decade economic growth has been slow, averag- as a transitional government to share power ing about 3 percent in the first eight years but in- between the country’s two main ethnic groups. creasing to 3.9 percent in 2010 and 4.5 percent Peace agreements have b been signed with all in 2013. Burundi’s gross domestic product per the rebel movements, respectively by the end of capita of $280 in 2013 is still far below the gross 2003 (with CNDD-FDD) and in 2007 (with the domestic product (GDP) per capita of $286 be- FNL). In February 2005, the country approved, fore the crises in 1993. Even though the country by referendum, a new constitution emphasizing is endowed with natural resources and a large power sharing and the protection of minority network of rivers, economic growth is held back rights. Multiparty competitive democratic elec- by several factors including poor diversification, tions have been organized in 2005 and 2010. lack of adequate infrastructure, an unattractive business climate, and a low-capacity labor force. 2. Despite significant progress in the The increasing cost of living, triggered by recent consolidation of peace and stability—includ- volatility in fuel and basic commodity prices, 14 have placed severe strain on the poor. Furthermore, there is pres- Figure 1.1 Burundi Governance at a Glance, 2012 Data sure on land and natural resources due to a high population growth Government rate (2.6 percent in 2011) and high effectiveness population density (300 inhabit- Burundi ants per square kilometer). In par- Voice and ticular, scarcity of land and com- accountability Rule of law petition for land ownership under Sub-Saharan conditions of weak land rights is Africa a continued underlying driver of Regulatory Control of conflict and fragility at the local quality corruption Low income level. countries Political 3. Burundi still faces signif- stability icant development challenges, and is unlikely to meet most of Source: World Governance Indicators, 2011. the Millennium Development Note: Percentile ranks (0–100). Goals (MDGs) by 2015. Prior to the conflict, the percentage of peo- the Government of Burundi (GoB) from ple living below the national poverty line was effectively reducing poverty and improving 36.4 percent (1990). During the conflict, this in- human development. The 2012 World creased to 68 percent (in 1998). By 2006 (last Governance Indicators (figure 1.1) provide a household survey) the rate was still high: 67 quick snapshot of the challenges confronting percent of the population was living below the Burundi. The country lags well behind the poverty line.10 Education and health outcomes African regional average on most dimensions improved significantly over the past few years. Life expectancy rose from 43 years in 2000 to of governance. In particular, Burundi ranks 50.4 years in 2011. But access to basic services in the lowest tenth percentile on government and infrastructure such as potable water and effectiveness, the rule of law, regulatory electricity remain limited (for example, less quality, and control of corruption. The country, than 3 percent of the population has access to however, performs relatively well on Voice and electricity). Moreover, high population density Accountability, in itself a major achievement in and growth rates—with about 90 percent of the a fragile post-conflict setting. This is also in line population living in rural areas—increase com- with the average performance of low-income petition for scares resources, especially land. As countries. a result, Burundi still ranks low in the United 5. A recent Public Expenditure Review Nations Development Programme’s (UNDP’s) (2013) emphasized the fiscal fragility of the Human Development Index (178th out of 187 country and its structural vulnerability to countries in 2013) and is unlikely to meet most external shocks. Burundi is one of the most of the MDGs by 2015. vulnerable countries in the world to external 4. The lack of progress on several shocks, including food and energy prices dimensions of governance has constrained fluctuations as well as aid inflows variations. This fragility has been exacerbated by a number of domestic factors such as the lack of transparency 10  The results from the latest household survey will be available in some government spending, the high level end-2013 and provide updated poverty figures. of tax exemptions—counting for as much as 3 Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 15 percent of GDP—and limited option to absorb postconflict countries in the region (including any fiscal shocks outside of external assistance Sierra Leone and Rwanda, among others), and/or Central Bank’s financing. As indicated where significant portions of public resources in the report, the government will continue and responsibilities have been transferred to to struggle to create fiscal space and reduce subnational governments. Previous studies debt from its currently high level. Although have focused on technical and structural factors the government is expected to limit current in attempting to identify obstacles that have expenditures growth in real terms, unless prevented decentralization from unfolding more additional revenue measures are introduced successfully in Burundi. Some explain delays to offset the falling grants, it will struggle to in transferring competencies as the result of reduce its deficit. While the deficit is expected insufficient institutional capacity, arguing that to be largely financed by external concessional communes are simply unable to offer the types loans, debt will remain high as a share of GDP, of services that might be assigned to them and will likely keep the country at a “high risk of under new legislation. Others have emphasized debt distress.” the low tax base and weak financial viability of communes, which denies subnational authorities 6. Within this context, the GoB has the resources required to generate economies embarked on a potentially transformative of scale that would enable them to provide process of decentralization, with the aim goods and services more effectively (Nihangaza to strengthen social cohesion, improve 2011). Others focus on the lack of clarify in the local governance, and promote access to assignment of expenditure responsibilities and basic infrastructure and service delivery. the absence of an effective system of public The process initiated with the Arusha Peace financial managements (EU report; Senate Agreement was successfully concluded in Report 2011). February 2005 with the promulgation of a new constitution and the installation of a 8. To complement previous analytical democratically elected government in August efforts, this chapter shifts the focus toward 2005. The newly elected government revised the an analysis of the key institutional and Electoral Code prior to communal elections, and political drivers underlying the design and later adopted a new Communal Law governing adoption of decentralization reform in the main level of devolved entities at the Burundi. Drawing upon a background paper municipal level in the form of 129 communes on the “Political Economy of Decentralization in (116 of which are outside the capital and thus Burundi,” the chapter will review historical and mostly rural), with the first communal elections recent developments and map the key actors taking place in June 2005.11 and institutions to better understand the main drivers of decentralization in Burundi, as well 7. Despite these ambitious goals, as their impact on the specific “trajectory” by Burundi’s decentralization experience has which it unfolds. Specifically, the chapter seeks confronted a number of challenges, failing to to provide an overview of the development of meet popular expectations and substantially the decentralization process since the end of the improve access to basic social services at the civil war and clarify (i) why a decentralization local level. As a result, the institutional reform program has been adopted in the first place, (ii) process has proceeded less quickly than in other how it has been designed, (iii) what might prevent it from achieving its stated policy objectives, and finally (iv) what opportunities exist to “unlock” 11  In 2005 the first postwar elections were held, which were the reform agenda and move the process forward. won by CNDD-FDD. Other big parties were FRODEBU and UPRONA (Uvin 2008). People elected the Colline Councils, the commune This overview will set the stage for the analysis council, and the national parliament. of the fiscal sustainability of the current system 16 (chapter 2), the opportunities and challenges to priority to promote social cohesion and provide land registration as key public service deliver a “peace dividend” to the poor. The for the rural communities (chapter 3), and the political economy of the civil war—with its most effective strategies to promote citizens’ deep-rooted urban-rural divide, as well as inter- engagement and social accountability through elite divide coupled with an interethnic political decentralization (chapter 4). struggle—deeply affected the collective memory of the population, generating large popular 1.1 The Decentralization Process demands and pressures to the postconflict since the End of the Conflict: government for defusing social tensions by reestablishing the state presence in rural areas. An Overview This urgency to deliver a rapid “peace dividend” 9. Despite early efforts to establish a throughout the country implied that effective decentralized model of governance, little state-capacity would have to be built in parallel progress was achieved in the early 1990s, at the central and local level. It was within this given endemic political instabilities. During context that the GoB launched a decentralization the late 1980s, several decentralization laws process to empower communes to improve and decrees12 confirmed government efforts local governance and social cohesion. The view to move on with the reform agenda. But the of decentralization as a national priority for period was marked by waves of ethnic-political Burundi peace-building efforts was also widely violence that undermined the reform agenda shared among the international community, and eventually led the country to a civil war. In with several donors providing technical and 1993 the first democratically elected president financial support to launch the process. (Melchior Ndadaye) was assassinated by high 11. The promotion of a new legal military officials. It took one year to form a new framework confirmed the government government; then, tragically, the new (Hutu) commitment to the reform process. In 2005 president (Cyprien Ntaryamira) was killed in a the Communal Law and Electoral code further plane crash in Kigali. This event fueled increased confirmed the commitment of the Central ethnic violence between the Burundian Armed Government to open the way for more direct Forces and numerous armed (rebel) groups. participation and accountability at the local level. Patterns of exclusion and interethnic competition The law recognizes communes as autonomous over control of state resources acted as the main and decentralized entities managed by driver of violence and war. (Insight on Conflict, communal council members. It also recognizes 2010) the communes’ deliberative and decision- 10. At the end of the conflict, the Arusha making powers, by allowing them to adopt Peace Agreement and the 2005 Constitution economic and financial measures, including the laid the foundations for an ambitious preparation of the communal budget and the decentralization program as a national communal community development plan (Plan Communal du Développement Communautaire, PCDC). The 2005 Electoral Code established the number of communal council members (25— 12  The decree-law No. 1/017 of June 1988 to transfer certain administrative revenues to communes as amended to date; changed to 15 in 2009)13 to be elected through the decree-law No. 1/026 of 21 July 1989 to transfer property taxes collected in the territory of Burundi to the communes and the Mayor of Bujumbura; the decree-law No. 1/027 of 21 July 1989 to transfer tax on rental income (impôts foncier) received 13  The Electoral Code was amended in 2009 with the introduction in the territory of Burundi to the communes and the Mayor of several changes, notably the reduction in the number of the of Bujumbura; and the law No. 1/004 of 23 March 1994 on the Communal council members from 25 to 15 to better balance the general organization of the communal administration. cost of financial incentives—jetons de presence—given to council Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 17 Box 1.1 Key Legislation The Arusha Accord for Peace and Reconciliation in Burundi indicates (Art. 8, al.2 IIème Protocole) that: “La commune forme l’unique collectivité territorial décentralisée” (communes are the only decentralized entity) while Art. 20, al. 13 of the same Protocole II stipulates that communes (municipal entities) and collines (hills which can be seen as wards) are administered by councils (conseils communaux and conseils collinaires), with both elected by direct voting. The 2005 Constitution of Burundi lays the rules for the decentralization process in its title XI “Collectivités locales” (local communities) where the organization of the central and local administrations are described. Title XI in its section on Communes recognized the Commune as a decentralized body and states that: “the Commune is a decentralized administrative entity and is divided into entities provided by an organic law.” Moreover, the Constitution states the organization of the communes as: “the commune is to be administered by the Communal Council and the Communal Administrator.” Title XI also include another important provision (article 267) stating that: “L’État veille au développement harmonieux et équilibré de toutes les communes du pays sur base de la solidarité nationale” (the Central Government is responsible for the development of all communes within a context of national solidarity; could be interpreted as justifying some form of equalization). Source : Le Sénat et la Politique de décentralisation au Burundi août 2010 partie IV. direct universal suffrage within each commune, implementation of decentralization reform and regulated the election of the communal has also been shaped by the constant—and administrator among the councilors (via increasingly coordinated—technical and secret ballot) and nomination by a Presidential financial support received from key players decree. within the international community, namely from the World Bank, European Union (EU), 12. In 2007 the government created the German Agency for International Cooperation National Fund for Communal Investment (GIZ), and Swiss Agency for Development and (FONIC) to address the resource constraints Cooperation (SDC).14 of communes. The decree that established FONIC stated that 15 percent of the national 13. In 2009 a National Decentralization budget should be transferred to the institution Strategy (NDS) was approved, providing the to channel funds to the communal level, but no basis for the institutional and regulatory analytical reasoning seemed to underpin this framework guiding the decentralization percentage. Moreover, these transfers have not reform process. This policy framework was been realized and FONIC remained until recently an institution with little means to function as designed. As a result, communal budgets remain 14  In particular, the Bank’s CDD and decentralization project, very limited, and investment resources are PRADECS (2007–12) represented one of the key instruments of the coordinated donor response to support decentralization, mainly channeled via donor-funded grants and with a key focus on social service delivery, capacity-building of projects. In part as a consequence of persistent communal level officials, and support to develop the first communal development plans in the country. GIZ provides support to the funding gaps from the central government, the Ministry of Communal Development and decentralization in key provinces, as well as taking the lead on the FONIC reform process. Swiss Cooperation provides similar support, as well as taking the lead on land issues. The EU, through their project Gutwara Neza members with the poor financial resources of communes. A (“good governance” in Kirundi) focuses on capacity building and gender dimension has also been streamlined into the revised law technical assistance to communal officials, particularly around requiring 30 percent of the 15 council members to be women. themes on accounting, budgeting, and procurement. 18 Figure 1.2 A Timeline of the Decentralization Process in Burundi, 2000–15 2005 New constitution 2010 2000 2009 First national 2nd Local Arusha Peace Amendement and local & National Agreement to the elections Elections Electoral Code 2005 2007 2010 The communal Creation of Amendment 2015 Law and the FONIC to the 3rd National & Electoral Code Communal Local Elections are adopted Law Note: The events listed are major historical landmarks that have shaped the process of decentralization in Burundi. then operationalized the same year through pillars identified.16 the creation of a three-year action plan by the Ministry of Communal Development. The 1.2 The Decision to Decentralize: action plan focuses on five pillars, including: Understanding Key Incentives (i) strengthening the legal framework for 14. Standard, normative arguments about decentralization; (ii) strengthening the the benefits of decentralization focus on issues institutional framework; (iii) capacity building surrounding the efficiency of service delivery. of actors in the sector and the articulation of a The basic idea is that decentralization, by communication policy; (iv) the effectiveness of making representatives closer to the people they local government in promoting local economic serve leads to socially more desirable service development, poverty reduction, and the provision, either because local politicians are provision of social services; and (v) fiscal more accountable or because they have better decentralization.15 In April 2012 an overall information about what needs to be done or review of the decentralization process was what people’s preferences are. Arguments that conducted to assess the achievements of the link decentralization with better service delivery first action plan and lay the groundwork for may be particularly compelling in situations a second set of action plan for 2012–14. The where there are ethno-regional divides over new action plan adopted by the government which policies to follow or public goods to was grounded in the previous National Policy supply (see Robinson [2010b: 1–2]; Bardhan Document and proposes new strategies to [2002]; and Eaton, Kaiser, and Smoke [2010] for advance the decentralization policy on the five a review of this argument). 15  République du Burundi, Ministère de la Décentralisation et 16 République du Burundi, Ministère du Développement du Développement communal, Document de Politique Nationale de communal, Plan d’action triennal 2012-2014 de mise en œuvre de la Décentralisation du Burundi, Bujumbura, mai 2009. la politique de décentralisation, Bujumbura, Septembre 2012. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 19 15. In the case of Burundi, however, the struggle for control and access to economic role played by standard normative arguments resources and maintenance of networks of was probably marginal and subordinated to clients. The events associated with the recent political objectives. More likely, the political political crisis have created some splits within interests of national politicians in the ruling the ruling coalition and have somehow remapped coalition government did more to determine the political configurations at the local level. the government’s decision to decentralize. This might contribute to explain the resistance To some extent, those interests might have of central government to further devolve been the result of pressures from the donors, power and delay the approval of key pieces of who pushed for decentralization to enhance legislations to promote fiscal decentralization political stability, promote state building, and (see chapter 2 for technical analysis), likely after ultimately address the root causes of the war. the 2015 elections. But there are reasons to believe that the interest of national government in decentralization was 1.3 Institutional Design: Mapping equally strong—although shaped by different Key Actors incentives. 18. The territorial organization of 16. The national government seems to Burundi is two tiered. There are 129 have perceived decentralization as a good communes that are formally divided into opportunity to meet popular expectations the 13 communes of Bujumbura—capital while—at the same time—establishing and the largest city in Burundi— and the and maintaining political control in the 116 communes outside Bujumbura. The countryside. This is not surprising, given population of Bujumbura is about 650,000 that a key source of popular discontent prior (average commune = 50,000) while communes to the war was the marginalization of rural outside Bujumbura have on average 90,000 communities and the centralization of economic inhabitants in 2014. The communes of and political power in the hand of urban elites. Bujumbura have less autonomy than the other Another potential motivation—decentralization communes. We thus distinguish between them as a strategy of political survival at the local in our analysis. There are also 17 provinces18 level17 (O’Neill 2003, 2005)—does not seem to which host deconcentrated entities of various be relevant in the context of Burundi. Historical ministries and a governorate at the provincial evidence suggests that the ruling coalition when level. One of them, Bujumbura—Mairie, is decentralization was introduced did not fear to both a province and a municipal body acting lose power in the next (2010) national and local as a commune. Table 1.1 presents information elections, and was instead confident of winning. on the size of communes by province outside 17. Recent political developments might Bujumbura. change the balance of power and shift 19. Each commune has a deliberative central government incentives relevant to body (the communal council), elected decentralization. In Burundi, like in other through direct universal suffrage by its countries, the political struggle is also the citizens; an executive body (the communal administrator), elected through indirect 17  For instance, O.Neill (2005) argues that in Bolivia the ruling party, anticipating that it would lose power at the national level, but confident of doing better in local elections, introduced local 18  Bubanza, Bujumbura Mairie, Bujumbura Rural, Bururi, elections and decentralized resources as a way of maintaining Cankuzo, Cibitoke, Gitega, Karuzi, Kayanza, Kirundo, Makamba, some amount of power and control. Muramvya, Muyinga, Mwaro, Ngozi, Rutana, Ruyigi. 20 universal suffrage by and among the nistrative capacity. The current organizational municipal councilors; and administrative structure (see figure 1.3) does not have a top services provided by two technical advisors. communal civil servant that would be the Chief Communes are then divided into administrative Operation Officer of the commune. This function units known as “zones”; these are a grouping is carried out by the communal administrator,19 of collines (hills) in rural parts of the country who is simultaneously an elected official, the top (neighborhoods in urban areas). These hills local civil servant, as well as the representative are the basic natural grouping in Burundi. They of the central government in the commune. This person receives a base salary from the have a five-member council elected directly by central government and a supplement from the the population; the councillor with the largest commune. Two technical advisers are selected number of votes becomes the local chief (chef by the communal council among central civil de colline). Moreover, colline councillors run servants paid and made available by the central as independents (they are on no party list). government to the commune.20 The communal Tasked to represent their residents and voice their concern, councillors do not deliver public services but their head may act as a tax 19  Loi 1/02 2010 portant sur l’administration communale art collector. 25 et 30. 20  In 2009 the cost of the two technical advisers and the mayor is 20. The current organizational struc- reported at 420 million FBu; p6 Mécanisme de Transferts Financiers ture of communes reflects their weak admi- Aux Communes (Fonds National d’Investissement Communal) by Table 1.1 Information on the Size of Communes, Rural Provinces, Burundi, 2008 Standard Standard # Average area Average Smallest Largest Provinces deviation deviation communes (km²) Population population population (km²) population Bubanza 5 21 070,2 7 559,8 6 7604,6 14 071,2 55 344 83 678 Bujumbura 11 10 382,2 2 877,8 50 539,4 19 402,8 24 660 78 823 rural Bururi 9 26 608,1 6 028,4 63 779,2 33 824,1 29 685 145 074 Cankuzo 5 37 855,2 13 933,8 45 774,6 9 505,7 32 458 57 322 Cibitoke 6 26 103,7 6 264,4 76 739,2 8 228,6 66 367 88 451 Gitega 11 17 285,8 6 533,5 65 094,9 36 134,2 32 714 155 005 Karuzi 7 20 200,3 3 313,8 62 349,0 16 324,6 41 476 81 938 Kayanza 9 13 145,4 3 698,7 65 045,8 162 61,6 39 633 97 252 Kirundo 7 25 448,4 11 818,3 89 750,9 28 618,6 57 326 145 424 Makamba 6 31 521,3 8 395,4 71 816,5 30 103,1 44 978 112 752 Muramvya 5 13 341,4 4 684,9 58 517,8 15 662,4 41 170 81 257 Muyinga 7 25 679,1 9 296,0 45 231,1 11 637,3 27 954 60 311 Mwaro 6 13 700,0 2 350,7 98 155,7 39 037,9 54 482 151 230 Ngozi 9 15 698,4 4 454,8 73 413,0 20 964,1 57 086 120 557 Rutana 6 30 865,3 12 785,2 55 585,0 19 936,2 35 471 93 365 Ruyigi 7 32 216,3 9 630,7 57 218,6 20 754,0 37 502 99 461 Source : Author’s compilations, using data from IGEBU and ISTEEBU. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 21 staffs paid through communal revenues 22. The communal administrator— comprise zonal leaders (chefs de zone), one hereby referred to as the administrator—is communal secretary, one civil registrar, and one appointed by presidential decree before communal accountant. A problem often cited by taking office, although he or she is elected central government officials is that the quality among the communal councilors through of the personnel varies between communes. a secret ballot within the council.22 The In particular technical advisers are often local administrator is the legal representative of the teachers not trained for the duties of technical commune and its population. As a civil servant advisers. working fulltime and remunerated by the state (base salary) with communal supplement, the 21. Communes in rural provinces administrator is responsible for the management are governed by a communal council of and administration of the communes, the 15 members elected by direct universal application of laws and regulations, public suffrage, based on a list of candidates from security, domain management, and the different political parties (block list approved preparation and execution of the budget. In by the party) reflecting a gender balance (set at terms of local development, the Communal 30 percent by the Electoral code) and the ethnic Law assigns to the administrator the task of balance within the represented commune. The preparing “the communal development plan and national independent electoral commission community development” (PCDC) and managing (CENI) can intervene ex post to ensure gender its implementation process while consulting and ethnic balance on the council.21 Council with the local population. He regularly reports to members serve a five-year term. Within a week the Communal Council as well as the respective after the elections, the newly elected commune tutelle authority. He is also mandated to submit council must have its first meeting, during an annual report on the management of his/ which its president, its vice president and its her commune to the communal council and secretary, and the communal administrator to the governor of the province, which will be (that is, the mayor), are elected through secret made public after adoption by the communal ballot. The communal council regulates all the council. The administrator delegates some of internal affairs of the communes including his/her power to the zone manager. The zone is local development. For instance, it sets the a decentralized administrative district between community development program, monitors the commune and the collines/neighborhoods. its implementation and ensures evaluations 23. To support the overall management are carried out. Moreover, through its of the communes, the administrator and administrative branch, it promotes employment the commune council are assisted by and handles personnel recruitments as well two technical advisors. These are central as all transactions pertaining to the operation government employees seconded to the of the commune. Oversight authority over the commune (détaché auprès de la commune). Council is exercised through a tutelle system They are selected by the communal council. One by both the provincial authorities (through the advisor is in charge of local development, while Governor) and the central government (through the other is in charge of social issues (health, the Ministry of Interior). 22  The communal law provides for cases in which an Mohamed El-Mensi and Déo Banderembako LDi/UNCDF 2009. administrator may be removed from office or suspended by the central government or the commune councilors for valid reasons 21  Plus a councillor from the Twa group if one ran but was not including incompetence, corruption, misconduct, or misuse of elected. funds. 22 Figure 1.3 Institutional architecture of decentralization in Burundi Presidency and Vice Presidencies Sectoral Ministry of Ministry of Citizens Ministries Communal Dev. Intérieur Governor Provincial departments (Education, health, etc.) Conseil Communal Tutelle (15 membres) Communal committee on Communal Administrateur Communal (1) communal dev. units Commune Technical Head of the Zone Colline/ Personnel advisors (2) rural/ urban (1) neighborhood council (5) Colline committee Head of colline/ on development neighborhood (1) Legend : Designates the “tutelle” role of the Ministry of Interior Designates the policy oversight and coordination role of the Ministry of Communal Development Designates (hierarchical) relationships of accountability. Note: The chart above (figure 1.3) identifies the institutional architecture of decentralization in Burundi, mapping all key actors, institutions and relationships between central government, deconcentrated structures and decentralized authorities. The black borders identify central government agencies in charge of policy oversight, coordination and tutelle over the communal administration. The red borders identify deconcentrated structures of the central public administration. The yellow borders represent decentralized actors and institutions at the communal and colline level. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 23 Box 1.2 Legal Provisions Regulating the “Tutelle” System Control over communal decisions (“mise sous contrôle” or ‘tutelle”) is already reflected in Articles 96, 98, and 99 of the Communal Act of 25 January 2010: “The Minister in charge of territorial administration in its attributions may cancel any regulations or resolutions of the communal authorities which is beyond their duties or are contrary to law or to the public interest. The cancellation must be made within thirty days after the date on which the Minister became aware of such resolution.”23 Moreover, Section 100 of the communal law reiterates that: “When Local authorities are failing to perform the duties bestowed upon them under the laws and regulations, the Minister in charge of territorial administration, the provincial Governor or the Mayor may, after two warnings, take any actions to replace them.”24 education, assistance for the poor, and so on) by a governor, who is supported by three as well as cultural and sporting activities. In the councilors (appointed by decree) and is case of the absence of the Administrator, these responsible for the design, coordination, advisors can step into his/her role. and monitoring of development activities at the provincial level and who also holds 24. The Committees on Communal an administrative and financial tutelle over Development (CCDCs) are formed on a communes (see box 1.2). A provincial committee sectoral basis and include representatives on communal development (PCCD) is also of deconcentrated government entities to established. Chaired by the Governor who is the communes. The CCDC is an advisory body supported by a principal advisor, the PCCD composed of individuals involved in community consists of two representatives per communes development and established by the communal appointed by the Minister of the Interior council based on a request by the administrator. to support the provincial administration in Its role is to provide technical expertise to local providing advice on administrative and policy authorities. The CCDCs are usually involved in issues including those relating to decentralization the planning of community development. and the development of communes. 25. The colline, or the neighborhood, is 27. Key sectoral departments such as the basic entity of territorial administration. agriculture or education have both provincial Similar to the communal level, the Collines/ and communal deconcentrated branches neighborhood council is support by an advisory (direction provincial and direction communal). body, which, however, is not recognized by the Their activities vary across sectors. For example, law: the Conseil de Développement collinaire in education the communal branch supervises (CDC). It is composed of all five members of the schools offering primary education (école the collines council, as well as seven persons de base et fondamentale) and plays a role in the selected by the communal general assembly for recruitment of teachers and the nomination of their commitment to community development principals while the provincial branch acts more activities and conflict resolution. CDC has as an interface between the central departments among others mission the role to participate and the communal branches. In both cases, they in the development of collines/neighborhood are assisted by an advisory council. community development plans which establish the priority needs of the collines/neighborhoods and are to be integrated into the PCDC. 23  La loi communale du 25 janvier 2010, Art.99, p.25. 26. The provincial government is headed 24  Ibid p. 26. 24 1.4 Institutional Design: La Décentralisation exige une nouvelle culture étatique, un nouvel état d’esprit Promoting Citizens’ Engagement politico-administratif… Tout le monde, to Strengthen State-Society pas seulement les élus locaux comme Relationships on a tendance à le croire un peu partout, est acteur de la Décentralisation. 28. The first defining design feature of the new legal framework was the promotion [Decentralization] favors a “silent revo- of an open partnership and collaboration lution” which brings about a qualitative between the state and citizens. The 2005 change in society. Decentralization decentralization initiative does hold some demands a new state culture, a new similarities to former historical attempts. politico-administrative spirit within For example, locally elected councils led by the state . . . Everybody, not just administrators who, although now elected by locally elected leaders as we all are and amongst the councilors themselves, continue inclined to think, is an actor within to be accountable to the Provincial Governor Decentralization. and central authority.25 However important (Government of Burundi 2009: 56—translation differences also exist. These relate to the overall and emphasis by Gaynor [2011]) aims and aspirations of the new power-sharing 29. A second design feature that departs arrangements, and are clearly articulated in the from the ethos, policy, and practice of National Policy on Decentralization adopted in decentralization in the pre-war history of 2009 (Government of Burundi 2009). Two key Burundi is the active role accorded to citizens aspects in this regard are worth noting. The in the process. Given that the decentralization first is the fundamental change in the political process aimed at the interconnecting areas of culture required wherein elected and appointed governance and development, citizens now have state officials begin to work in open partnership a role to play in both determining development and collaboration with local citizens rather than priorities and overseeing initiatives to address exercising centralized control as in the past. This these priorities. According to the government’s is described as requiring nothing less than a texts: “silent revolution”—a complete bouleversement of traditional relations (revolution) in the La Décentralisation vise l’objectif de la absence of violent conflict (silent).26 participation active de l’ensemble de la population à la définition et à la mise [La Décentralisation] favorise une en œuvre des politiques de développement « révolution silencieuse » qui engendre économique et social de leur localité. un changement qualitatif de la société. Les résultats attendus d’un processus de Décentralisation sont d’une part le développement local et communautaire, 25 See OAG (2007), OAG (2009), and Gouvernement de Burundi et d’autre part la démocratie locale et la (2009) for more on this point. bonne gouvernance. 26  As indicated in Gaynor (2011): “the term “silent revolution‟, when cited during the research feedback presentation generated Decentralization aims at the active a lengthy debate as to what this meant and whether the word participation of all the population in “revolution” was appropriate in the Burundian context. It was noted that this was the state’s own wording (appearing as it does defining and implementing economic in the Government’s National Policy) and that the use of the word and social development policies in their reflected the enormity of the political, attitudinal and behavioral localities. The envisaged outcomes from a changes required in how local state and civic actors relate to each other and conduct affairs in an open, transparent and participatory process of Decentralization are, on the one manner.” hand, local and community development Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 25 and, on the other, local democracy and ruled no longer consist of ruling over but ruling good governance. with. These transformations, as clearly set out in the Government’s 2009 National policy, certainly (Government of Burundi 2009: 10—translation and emphasis by Gaynor [2011]) require nothing less than the “silent revolution” called for in the same policy. 30. There is a radical shift in state-citizen engagement, from consultation to active 32. A review of the structures and proce- participation. Within the national policy, it has dures as they are set out in the relevant been emphasized that citizen participation is not legislation of 2005 and 2010, as well as within merely about consulting local citizens on their the Manual of Administrative and Financial views and ideas but instead moving toward the Procedures published by the government in higher end of the participation spectrum, where 2011 and found in box 1.3, highlights the current citizen participation within local governance design to strengthen state-society relationship in Burundi involves shared decision-making at the local level. between state officials and citizens in relation 33. While several opportunities exist for to development programmes and interventions accountability and participation, these legal within their localities. provisions fall short of shaping daily practices Le Gouvernement de la République on the ground. The current decentralization du Burundi a pris l’option politique framework as enumerated in box 1.3 offers de rapprocher les services publics de numerous opportunities for citizen oversight la population et d’impliquer cette and engagement in the process. Figure 1.4 gives dernière dans la prise des décisions a clear picture of such process as envisioned by et le choix des programmes et projets de the state. At the local level, opportunities for both développement de leurs collectivités. accountability from state actors (highlighted in red), as well as openings for participation The Government of the Republic of from civic actors (highlighted in blue) exists. Burundi has taken the political option In theory and institutionally, a sound structure of bringing public services closer to the has been devised. In practice however, these people and of involving the people in opportunities are not exercised. For example, decision-making and choice in relation to the two public consultative meetings chaired development programmes and projects in by the communal council each year focus on their localities. accountability—literally referred to in Kirundi as “accountability meetings”—do not garner (Government of Burundi 2009: 61—translation and emphasis by Gaynor [2011]) wide participation. The framework below, in theory, could be very effective at bolstering 31. The call for a “silent revolution” in social accountability at the local level. There the way state and citizens interact emerged is a clear division of mechanism among the as a clear policy objective of the new communal council, the administrator, and the hill decentralization process. As enumerated council (conseil de colline), as well as a balance in the Government’s new decentralization between opportunities for participation and policy, decentralization within Burundi today opportunities for greater communication. The represents a fundamental shift from previous main question worth investigating is how does attempts to this process. This transformation this policy translate into procedure and practice embodies a shift from top-down to bottom-up on the ground in the communes and collines planning where citizens are no longer considered throughout the country? In practice, evidence subjects but partners in governance and where suggests that various factors contribute to act as the societal divide between the ruler and the binding constraints to mobilize effective citizens’ 26 Box 1.3 Current Decentralization Procedures in Burundi The larger objectives and purposes articulated in the 2009 policy document find more concrete expression in the associated texts produced over the last few years—notably the Decentralization Law of 2005 and the Manual of Administrative and Financial Procedures of 2011, both produced by the Government of Burundi (Government of Burundi 2005b, 2011). A number of the key features of both the commune and hills councils (conseils de colline), deriving from these associated texts, are outlined below: Procedures for commune councils • The commune council comprises 15 members elected through proportional representation from party block lists of candidates. • Co-options are permitted to achieve power-sharing quotas—no more than 67 percent from any one ethnic group; no less than 30 percent female representation. • The role of the commune council is to ensure that public services respond to the needs of the population (Article 5*). • The commune council meets three times a year in ordinary session where a quorum of 10 or more members is required (Article 10). • These meetings are public (Article 19). • Before March 31st each year, the communal administrator presents a progress report to the council. The report is sent to the Provincial Governor and is made publicly available following validation by the communal council (Article 31). • Decisions at council meetings are by simple majority, public voting or, for sensitive issues, by secret ballot of all members (Government of Burundi 2011: 15). • A record of deliberations of each meeting is kept and extracts from all deliberations are posted on the public notice board at the commune offices (Government of Burundi 2011: 15). • Twice a year, the council holds open information sessions with hill council members and members of local associations. Participants at these meetings may pose questions and propose ideas and solutions to the commune council (Article 15). • The commune council is responsible for appointing a consultative committee (following propositions from the Administrator) to advise on priorities for the communal community development plan (PCDC) (Article 38). • The Administrator presents a bimonthly progress report to the commune council on implementation of the PCDC (Article 39). • The annual budget is adopted by the council and transmitted to the Governor for approval, at the latest by October 31st (Article 58). • The transfer of responsibilities is accompanied by a transfer of the necessary financial and human resources to carry these out (Articles 77 and 71). Procedures for hill councils • The colline council comprises five members elected on an individual first-past-the-post basis. The candidate with the most votes is the head of the council / chef collinaire (Article 34). • The colline council meets three times a year following notification from the council head (Article 35). • The role of the colline council is . . . (Article 16): -- to propose to the commune councils actions leading to development and the preservation of peace; -- to mediate / arbitrate on conflicts between neighbors; -- to advise the commune council on possible projects for the colline; -- to monitor the implementation of commune activities on the hill. • At least three times a year, the head of the council organizes an open meeting for all colline residents to analyze the political, social, economic, and security situation on the colline (Article 37). Source : Trocaire Report 2011. * All articles cited in this section refer to articles from the relevant law (Government of Burundi 2005). Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 27 Figure 1.4 Opportunities for Engagement at Different Levels of Burundi’s Decentralization Process COMMUNE COUNCIL Annual Budget (15 Members) Meets 3 times a year (Open to public) Public Consultative Development meetings twice a Committee year public can ask (consultative body to questions) council) Minutes y compris les decisions sont affichées sur panneaux publics Administrator Biannual Report on Annual Report to commune community Council dev. Plan Colline Council Head of Colline Council (5 members) Development committee Council meetings 3 times a year Legend: Open/Public meeting Institutions 3 times a year Accountability Mechanisms Participation mechanisms Source: Trocaire report (2011) adapted by the authors 28 participation and downward accountability. human and financial, to communes necessary (The main factors accounting for the observed for its proper implementation. The law on situation are analyzed in chapter 4.) communal finances has not been submitted to parliament (still being prepared by the Ministry 1.5 Obstacles to Moving of Interior) while the law on a territorial civil the Reform Agenda Forward: service (defining the legal status of communal staff) is not much more advanced and the law Understanding Institutional revising the role and of local administrator is Constraints and Political equally on hold. Finally, the law on revising the Incentives legal status of Bujumbura Mairie is also under discussion. In addition, all these texts will need 34. Since 2010, when a few amendments to be harmonized and made consistent with to the 2005 communal law were introduced, other legislation governing local development there has been little progress toward and planning (Environmental Code, Land Code, effective decentralization in Burundi. On Forest Code, and so on) as well as legislation the one hand, the Government has recently governing deconcentrated administrative adopted important legislative documents unites. 27 such as the Decree on intercommunality; the Decree on Burundi’s decentralized cooperation; 35. Institutional constrains combined the Decree establishing the Interministerial with political and bureaucratic incentives Orientation and Steering Committee for are critical obstacles to significant progress. Decentralization in Burundi and the Technical In fact, the uncertain political dynamics may Monitoring Committee; the Decree on the suggest that the introduction and implementation Reorganization of the FONIC; the Order of some of the new legal provisions might now establishing a Technical Monitoring Committee be perceived as politically risky by central of the Comité Interministeriel d’Orientation et de authorities, as it might strengthen the autonomy Pilotage de la Décentralisation (CIOPD). On the of communes and reduce their dependency on other hand, many basic legal texts to support the decentralization process have yet to be adopted. Notably, the law on the responsibilities 27  This is the case, for examples, of bills that have been initiated by the Ministries of Education and Health and created of communes remains to be adopted since new local structures (departments of health districts, municipal parliamentarians deplore the lack of transfers, departments of Education). Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 29 the center. The section below reviews these opment.28 However, currently the Communes incentives and further articulates on the effects are expected to fulfill the missions of multiple that a changing political environmentt might ministries. Moreover, as discussed above (para have in reshaping the institutional framework 21 and 22), the communes fall under the of decentralization. tutelle of the Ministry of the Interior as well as the provincial authorities (Governor). This 1.5.1 Institutional Constraints “fragmented accountability” and overlapping 36. Institutional factors seem to have oversight creates a functional duplication, which played a key role in the delays, hesitations, in turn generates rivalry and competition over and contradictions that characterize the leadership between ministries. For example, process of promoting the legal framework while the National Policy Strategy assigns a for decentralization. At the institutional leading role to the Ministry of Decentralization level, the lack of a strategic vision is often and Communal Development,29 in practice the reflected in the fragmented legislative process, Ministry of the Interior plays a leading role in with several actors within and outside the regulating the competition among communes government (including parliament, local actors on local development projects. In particular, and civil society organizations), lamenting the the daily management of communes and “piecemeal approach” to legislative initiatives, participation in initiatives of political nature once called “party works” and now “community which in turn makes difficult for national policy works,” depends largely on the relationship makers to understand the overall direction and of the communal administrator with central coherence of the decentralization program. government officials. This relationship is more Moreover, ongoing competition between political than technical in nature and shaped by central ministries and vice presidencies over partisan relationship (see section below). the political leadership of the reform causes blockages and mutually reinforcing deadlocks 38. The National Interministerial during the processes of analyzing and adopting Steering and Orientation Committee on legal texts. decentralization (Comité Interministeriel d’Orientation et de pilotage de la 37. The institutional architecture of decentralization –CIOPD) has yet to act as the reform process remains to be clarified, an effective coordinating body and provide undermining efforts to clarify political leadership of the reform process. The leadership and define responsibilities across decentralization policy is transversal and involves ministries. Since 2005 the reform process has been housed in changing departments and ministries, fragmenting reform plans and eventually undermining the overall coherence 28 To achieve its mission, the Ministry in charge of decentralization has created offices namely the Direction of the whole process. It started within the Generale de la Décentralisation and the Mobilisation pour l’Auto- Ministry of Interior and Public Security (2005– développement. General Directorate on Decentralization and the Mobilization for Self-development. These offices have specialized 07) later reorganized as Ministry of Interior technical directions. In addition, the ministry has a funding and Communal Development (2007–09). It then mechanism or transfer of funds to communes through the Fonds National d’Investissement Communal (FONIC) which was created moved to the Ministry of Planning and Communal in 2007 to replace the Fonds de Développement Communal. Development (2009–11), and was finally given to the Ministry of Communal Development 29  Sspecifically, the Ministry has the mandate “to ensure the coordination of all activities related to decentralization, monitoring (2011–13). According to the latest Decree on and evaluation of local communities’ development, to develop the issue (No. 100/323 of December 27, 2011) and implement the national policy on community development, design and implement government policy on decentralization and the mission to lead the decentralization process Communal development and contribute to the mobilization of is now under the Ministry of Communal Devel- resources to implement the policies advocated and so on.” 30 Box 1.4 Sectoral Ministries Involved in the Decentralization Process -- Ministry to the Presidency in charge of Good Governance and Privatization, within the framework of the implementation of the National Strategy for Good Governance and Fight against Corruption under the aspects of control and inspection of the communes, and the development of standards to improve the management, promotion of ethics of good governance; -- Ministry of Interior through the Decree of December 27, 2011, Section 1, Art.13, which assigns the role of designing and implementing the national policy of territorial administration while ensuring the supervision of the decentralization process and monitoring of territorial administration, as well as all services under its jurisdiction, to ensure, in collaboration with relevant ministries, the coordination of territorial technical services and decentralized government services. -- Ministry of Public Security with a mission to execute the national policy on public security must collaborate with local communities, and especially because the Communal Law provides that the communal administrator practices in the territorial limits of its jurisdiction a general police power. The law provides that the administrator has direct line authority over the police jurisdiction assigned in his commune. -- Ministry of External Relations and International Cooperation, which helps in the pipeline projects of decentralized cooperation. -- Minister of Finance and Economic Development Planning intervene to ensure support for joint planning and the financing of communal projects, particularly through the FONIC. -- Ministry in charge for public works and equipment involved in the planning and supervision of rural development in the context of the design and implementation of government policy on road infrastructure and housing. -- Ministry in charge of environment and spatial planning in its duties as a direct manager of the Land Code, the Code of Environment, and Forest Code which important steps are involved in the transfer of skills. -- Ministry of Energy and Mines, which intervenes in the planning and supervision of rural development in the context of hydraulic and electrification. -- All departments with decentralized services in particular the ministries of agriculture and livestock, health, and education. two main ministries (Interior and Communal first, it has been established by a Ministerial Development) while being also fragmented Order to involve actors from the Presidency and across several sectoral ministries engaged in other ministries, which raises issues of lack of local development (box 1.3). It involves the first authority and “inverse accountability.” Second, vice presidency on one hand and the second the CIOPD has no real decision-making powers vice presidency on the other. This poses a huge and its meetings often turn into opportunities problem of overall strategic orientation and for sectoral ministries to push for their own undermines coordination efforts, since not all agenda. ministries involved in decentralization reform are 1.5.2 Political Incentives within the same vice-presidency. To address this problem, in 2011 the Government established 39. Decentralization has been perceived the CIOPD, through the National Policy Paper on by national elite as an opportunity to Decentralization. This mechanism failed to play establish and maintain political control and effective role so far, for at least two reasons: at the local level. As indicated above, the Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 31 political motivation to establish control and selection of candidates has been controlled power at the local level acted as a key driver by national party leadership, which favors in the politicization of the deconcentrated and party activists able to mobilize votes rather decentralized administration, particularly than independent candidates running their with regard to key decision-making figures campaigns around local development issues. (governors, administrators, and so on). The This in turn risks undermining downward primary interests of national elite to use electoral accountability, reducing local decentralization as a system to establish and politicians’ incentives to address popular maintain political control at the local level helps needs. to make sense of specific design features of the decentralization process, as well as resistance to ii. Human resource management has proposals that might change the current political been politicized, undermining the equilibrium. In particular, incentives have shaped institutional capacity of local state and are reflected in (i) choices in the electoral actors. As indicated above, national law (proportional system with a ‘blocked’ lists political parties still consider that locally of candidatesapproved by political parties) as elected officials (communal agronomists, a means to strengthen electoral support and teachers, principals and inspectors of loyalty to the national party leadership; (ii) primary schools, nurses, and so on) are the tendency to reward party supporters to fill effective intermediaries in mobilizing out civil service positions at the provincial or voters. In this context, requests for transfer communal level; (iii) the resistance to change of local public officials are often addressed the status of the Communal administrator; and more as a mechanism to rewards political (iv) the delays in approving the draft laws on supporters than as a management tool to the transfer of competences and resources to initiate a genuine policy of redeployment communes. of competent officials of the public administration to the communes. Overall, i. The choice of the electoral law governing this approach has favored political loyalty local elections is likely to favor party over competence and merit, undermining activists at the expense of independent the institutional capacity of local authorities candidates outside the control of national to deliver. 31 party leadership. Despite the democratic elections of 2005 and 201030 marking a 40. Resistance to changing the status turning point in political decentralization of the local administrator illustrates this and empowering local authorities with point. As discussed above, the communal legitimate power before local communities, administrator is currently stuck between the communal councilors are elected via a status of an elected official and the manager of proportional system in a “blocked list” of the commune as a civil servant who represents candidates presented by political parties. the central government, while the commune Focus group research has confirmed that this mechanism tends to produce a sharp politicization of local elites, since the 31  Lack of technical capacity, however, is not only the result of calculated political choices, and seems also to be the by-product of poor human capital across rural communities. Technical and financial partners have supported the government in this issue, 30  The 2010 elections were characterized by a high turnover including the establishment of a Centre National de Formation of local elites, undermining capacity building efforts supported des Elus Locaux (CNFEL) National Training Centre for Elected under several donor-funded initiatives and contributing to the Officials. But a better coordination with the National School of erosion of previously acquired skills in development planning and Administration (ENA) and other institutions specialized in training management. of civil servants seems urgently needed. 32 does not have the equivalent of the chief 1.6 Recommendations for Policy executive officer. Moreover, the administrator is subject to the control of the commune councils Dialogue: “Unlocking” the and acts under the supervision of the provincial Reform Agenda authority, the governor (the authority of the 42. The analysis discussed above tutelle), who represents the interests of the suggests that—at least in the short term and central government. The political proximity up till the national elections—key political between the two levels of authority shapes the and bureaucratic incentives are unlikely to party allegiances of the administrator, who might change significantly. This implies the need for act in the interest of the Governor rather than development partners to lower expectation and promoting the interests of the commune.32 It is set realistic targets for support. That said, it might therefore not surprising that demands expressed still be possible (and strategically meaningful) by several interests groups (including ABELO to focus efforts in addressing the institutional and representatives of local communities) to constraints that might not necessarily reflect a change the status of the administrator, making well-defined political strategy and instead be it an elected (by direct universal suffrage) have the undesired effect of poor coordination across been blocked. According to some observers, it ministries. is feared that this proposal—by granting the administrator greater legitimacy and autonomy 43. Efforts are still needed to reach a to exercise his power—might open the way for consensus on the strategic direction of the greater autonomy of the local authorities from reform. As discussed above, the time taken the center. to achieve consensus on the decentralization framework is explained not only by the technical 1.5.3 Bureaucratic Incentives complexity of the issue, but also by the inherently 41. Among civil servants prospects of political nature of the reform progress. In decentralization create fears about the this perspective, the frequent changes and security and location of employment. In amendment in the legal framework taking place particular, civil servants are reluctant to leave since the early 1990s33 also reflect the difficulty their civil service positions to join communal of reaching a consensus among decision makers administration without assurance that they will on the overall strategic direction of a reform retain full status. (While the center guarantees that has important political and institutional secure employment, it might not be the case at the consequences. level of most of the rural communes, depending 44. A coherent vision needs to be devel- on their fiscal situation). Furthermore, most civil oped as the foundation for a comprehensive servants have no intention to be posted outside and coordinated legal and institutional main urban centers and seem to be especially framework. After almost a decade of efforts, it reluctant to be accountable to poorly educated has become clear that the sectoral approach to local executives. pilot the decentralization reform through two ministries is not performing as desired given the transversal nature of decentralization as well as the lack of a conflict regulation mechanism to resolve interministerial rivalries, or “by-pass” 32  This explains why for example, even after the new manual of administrative and financial procedures has abolished the 2 percent “tax levies” by provincial authorities on communal budgets, some governors continue to benefit from this practice based on 33 The list of the legislative acts (decrees and enacting the political affiliation of their communal administrators. regulations) is available from the authors upon request. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 33 sectoral resistances to push decentralization 45. The Presidential decree should forward. It is equally clear that decentralization clarify the power and competencies of the can no longer be presented as a series of texts steering committee (CIOPD), to better signal or isolated actions as done in the recent past to the government’s political will and long-term (“piecemeal legislation”). This comprehensive vision vis-à-vis decentralization reform. The and systemic approach should include a CIOPD could be under the supervision of the robust legal framework governing the reform. second vice president . The CIOPD should then These legal texts should be consistent and adopt an action plan and report its achievements harmonized with other existing laws governing periodically (once per semester) to the head of decentralization. the government. 34 Chapter 2 Fiscal Decentralization: Context, Policy Issues, and Possible Changes A Summary of Main might contribute to increase subnational inequalities. For capital expenditures, Fonds Arguments and Key Findings National d’Investissement Communal (National 1. Despite its importance, fiscal Investment Funds for Communes, FONIC) decentralization represents the weakest provides some support with an equal transfer link in the decentralization framework so by communes that does not take into account far. An adequate transfer system for current their taxable capacity or needs. Moreover, expenditures is yet to be defined: to date, there while incentive mechanisms exist in the FONIC is no mechanism in place to ensure equitable allocation formula, they are not applied given and transparent transfers for current spending the limited envelope available. by communes34 in Burundi.35 This is unusual as territorial inequalities make this necessary in 2. In the absence of a predictable and most countries of the world and—in the case of transparent intergovernmental transfer Burundi—the uneven distribution of resources system, most communes struggle to remain fiscally viable and finance even basic operating costs. The socioeconomic diversity 34  Au Burundi les communes vivent seulement de leurs of communes (in terms of size and economic ressources propres ; il n’existe pas de ressources transférées ni par le potential) represents a technical challenge for Gouvernement, ni par les Bailleurs de fonds. Les communes collectent an equitable distribution of public resources leurs ressources en appliquant en particulier l’OM N°530/540/312 du 04/08/1997 across the national territory, in a context where local revenues represent only 2.7 percent of 35 Portant sur les taxes des communes. ETUDE SUR LA VIABILTE INSTITUTIONNELLE ET FINANCIERE DES COMMUNES RURALES total government revenues and remain largely AU BURUNDI, p. 39. concentrated in Bujumbura (25 percent of total Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 35 local revenues). The average annual communal that communes mainly provide local services revenue outside Bujumbura is FBu 103 million; that require proximity to residents with little equivalent to $68,000 or somewhat less than $1 technical content (high technical content per capita. Resources to finance the execution is associated with services such as water of Communal Development Plans (CDPs) are treatment or composting, for example) or high therefore largely insufficient, with the result fixed costs (similar services), the case for a the result that available social infrastructure, as national consolidation policy does not seem to well as the maintenance and operations of these be justified. Meanwhile, the legal status of the facilities (health clinics, primary and secondary communes of Bujumbura should be clarified. schools, and water points) remains limited to address citizen needs and improve access to 5. Policy issue #2: General responsibili- social services. ties assigned to communes should be clarified, given possible changes before 3. Given the fiscal fragility of the country finalizing more robust financing arrange- and its structural vulnerability to external ments. Also, the proposal for transferring shocks, the central government has little competencies (draft law) should be fiscal room for maneuver to improve the accompanied by a clear assessment of the fiscal viability of the communes. Under these required financial and human resources structural constraints, the financial implications required. These current responsibilities include of proposed legislation to increase the functional maintenance of communal roads, provision of responsibilities of communes need to be carefully official documents for residents; preparation considered. The government is aware of such of communal development plans; provision of trade-offs and has requested the assistance of the sport and cultural activities; and playing a role Bank and a group of donors to provide technical in education, health, and social services at the assistance in identifying feasible options to local level. This approach is less precise than in manage the fiscal decentralization process and other countries where communes either have the policy actions needed improve the medium- a specific list of responsibilities or a general term sustainability of the current system, under competence in local matters.36 The absence of conditions of macro-fiscal fragility. Specifically, clear and specific mandates reduces the scope this chapter aims to review the current status of for communes to play a key role in improving fiscal decentralization in Burundi and directly local service delivery. The new draft law address a series of five key policy questions that currently under preparation is a step forward have been raised by the national stakeholders to address current uncertainty and improve the and currently are at the center of public debate institutional viability of communes. However, it surrounding fiscal decentralization in Burundi. might also represent a potential trap since the financial and human resources implications 4. Policy issue #1: Given the existing have not been adequately assessed. assignment of functional responsibilities, the current territorial organization of Burundi is 6. Policy issue #3: On average, communal reasonable, and there is no need to reduce the revenue sources barely cover the existing number of communes. Moreover, regrouping responsibilities of communes. The legal communes as such does not necessarily alleviate framework for local tax collection needs to the lack of resources and survey results suggest be updated to ensure that local revenues are it may have costs in terms of accessibility of services for the population. Compared with other African countries, Burundi—and its 116 36  Mécanisme de Transferts Financiers Aux Communes (Fonds rural communes, with an average population National d’Investissement Communal) Mohamed El-Mensi Déo of roughly 90,000—is not an outlier. Given Banderembako, 2009 UNCDF. 36 adjusted by inflation. At the same time, efforts and (iii) shares some of the resource revenues are needed to strengthen local tax administration between producing and nonproducing system, improve tax compliance and strengthen communes, according to different scenarios. the link between improved revenue collections and better services for the citizens. Collaboration 2.1 Fiscal Decentralization Today with the Office Burundais des Recettes (Burundi Revenue Authority, OBR) should be promoted to share information on tax payers—in line with Important note for the reader: the recent experience of other countries such Data limitations hamper the scope as Rwanda—and/or use OBR as a tax collection of the analysis. There are no recent agent for the communes. comprehensive fiscal data for communes 7. Policy issue #4: In contrast to some in Burundi. For older years (2000-2010) of its neighbors in the region, communes do the team was able to obtain information not receive any central transfer to finance from ISTEEBU for almost all communes, but recurrent spending, indicating the need for recent ones, data is either completely for a current expenditure funding transfer missing (2011) or partial both in terms of scheme. The lack of such scheme is unusual geographical coverage or itemized coverage, as territorial inequalities make this necessary at least when in electronic form. In close in most countries of the world, and has collaboration with the Observatoire sur major implications for the efficiency and les finances communales located in the the effectiveness of the services provided Ministry of Interior – and operating under by communes. For example, generalizing the EU-funded Gtwara Neza Project – the the communal land service (Service Foncier team has collected data for a large set (100 Communal—see chapter 3) requires financial or so) of communes for 2012 on top level resource that fees alone are unable to provide. line items for expenditures (organized by Also, the expected increase in FONIC capital economic classification) and aggregate transfers will create greater demand for revenues. However, there is also no data on recurrent spending that will remain unfulfilled. expenditures by functional classification. A current proposal is under discussion, but does not specify the allocation criteria. Simulations presented in the chapter suggest that a more 9. The current legal framework equitable recurrent transfer system is quite indicates that communes have general feasible and appropriate, and would produce on responsibilities for local affairs. The first average a 45 percent increase in total communal conceptual pillar of fiscal decentralization is revenues. the assignment of functional and expenditure- related responsibilities. The functional 8. Policy issue #5: Given the expected responsibilities of communes in Burundi are increase in the importance of large-scale set out in the Law on Communal Administration mining activities and their geographic concentration, it may be important to (Law 1/02 2010). Most pertain to public services consider a revenue-sharing arrangement that are not of a national scale or already across communes. There is an opportunity assigned by law to the central government. But to integrate both decentralization and mining the law does not explicitly assign functions or through a mechanism that: (i) compensates responsibilities. Instead, Article 76 enumerates communes with mining related activities for items to be covered by communal spending: environmental degradation, (ii) sets aside salaries, recurrent costs, development plans, some resource revenues for future generations, maintenance costs of infrastructure belonging Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 37 to the commune, annuities for loans, waste 10. The role of communes is limited; many and wastewater removal, help to the poor, and local services are provided by deconcentrated spending on transferred responsibilities. Articles branches of the central government. The 91–95 cite local road maintenance. Furthermore, provision of basic education and health from the description of the tasks of the two top services, two activities that are devolved to civil servants (conseillers techniques), one infers the communal level in some countries, are that communes also prepare official documents carried out by deconcentrated branches of the (such as birth certificates, Article 51), prepare education and health ministries in Burundi, with the communal development plan (PCDC; Article few contributions (if any) from the commune 52), and promote sport and cultural activities level (but local input may be obtained through, and play a role in education, health, and social for example, parents’ committees for primary services (Article 53). An overall lack of detail in schools). The responsibility for road maintenance the legal framework has advantages as well as is shared by communes, central government disadvantages: whereas in principle the broad agencies, and the private sector, with instances assignment of responsibility for local affairs of national agencies (for example, those for allows communes considerable leeway to coffee and tea) carrying out some maintenance contribute to a broad range of local services, at work to ensure that relevant products can gain the same time, the absence of clear and specific market access. The provision of electricity and mandates reduces the scope for communes piped water (when available) is carried out by a to play a key role in improving local service national agency (Regideso) with direct payment delivery. As a result, in practice, the central by households. The provision of police and civil government plays a leading role in the provision protection (including firefighting) services is of many “local” services. national in nature. 11. The limited role of communes in public service delivery is reflected by the small share of public expenditures that are devolved to the commune level. In the absence of detailed expenditure data for all communes, it is difficult to determine the relative importance of commune-level expenditures with precision. Given that essentially all commune expenditures are funded from own-revenue sources, total commune expenditures can be estimated based on the level of communal revenue collections. In 2012 communes accounted for about 3 percent of public revenues and expenditures in Burundi. This percentage is derived from financial information for the city (Mairie) of Bujumbura (comprising 13 “communes”) and for 94 of the 116 communes outside of Bujumbura for which financial information is available. For the 107 communes for which data are available, revenues amount to FBu 15.4 billion. Adding an estimated FBu 600 million in additional revenue for the 22 missing communes yields a total of FBu 16 billion or 2.9 percent of total public revenues 38 for that year.37 From our analysis of communal only devolved expenditures and revenues that budget data, it appears that communes tend run through the accounts of the communes, to spend more or less what they collect in a which, in the case of Burundi, would suggest given year (especially since they do not receive a limited level of decentralization in terms of any intergovernmental fiscal transfers). For expenditure responsibilities. One can argue, example, in 2012, the ratio of total spending to given international evidence (see figure 2.1), total revenue was 1.04, a result consistent with however, that looking at devolved expenditures the presence of communal arrears. alone provides an incomplete picture of Burundi’s ongoing decentralization process. 12. Devolved public spending is one As elsewhere, central line ministries continue element of local public spending. Traditional to play an important role in the financing and measures of fiscal decentralization measure delivery of local services. Most importantly, all only devolved expenditures and revenues that sectoral staff posted at the local level in Burundi run through the accounts of the communes, continues to be employed and financed by their which, in the case of Burundi, would suggest respective central line ministries. Likewise, a limited level of decentralization in terms of central line ministries continue to be responsible expenditure responsibilities. One can argue, for an important share of sectoral development given international evidence (see figure 2.1), expenditures relevant to local public services. however, that looking at devolved expenditures In sum, devolved communal expenditures alone alone provides an incomplete picture of do not accurately convey the magnitude of Burundi’s ongoing decentralization process. public services, infrastructure, and government As elsewhere, central line ministries continue activities that takes place at the local level. to play an important role in the financing and delivery of local services. Most importantly, all 13. As a next step, a comprehensive sectoral staff posted at the local level in Burundi analysis of fiscal decentralization in Burundi continues to be employed and financed by their ought to look at all funding that flows to respective central line ministries. Likewise, the local level, including both devolved central line ministries continue to be responsible and deconcentrated spending. Measures for an important share of sectoral development to capture the geographic distribution of expenditures relevant to local public services. expenditures in the central national budget In sum, devolved communal expenditures alone should also be promoted. The current level do not accurately convey the magnitude of of communal expenditures does not reveal public services, infrastructure, and government the amount of deconcentrated spending that activities that takes place at the local level. takes place at the local level for the delivery of key local services, such as education and 12. Devolved public spending is one health service. Although preparing a complete element of local public spending. Traditional expenditure profile for the local public sector measures of fiscal decentralization measure falls beyond the scope of this study, a partial analysis was carried out covering only the education and health sectors to ascertain the 37  It is worth noting that this picture does not change much size of the “local public sector.”38 This analysis even if we take into account the in-kind transfers (for example, suggests that central funding of local public construction materials) that communes received from the FONIC prior to 2013, which were worth FBu 460 million in 2009 and services is important: central government FBu 934 million in 2010. Adding the FONIC contributions in 2010 increased the revenues of communes by 10 percent. As of 2013, these contributions should appear as increased communal revenues. Cash transfers by FONIC should be about FBu 8 billion, 38  The concept of a “local public sector” has recently come to increasing communal revenues by about 50 percent in 2014 (or by the fore as a complement to standard decentralization measures. about 1 percent of all government revenues). See http://www.localpublicsector.org/ for more details. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 39 Figure 2.1 La composition du secteur public local dans 15 pays, 2010–12 (%) 120.0 100.0 80.0 60.0 40.0 20.0 0.0 In azil au ia am ia e ia ue o sh ia m a a li n l a pa on d di Ba Fas er Ma ni ric s oz ger an an de Bu biq an ne Ne bo Be Br Le rit nz Af la Ug Ni a do in ng ra Ta h Ca rk ut M So Si M Devolved Deconcentrated Centrally-controlled Central government expenditures spending 'local' spending spending Source: Report No: 85735-SL Decentralization, Accountability and Local Services in Sierra Leone: Situation Analysis, Key Challenges and Opportunities for Reform, April 2, 2014, World Bank. spending—whether deconcentrated or centrally other minor taxes.40 As already noted, in 2012 controlled—on education and health accounts communes in Burundi collected only about 3 for 40 percent of total public spending.39 More percent of total government revenues. Indeed, systematic work can be proposed as part of analysis of revenue collection data over a longer ongoing analytical work on decentralization and time span (2000–10) confirms that the relative public financial management in Burundi. In this and absolute revenue importance of communes in Burundi is small. While communal income context, government might consider the need to increased in nominal terms from FBu 3.6 billion introduce geographic monitoring of subnational in 2001 to FBu 4.5 billion in 2005 and 9.9 billion public expenditure in the central budget—on a in 2010, much of this nominal increase was the provincial and communal basis—so as to permit result of inflationary pressure. Since inflation better monitoring of expenditure executed by was 225 percent over the period,41 in real terms, sector and its financing source. the amount of revenue collected in 2010 was actually only 4.4 billion, as expressed in 2001 14. Revenue sources assigned to FBu. Furthermore, when communal revenue communes in Burundi yield limited is expressed as a share of total government communal revenue. The second pillar of fiscal revenue (see figure 2.2), the trend shows a drop decentralization is the assignment of revenue in the relative importance of communal revenues sources to the local level. The main revenues between 2000 and 2010. The population of assigned to the communal level include a local Burundi grew by 52 percent in the 18 years property tax, business taxes, taxes on bicycles and motorbikes, taxes on cattle, and a variety of 40  A complete overview of communal revenue sources is available upon request from the authors. 39  Results from the analysis are available from the authors 41  Annuaire statistique du Burundi janvier 2001-janvier 2010 upon request. table 11.01 Indice général. 40 between the 1990 census and the 2008 census; thus, for the years since, we Figure 2.2 Communal Revenue as a Percentage of Total Government Revenue, Burundi: 2001, 2005, and 2010 assume a growth rate of 25 percent. Therefore, in real terms, corrected 3.7% for population growth, the financial 4.0% resources of communes in Burundi 3.5% 3.1% remained essentially unchanged 3.0% 2.7% between 2001 and 2010. 2.5% 2.0% % Communes/ 15. The size of devolved totales 1.5% revenue collection in Burundi is 1.0% small by international standards. Further confirmation of the limit- 0.5% ed role that communes play in 0.0% government revenue collection is 2001 2005 2010 provided when devolved revenue Source : Annuaire statistique du Burundi 2010, 2012 tableaux 14.01 et 14.11. collections in Burundi are compared to revenues collected by other weaknesses in communal tax collection such countries in the region. Figure 2.3 indicates the as the lack of lists of taxpayers, the lack of a tax size of devolved revenues in various African identifier for taxpayers used for various taxes, countries; Burundi would have ranked fifth the poor knowledge of tax laws by tax collectors, from last if it had been included in this figure. and poor enforcement of penalties against tax 16. Communal revenues vary by delinquents. This has also been confirmed by commune. Even though communes collect the focus group discussions undertaken for this only roughly 3 percent of government revenues study (see chapter 4) and suggests that there is overall (thus limiting their ability to deliver scope for improving communes’ revenue yield public services from own-source revenues), by improving local tax administration practices, this statistic understates the revenue potential and by strengthening the link between revenue of many rural communes. In fact, there is collection and local services. considerable variation in revenue potential 18. The central government provides no across communes (see figure 2.4), with many lacking meaningful revenue potential. In fact, recurrent intergovernmental fiscal transfers. 40 percent of communes collect no more than In many countries, intergovernmental fiscal FBu 1,326 (less than $0.90) in revenues per transfers (considered the third pillar of fiscal capita. The northeast of the country appears to decentralization) are an important source of contain better-performing communes in terms funds for communes. In contrast to some of its of per capita revenue. A similar pattern can be neighbors in the region (for example, Rwanda, observed looking at the expenditure side (see Tanzania, and Uganda), Burundi does not figure 2.5). supplement communes’ own source revenues by transferring additional funds for recurrent 17. Communal revenue collection commune expenditures (except for the salaries practices are weak. The main sources of of three commune officials. This lack of central communal revenues are taxes and fees levied government support for recurrent expenditures by communes. A recent study42 notes several has major implications for the efficiency and efficacy of the intergovernmental fiscal system: 42  Étude Sur La Viabilité Institutionnelle et Financière Des Communes Rurales Au Burundi Swiss Cooperation 2012 Charles Nihangaza. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 41 Figure 2.3 Local Revenue as a Share of Total Public Revenue in 22 African Countries: 2006–08 30.0% 25.0% 24,2% 19% 20.0% 17,9% 15.0% 15% 9,8% 10,4% 10.0% 7,4% 6,8% 6,6% 5.0% 4,2% 4,9% 5% 3,7% 3,3% 3,8% 2,6% 3% 2,3% 1,7% 1,3% 0,8% 1,8% 0.0% Ga ia Gh on Ke a M ya am co i Ni er A l go Ug isia m ia Af e Et pt nz a a Eg o Ni e ia n M or i So Sen a Z a a aw an bw M al Ta fric ric qu ri d s ut eg op Zi amb ni y g oz oc an n To b Fa M an ge n al Be bi ba hi Tu a h in rk Bu Source : figure 3, p. 25 l’Afrique (avec François Yatta) in Local Government Finance: The Challenges of the 21st Century (United Cities and Local Government Ed) (GOLD II 2010). Barcelone : Edward Elgar, 2011 : p. 25–59. Figure 2.4 Fiscal Revenue of Communes, Figure 2.5 Fiscal Expenditures of Communes, by Quintile, 2012 by Quintile, 2012 5th (highest) 5th (highest) 4th 4th 3rd 3rd 2nd 2nd 1st (lowest) 1st (lowest) Bujumbara Mairie Bujumbara Mairie No data No data Note: Note: 1st quintile: 395-1050 1st quintile: 168-958 2nd quintile: 1051-1326 2nd quintile: 959-1306 3rd quintile: 1327-1948 3rd quintile: 1307-1922 4th quintile: 1949-2931 4th quintile: 1923-3629 5th quintile: 3067-30344 5th quintile: 3630-30283 Source :Burundi 2012 Execution Budget Source: Burundi 2012 Execution Budget 42 • Generalizing the communal land service Figure 2.6 Poverty Rates (%) Burundi Communes, by Quintile, year 2008 (Service foncier communal, SFC, discussed in chapter 3) requires financial resources that fees for land services are incapable of providing. • The national goal of equitable regional development—as enshrined in the constitution—is unlikely to be achieved without an equalization transfer. 19. Communal investments are supported by communal savings and by transfers from the central government, both in kind (materials) and in cash. The main capital transfer mechanism for supporting communal investments is FONIC, created in 2007 following the merger of two preexisting funds. In 2013 the role of FONIC was changed from being a 5th (highest) 4th 3rd distributor of building materials and an agency 2nd 1st (lowest) Bujumbara Mairie supervising construction activities to a financing agency that provides cash capital grants to Note: communes. It is expected that communes now 1st quintile: 41.2%—54.2% supervise their own building activities.44 2nd quintile: 54.3%-60.8% 3rd quintile: 60.9%-65.9% 4th quintile: 66.0%-69.4% 20. Looking at the FONIC budget for 2013 5th quintile: 69.5%-76.2% and 2014, we first note projected growth in Source: ISTEEBU official report based on 2006 poverty survey the size of the total FONIC grant pool of 12 and 2008 census. percent for national resources allocated to FONIC and 26 percent for overall FONIC resources (see • Poor communes are unlikely to be able to table 2.1). The most important budget line is a provide all basic services to their residents, transfer of an equal amount to each commune. since supplementary tax revenues—while That budget line represents 70 percent of needed—are difficult to obtain. the total in 2013 and 77 percent of Burundi FONIC financing and 80 percent of total FONIC • FONIC transfers offer an incentive to use financing in 2014. Financing for performance subsidized capital spending rather than and “villagisation” increased between 2013 and proper maintenance spending. 2014 while that for building primary schools • Complementary spending to make full use dropped. of communal capital will be underfunded. As the FONIC funded capital increases, it will create a demand for recurrent spending Thus a typical FONIC project worth FBu 50 million in 2014 that will either crowd out other spending or generates a 2.5 million need for recurrent spending in all future remain unfulfilled.43 years until the capital is no longer useful. For more information, see: Recurrent Expenditure Requirements of Capital Projects Estimation for Budget Purposes, Ron Hood David Husband Fei Yu, 2002, World Bank Policy Research Working Paper 2938. 43  The exact relationship varies by type of project; assuming a 44  A description of the FONIC transfer rules and formula is .05 ratio (that is, FBu 50,000 per 1 million) is probably reasonable. available upon request from the authors. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 43 2.2 Managing Fiscal 24. There are various drivers behind the interest in reducing the number of Decentralization under Fiscal communes. One reason for a reduction is Constraints: Key Policy Issues financial viability (or, more exactly, its lack). Some communes have small budgets. One study 22. After almost 10 years of effort to argues that—even that considering communes’ achieve greater fiscal decentralization, functional responsibilities are few in number— a number of issues still require strategic the minimum annual budget for an effective and policy decisions to be made. The way in efficient rural commune is FBu 600 million.45 which these policy questions are addressed Others compare Burundi with Rwanda, where will have a considerable impact on the ability the number of districts has been reduced to 30 of Burundi’s public sector to achieve greater since 2006, with a mean population of 400,000 efficiency and improve access to basic public and responsibilities in the delivery of education services. Responding to a specific request of and health services. technical assistance, this section will address policy questions that have been raised by the 25. The communes of Burundi are authorities and currently are at the center of not that small in population terms, given public debate surrounding fiscal decentralization their responsibilities. Table 1.1 (chapter 1) in Burundi: (i) is the existing number of communes presents information on the size of communes appropriate, or should the number of communes by province. It shows variations in population be reduced; (ii) exactly which functional size that may warrant some mergers, but given responsibilities should be assigned to communes; that communes mainly provide local services (iii) how can communes increase their own- that require proximity to residents with little source revenues; (iv) how to put in place pro-poor technical content (high technical content is transfer schemes that can supplement commune associated with services such as water treatment revenue collections; (v) how to improve the or composting, for example) or high fixed costs efficiency of capital transfer mechanism (FONIC); (similar services), the case for a national merger and (vi) what are the potential impact of mining policy is not easily made. Table 2.2 presents data revenues for communes finances. The following on 17 African countries. Compared with them, sections address each issue in turn. Burundi—and its 116 rural communes, with an average population of roughly 90,000—is not an 2.2.1 Policy Question 1: Is the Number of outlier. Communes in Burundi Optimal? 26. There is no stable or natural number 23. The number (and average size) of of communes.46 In 1960 there were 18 provinces communes is an important policy issue. As and 181 communes. But in 1965 the number of noted in chapter 1, the number of communes in communes dropped to 79. In 1981 there were Burundi has evolved over time. It is not unusual 15 provinces and 114 communes; 10 years later, for countries to align the structure of local in 1991, there were 113 communes outside governance with two competing needs: on the Bujumbura. Finally, in 1998, 3 new communes one hand, to ensure that local officials are close were created and Bujumbura was divided into to the communities that they serve, and on the other hand, to ensure that commune jurisdictions are sufficiently large in scale that they are 45 Étude Sur La Viabilité Institutionnelle Et Financière Des able to perform their functions effectively and Communes Rurales Au Burundi Swiss Cooperation 2012 Charles efficiently. Whether the number of communes Nihangaza. ought to be reduced has recently received some 46  Details from Le sénat et la politique de décentralisation au attention in policy circles in Burundi. Burundi août 2010, partie II. 44 Table 2.1 FONIC Budgetary Summary, 2013 and 2014 Items 2013 2014 Observations Equal transfers to Primary schools Investments in communal plan 40 million per commune in 2013; communes 5,160 million (PCDC) 6450 millions 50 million in 2014 Bonus for performance 90 million thus 5 million to 200 million for 65 communes Amounts per commune in 2014 each of 18 communes (63 regular + 2 Bujumbura); are in proportion to the score amount varies between FBu obtained at the annual evaluation 2,370,000 and 3,863,000 of communes Finishing 728 million 607 million Remainder of 2013 budget to be abandoned schools spent in 2014 Project villagization 341 million 500 million For housing Building classrooms 501 million: 3.9 for each GIZ/ADLP 0 1 020 millions New in 2014 Total 7 417 million (sum of items 9 358 million or 8,338 without above–our calculations) GIZ Source: 2013 Plan d’action du FONIC révisé 9 juillet 2013; 2014 Budget d’investissement du FONIC, 16 janvier 2014. The equal amount to each commune of 40 (in 2013) and 50 (in 2014) million are a drawing right paid against proof of disbursement by communes; the total is broken down between spending on manpower and materials. Communes who spent less than 40 million in 2013 see the remainder made available in 2014. GIZ in 2014 made available €500,000 to FONIC to provide FBu 60 million to each of the communes in the provinces of Gitega and Monoro. FONIC has chosen not to reduce funding of communes from Burundian revenue in these provinces so they will have access to 110 million. a further 13, yielding the current number of and of other countries.48 As a result, voluntary, 129. There is no evidence to suggest that these intercommunal collaborative arrangements may changes in commune numbers had a meaningful well emerge as an alternative to mergers. impact on service delivery outcomes. This point may be used to argue against a change in the 28. Regrouping communes does not number of communes (unless accompanied by necessarily alleviate an overall lack of a major change in the functional responsibilities resources and may create intracommunal of communes). tensions. Merging three communes with the same population and each with a budget of FBu 27. Intercommunal groupings are 200 million would increase the absolute budget almost nonexistent. The National Policy on size to FBu 600 million but would have no Decentralisation of 2009 argues for the use impact on the per capita amount of resources. of intercommunal groupings (collaborative Unless accompanied by more far-reaching arrangements between two or more communes) administrative reforms, it is unlikely that the as a means of achieving development goals. The consolidation of communes would create more emergence of such groupings has been slowed fiscal space for local services through cost by the lack of a clear legal framework. A decree savings in local administration. While it could aimed at facilitating progress in this area was be argued that regrouping communes might adopted by the Government of Burundi in reduce some of the fiscal inequalities between October 2013,47 along with one on international them, such mergers would not be without cooperation between communes of Burundi social and political costs. For example, consider 47  DECRET N° 100/ 243 DU 06 OCTOBRE 2013 PORTANT 48  DECRET N° 100/ 244 DU 06 OCTOBRE 2013 PORTANT REGLEMENTATION DE L’INTERCOMMUNALITE AU BURUNDI REGLEMENTATION DE LA COOPERATION DECENTRALISEE AU http://www.presidence.bi/spip.php?article4118. BURUNDI http://www.presidence.bi/spip.php?article4119. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 45 Table 2.2 Number of Local Governments (Lowest Tier) and Average Populations in 17 African Countries, 2006–08 Number of local Average local Population Country Population country 000 governments units 000 Algeria 1 541 21 33 800 Bénin 77 120 9 300 Burkina Faso 351 43 15 200 Côte d’Ivoire 987 20 19 600 Malawi 40 330 13 200 Mali 703 18 12 700 Maroc 1,497 21 31 478 Mauritania 216 14 3 200 Mozambique 43 477 20 530 Namibia 54 33 1 830 Niger 265 55 14 700 Rwanda 30 310 9300 Sénégal 431 29 12 700 South Africa 283 172 48 687 Tunisia 264 39 10 326 Togo 354 19 6 800 Zimbabwe 90 129 11 631 Source : Preliminary unpublished work for “Africa” (with Francois Yatta) in Local Governance Finance: The Challenges of the 21st Century (United Cities and Local Government Ed) (GOLD II 2010) Barcelona: Edward Elgar, 2011: p25-59. merging three communes: all have the same at the household level. This issue has not been population, but one has a budget of FBu 800 studied by the Government of Burundi. Data million (a “wealthy” commune) and two “poor” were gathered in 2013 by the World Bank for an communes have only 200 million each. This analysis of the benefits and costs of a plausible merger would theoretically double the per capita reduction in the number of communes outside resources available to the “poor” communes. In Bujumbura, and the findings from the fieldwork practice, however, such mergers often create conducted at that time suggest that a reduction social and political tensions since, to achieve in the number of communes is not justified (see intracommunal equalization, the majority of box 2.1 and annex 5). voters in the new jurisdiction (located in the poor parts of the commune) would have to 30. The legal status of the 13 communes reduce (by half) the resources available to the of Bujumbura is under discussion. The minority of voters residing in the richer part of communes of Bujumbura have an elected council the commune. but do not have financial powers, which are instead exercised by the province/municipality 29. Regrouping affects residents’ access of Bujumbura. A proposal was put forward in to services. Increasing the size of communes January 2014 to transform the communes of may yield little in terms of financial gain but Bujumbura into full communes. Some argue that increase the time or monetary costs required to this proposal is a step toward a merger into three access these services—and even reduce their use communes; others believe that it is an attempt 46 Table 2.3 Boroughs Around the World: Strengths and Weaknesses in Three Areas Legal status Political status Budgetary Autonomy Montréal Strong Strong Strong Toronto Weak Strong Weak New York Strong Strong Weak Los Angeles Weak Weak Weak Mexico Strong Weak Weak Bogota Strong Weak Weak Buenos Aires Weak Strong Weak Paris, Lyon, Marseille Strong Strong Weak Source: Gouvernance urbaine, mégavilles et décentralisation sous-locale : un tour d’horizon des Amériques, Jean-Philippe Meloche 2014, Mimeo. Note: Legal status is considered “strong” were boroughs’ number, borders, and powers are protected by law and weak otherwise; political status is considered “strong” where borough politicians are elected directly by the population, and weak otherwise; budgetary autonomy is considered “strong” if the borough has a substantial budget under its control, and weak otherwise. Box 2.1 Assessing the Costs and Benefits of Reducing the Number of Communes: An Empirical Analysis with Unchanged Services To establish the benefits and costs of reducing the number of communes in Burundi, the following steps should be used: • Establish the annual number of interactions between the residents of communes and the communal office: 157,500 (for the communes outside Bujumbura) • Establish the cost in time and money of these interactions: FBu 1 billion (again, for the communes outside Bujumbura) • Establish a plausible, reduced number of communes (for example, 40) • Calculate the reduction in costs linked to this reduction in commune number: communal wage costs could drop by about FBu 200 million and central wage costs by about 400 million, for a total of 600 million • Calculate the impact of a reduced number of communes on individual access costs: accessing centralized communal headquarters would cost residents FBu 1 billion more The question then becomes what can be done with only the authorization of either communal administrators (l’administrateur communal) or advisers (conseillers techniques), and what can be done in their absence. If they need not be present, then services may be maintained in their existing locations, at a reduced cost of about 5 percent of communal revenues (even if some of this is saved centrally). The answers gathered in community focus groups carried out by the World Bank and discussed in chapter 4 indicated that people are concerned about access to the administrateur communal. In a survey of the staff of 16 communes, interviewees indicated that some services required the intervention of the administrateur communal and that increased distance between administrative offices and the communities they serve could reduce the demand for services. If the number of politicians is reduced but not the number of other related employees, a reduction in the total number of communes is not financially justified. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 47 to formalize equal treatment of all communes in • Equip, and maintain local tribunals and jails Burundi. • Build centers for play or professional 31. International experience indicates training for vulnerable groups (within their that designating Bujumbura’s communes budgetary means) as “boroughs” may be more appropriate. • Build, maintain, and manage markets Subunits of large urban areas typically operate • Provide welfare services for poor students under one of the following two legal statuses: (within their budgetary means) independent cities, such as those of Brussels or • Provide medical services for vulnerable London, and boroughs (see table 2.3). Because groups (within their budgetary means) Bujumbura’s communes are part of an integrated urban area, it seems appropriate to consider • Manage waste formally treating them as boroughs. • Prepare and deliver land titles49 2.2.2 Policy Question #2: Which • Manage communal land and forests Functional Responsibilities Should Be 33. The current draft law includes a list of Assigned to Communes? transferrable competencies in infrastructure 32. Amid the policy debate surrounding and services. If implemented, the transfer of the optimal number of communes, there responsibilities would place communes in charge is an ongoing process of transferring of basic communal services (roads, sports, waste) responsibilities from the central to the and would give them partial responsibility commune level. Since 2009 the exact for the delivery of local education and health responsibilities of communes have been under services (facility construction, equipment, and review. A draft law long under discussion on the maintenance). Also, and not necessarily common transfert des compétences de l’État aux communes internationally, the commune would have a role du Burundi was presented to a parliamentary in the construction and maintenance of local commission in the spring of 2013—and its tribunals and jails. What is surprising is the lack of explicit competencies in the provision of some study was suspended soon after. A new version typical community services, such as community was prepared during the summer of 2013, to be development and youth activities or sports. Even submitted to parliament in 2014. In the areas of though communes are assigned responsibility infrastructure and services, as far as we know to build, maintain, and manage (gérer) sports - since the official version has not yet been facilities, they need not provide services in this made public - this draft law specifies commune area. responsibilities as the following: 34. While the draft law is a step in • Build, equip, and maintain primary and the right directions in terms of clarifying secondary communal schools responsibilities of communes, it needs • Build public health centers, maintain them, to be accompanied by a plan of how the and provide nonmedical equipment required financial and human resources will • Build and maintain communal water outlets • Build and maintain public toilets, sewer and 49  It is important to observe that the responsibility to provide wastewater plants, and incinerators land certificates is already devolved to communes under the new land code (2011, article 384). This recognition reflects efforts to • Build and maintain communal roads repeal old legislation and harmonize the provisions of the Land Code with the provisions under the 2005 communal law. This is not • Build, equip, and maintain cultural, sport, an unusual process and reflects the recent shift of responsibilities and touristic communal infrastructure to communes. 48 be transferred. How will the communes be 36. A new communal finance law is staffed to take on their new responsibilities? For currently under study. Naturally, a more example, to remain revenue neutral, by how much appropriate solution to the erosion of the local will the health ministry’s budget be reduced if tax base is a formal revision of the structure and some health competencies are transferred to rates of local taxes. Communal taxes included in communes? These are all important questions the proposed revised local tax structure include (with both political and technical implications) the following: that would be best addressed before the passage of a law. Also of concern is a lack of connection to • L’impôt sur le gros bétail: tax on large cattle the laws of the relevant ministries. For example • La taxe sur la vente de bétail: tax on the sale in primary education a law was adopted, in of cattle 2013, with one goal: to “decentralize” education • La taxe d’abattage: slaughtering tax services to deconcentrated communal education • La taxe sur les cycles et cyclomoteurs: tax on directorates. bikes with/without engines 2.2.3 Policy Question #3: How Can • L’impôt foncier: property tax Communes Increase their Own Revenues? • L’impôt sur les revenus locatifs: tax on rental income 35. The existing legal framework of local • La taxe sur les pylônes: tax on cellphone taxation is responsible for considerable antennas erosion of the tax base, with the tendency of • La taxe forfaitaire sur les activités (patente): communal councils to increase local tax rates. lump sum tax on businesses Own-revenue sources are currently the main • La taxe sur les cultures industrielles: tax on funding source for communes (together with industrial agriculture FONIC and donor-funded projects)50. In most cases, tax rates are set in nominal terms (that The tax on cell-phone antennas (towers) is a is, a fixed amount per item such as a certificate, new tax, whereas the other communal taxes are a head of cattle or per square meter (m²) of already in existence. The proposed commune property); this was accomplished in 1997 by an finance law includes significant changes to both ordonnance. Since then inflation has been 500 the existing tax structure and rates. percent but local tax rates were never legally 37. The proposed legislation incorpo- changed, resulting in the considerable erosion rates a number of important changes to of the local tax base. To by-pass this problem, communal taxes and to the tax environment communal councils have increased local tax of communes.52 The most common proposed rates with the implicit/complicit agreement of change in the communal tax system is an increase the Ministry of Interior—based on the argument in the taxes on various items. For example, the that if the population agrees to pay the new rates, rate per head of cattle—now at a fixed amount then the tax rate hike is legitimate. This relies on an interpretation of the communal law that of FBu 300—would be set between FBu 500 and allows communes to overrule the ordonnance.51 FBu 1,500, to be determined by the communal council. In addition, the draft legislation includes a promise to regularly (periodicity not stated) update rates using a joint decree 50  The structure and rates of the revenue sources to which communes have access is available upon request from the authors 51  Voir Étude sur la viabilité Insititutionnelle et financiere de 52  These comments draw on NOTE SUR L’IMPACT DU PROJET communes rurales au Burundi. Coopération suisse 2012, Charles DE LOI SUR LA FISCALITE COMMUNALE mai 2013 Charles Nihangaza Nihangaza. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 49 from the ministries of finance and the interior.53 these key revenue sources included (i) lack of Additional proposed changes to the communal information about the property tax; (ii) lack tax system include: of knowledge of the legal provisions allowing communes to collect property tax; (iii) conflict of • A clarification of the role of communes and interests of the collectors (owners of properties); the national tax office (Office Burundais and (iv) resistance of local population to comply des recettes, OBR), by which communes are with the legislation and pay. stated to have the same powers over their taxes as does the OBR for its taxes (Article 39. The role of the OBR in local taxation 3). needs to be clarified to avoid perception of • The possibility for the communal council for overlaps and double taxation. There is some some taxes to choose a rate within an interval confusion as to the role of the OBR in communal rather than use a single centrally set rate. taxation; some communal and national officials believe that the OBR collects communal taxes. In • Important increases in the rates of the reality, however, OBR representatives confirmed property tax. The rate on buildings goes that they do not collect communal taxes. The from a fixed 36 FBu/m² to between 0 and perception by some who intervene in communal 500 FBu/m² bracket with the rate linked to finances that the OBR is present in this field the type of services available to the property appears to be due to two facts. First, the existence while for land attached to buildings it goes of two lump-sum taxes on small merchants from 4 FBu/m² to a 0–50 FBu/m² bracket. (one national and one communal) may lead • The highest marginal rate for the rental some to believe that the OBR is involved in local income tax drops from 60 percent to 30 taxation. This confusion is fueled in part by the percent, while an annual exoneration for fact that the national income tax in some cases the first FBu 1.800.000 of rental income is can be calculated using a formula similar to that introduced, both changes creating a loss in of the local taxes communales sur les activités revenue. économiques (patente).54 Some merchants argue with officials that the payment of one tax • The lump sum business tax is limited to (either to the OBR or to the commune) covers businesses with a turnover below FBu 24 the obligation for both. While that is incorrect million. This means large businesses no as of now, this (perceived) double taxation may longer contribute to the financing of local change under the proposed revision of communal public services, thus segmenting businesses taxes, as discussed above.55 A second reason for between OBR and communal taxpayers. It is the confusion over the OBR’s role in communal not clear how information on change of size taxation is that in some policy discussions OBR will be collected/handled. representatives have argued that the rental 38. One key challenge will be to overcome income tax (l’impôt sur les revenus locatif) given some communes’ general lack of interest in to communes at the time of decentralization collecting taxes. Our survey of 16 communes indicates that as of August 2013 only six communes collected rental income tax (impôt 54  Article 40 of the Loi relative aux impôts sur les revenus (2013) sur le revenu locatif) and four, property tax allows this for businesses with an annual turnover less than FBu (impôt foncier). Reasons given for not collecting 24 million with the following form, http://www.obr.bi/images/ stories/download/Formulaire_dclaration_Prlevement_IR_new. pdf. 55  In other cases, it appears that some border communes want 53  The legality of using a decree to change rates set by law may to charge taxes or market fees on transit goods (sealed trucks), be an issue. which they are not entitled to do. 50 should be centralized. percentage points of value added tax (VAT) (out of an 18 percent rate in 2013) to communes, 40. The OBR has a potential role to play to be distributed on the basis of their official in the collection of local taxes, and can help population. This would represent about 2.75 communes to strengthen the current system. in so far as the Impôt sur les revenus locatifs percent of VAT receipts. Although the proposed is one schedular tax among others applied to introduction of a recurrent grant is a welcome various types of income, it would be normal development, the population-based allocation that it be collected by the OBR to ensure a formula limits the ability of the proposed grant coherent treatment of individual and corporate to support poorer communes. One refinement income taxation. Under this scenario, the OBR would be to make the allocation formula pro- could act as a collection agent for communes, poor and thus help those communes in need of while amounts collected could be paid back greater help. Table 2.4 presents an example of a to communes on a derivation basis. This is more pro-poor transfer scheme currently used apparently argued for by the OBR as part of in the region. an effort to harmonize with the East African Community (As a counterargument, it should be noted that this tax is collected by districts in Table 2.4 An Example of a First-Generation Transfer Formula: Kenya Rwanda until 2015). This debate does not apply to the property tax—also ceded to communes at Parameter Weights (%) the time of decentralization—which remains an Parameter Formula Approved by appropriate tax for communes to collect. Parliament Population 45 2.2.4 Policy Question #4: How Can a Pro- Equal share 25 Poor Funding Transfer Scheme for Current Poverty 20 Spending Be Established to Complement Land area 8 FONIC? Fiscal responsibility 2 41. An unusual feature of Burundi’s Total 100 intergovernmental fiscal system is that Source: Commission on Revenue Allocation, Government communes do not receive any transfers to of Kenya, 2013, http://www.crakenya.org/information/ revenue-allocation-formula/>. finance recurrent spending. One explanation Note: This formula was implemented as of 2013; the fiscal offered for the absence of a recurrent transfer responsibility share is currently also allocated as an equal system by finance officials is that the central share as data to be used in assessing fiscal responsibility is government ceded two tax fields to communes not available given the recent creation of counties (March 2013). to fully pay for their spending needs: the property tax and the rental income tax (l’impôt foncier et l’impôt sur le revenu locatif). But 43. A more equitable recurrent transfer the concentration of these two tax bases in scheme is quite feasible in Burundi. By way of Bujumbura and a few urbanized communes illustration, box 2.2 presents a possible formula means that the revenue yield from these two for a communal transfer system for current taxes is inadequate to ensure a minimum level expenditures. The proposed formula contains of basic local services for most communes. four allocation factors: an equal amount per 42. The reform of communal taxes commune (10 percent); land area, adjusted under consideration includes a proposal to for topographic conditions (20 percent); create a recurrent transfer scheme. In the population below the poverty line (30 percent); proposed law, there is the intent to pay out 0.5 and population (40 percent). The simulations Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 51 based on the allocation formula56 show that the Figure 2.7 Projected Per Capita Transfers importance of such transfers for communes to Rural Communes, Using Hypothetical would be quite significant in relative terms: for Allocation Formulation 2010—the last year for which data are available for almost all communes—such a transfer scheme (with the envelope adjusted down to 2010 VAT revenues) would have meant a 45 percent increase in total communal revenues or one-third of total communal revenues, post transfer. 44. The relative ease with which a more equitable, formula-based grant can be simulated suggests that, in the planning of a future recurrent grant scheme, further consideration should be given to allocation. For illustrative purposes, simulated transfers were computed according to the proposed allocation formula. Figure 2.7 shows variations in the levels of simulated grants, ranging from 5th (highest) 4th 3rd FBu 558 to FBu 1,330 per capita; the communes 2nd 1st (lowest) most favored by the formula are concentrated to Bujumbara Mairie the east of the country. Source: Authors’ own calculations using the transfer formula discussed above. 45. In addition to the possible Note: Quintile boundaries (FBu) 1st: [558, 704] 2nd: [704, introduction of a recurrent grant system, the 738] 3rd: [738, 798] 4th: [798, 889] 5th: [889, 1330]. capital grant scheme (FONIC) will continue to play an important role in Burundi’s system of intergovernmental finance. In fact, For example, paying a bonus to the best- FONIC quickly grew in importance from 2010 performing communes in 2013 and 2014 does to 2012 (see figure 2.8). As such, the various not respect the FONIC internal procedures but is reforms in its governance—implemented in justified by FONIC management as ensuring the 2010–13 with support from the Gesellschaft continuity of past policies (PRADECS program). für Internationale Zusammenarbeit (German Such deviations from FONIC’s own operating Agency for International Cooperation, GIZ) procedures are best avoided since they offer and United Nations Development Programme (UNDP)—are most welcome. possibilities for potential abuse and institutional conflicts of interest. 46. To ensure a robust capital grant system, some FONIC choices need to be 47. Increasing transfers to communes reexamined. An objective and transparent should be accompanied by better public formula-based grant mechanism loses financial management practices. Recent credibility if it does not operate in accordance diagnostic work undeertaken by technical with its own formula and operating procedures. partners has highlighted difficulties in both the management of public procurement and in the management of local funds. 56  The formulas used for the simulations are available upon request from the authors. 52 Box 2.2 Establishing an Equitable and Transparent Transfer System for Communes: A Proposal with Simulations To implement a transparent and equitable transfer system for communes in Burundi policy makers must: (i) identify eligible communes, (ii) establish the resource envelope for transfers, and (iii) determine the allocation formula. 1. Identify eligible communes We exclude the 13 communes of the capital (Bujumbura Mairie) as transfer recipients, given that poverty is much less prevalent there than in the rest of the country. Given Bujumbura’s role as a capital city, these 13 communes both have special needs and special access to the central government. 2. Set the envelope for transfers As noted earlier, there is a policy reform under study that would allocate 0.5 percent points of VAT (18 percent rate in 2013) to communes, according to their official population. This is about 2.75 percent of VAT receipts, and is the transfer envelope used here. This is not a judgment on the adequacy or inadequacy of the proposed amount but simply reflects that it seems indicative of the evolving policy of the Government of Burundi. In the 2013 national budget VAT receipts comprise FBu 192.1 billion57 while for 2014 the predicted amount is approximately FBu 221.2 billion.58 It thus seems reasonable to use FBu 210 billion, which—multiplied by 2.75 percent—yields an estimate transfer pool of FBu 5.8 billion to allocate across communes. This is about 1 percent of expected tax receipts of the central government for 2014. Since some calculations are based on 2010 data, we note that the grant pool for that year would have been FBu 2.7 billion.59 Two additional items that need to be clarified before implementing such a scheme are, first, whether to rely on budgeted VAT revenues or realized VAT collections, and second, which entity will be the transferring agency: • Using the budgeted amount implies that all the macroeconomic fluctuations (either up or down) are supported by the central government. Using the collected amount shares the impact of macroeconomic fluctuations between the central government and the communes. • If one uses collected amounts, OBR lagged data could be used to set the monthly transfers. OBR could make the payments directly to communes or they could transit through the ministry of interior or finance. 3. Allocation formula A pro-poor, equitable distribution formula must account for: • Differences between communes in terms of expenditure needs (for example, differences in demand for public services caused by demographic factors) • Differences between communes in unit costs resulting from factors such as topography or climate • Differences between communes’ tax potential One could add incentives to improve performance in the supply of public services or tax collection; we suggest holding off on these items until the transfer of responsibilities and tax reforms discussed elsewhere have been implemented. For Burundi we have no indicators of tax capacity at the communal level, such as the value of properties; discussion with the OBR indicates that there are no indicators of tax capacity for national tax available at the communal level. In addition, communes in Burundi currently carry out municipal-level functions such as birth and death registry, local road maintenance, garbage removal, and cultural and sports activities but do not provide health or education services. Against this backdrop, the following points should be noted: 57  FBu 192 088 819 724, to be exact; Loi 1/13 , 30 juillet 2013 , Budget revisé , p16. 58  Fbu 221 776 438 178; Projet de loi, Budget 2014, article 1, p 23 59  Tableau A, loi 1/31 31 décembre 2009. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 53 • The weight for the fixed amount was derived using information from the survey of 16 communes on the wage costs of the compulsory employees of communes. This came out to FBu 475 million. Taking into account a 20 percent overhead yields FBu 570 million, or about 10 percent of the envelope noted above. The equal share window results in a lump-sum allocation of about FBu 5 million per commune. In the follow-up work on the formula to be implemented, care should be given to measure this with a larger data set. • The impact of topography is taken into account by weighting the size of each commune’s land area equally by two indicators: the degree of “slopiness” and the highest elevation in the commune. • The impact of poverty is taken into account by introducing the poor population into the formula. We use both monetary and nonmonetary poverty. Weight of Four Items in Allocation Formula Criteria for Equal per commune Area corrected for Poverty Population allocation topography Weight 10% 20% 30% 40% Note: Area and topography data are from the Institut de Géographie; population data are from the Burundi Statistical Handbook; poverty data are calculated by the Burundi statistical office (ISTEEBU) using data from a poverty survey for 2006 and the 2008 census. In the case of public procurement, one possible approach to improve procurement practices is Figure 2.8 FONIC Investment Funds, 2010–12 to use intermediate structures grouping several communes. These could be inter- communal bodies. In the case of local funds, evidence is 8000 less conclusive but for one market, one finds 7000 increasing arrears and some figures hard to 6000 reconcile.60Accordingly, it would be appropriate to reinforce the audit capacity of central bodies 5000 both at the provincial level and at the central 4000 level using the auditor general office (cour des 3000 comptes). 2000 2.2.5 What Are the Potential Implications 1000 of Mining Revenues for Communes Finances? 0 2010 2011 2012 48. The mining sector promises to be 000 000 FBu an important driver of economic growth in Burundi in the next 10 years. As such, Source: Authors’ own calculations using BOOST data. the impact of mining expansion on commune finances merits some discussion in the context of the country’s fiscal decentralization strategy. Burundi is believed to have considerable mineral resources, with the most important potential in nickel, gold, and coltan. In particular, the country 60  See AUDIT DU CONTRAT D’AFFERMAGE DU MARCHE DE has considerable potential in nickel exploitation. MATANA, FenRaj conseil s.a2013 Musongati, the main laterite nickel deposit in the 54 country, is among the 10 largest undeveloped are particularly unreliable. It is estimated that laterite deposits known, with an inferred export data represent less than 20 percent resource of approximately 180 metric tons.61 of the quantities actually exported. Available National reserves are estimated at 285 million estimates place the value of gold exported tons, and over the past five years a number from Burundi at $400 million.64 With an ad of international companies have expressed valorem tax of 2 percent on the artisanal mining interest in commercial exploration of these of gold, the Burundian government could be resources.62 Best-case scenario analysis (under- entitled to receive up to $8 million in 2014, taken in the Country Economic Memorandum for without including other fees and charges due, Burundi, CEM, in 2011) estimates tax revenues if collection and declaration mechanisms were from industrial-scale exploitation of Musongati strenghtened.65 alone to range from $50 million to $233 million (2 to 10 percent of 2011 gross domestic product 50. Mining activities are concentrated in [GDP]), depending on global nickel prices. several communes. Ongoing mining activities Accordingly, the Government of Burundi has in Burundi target gold, Coltan, Cassiterite, and identified mining as a potential source of growth Wolframite and are concentrated in the north and income creation in its Second Poverty of the country in the provinces of Cibitoke, Reduction Strategy Paper. As part of its efforts Kayanza, Bubanza, Ngozi, Kirundo, and Muyinga to develop the sector and attract investment, as well as Ruyigi to the east. Most recently, a Burundi passed a new mining code in October mining site was opened in Gitega (see figure 2013. The mining sector is also at the heart of 2.9). Mining exploration is concentrated in the the first and third pillars of the World Bank’s north of the country and also along the shores FY13–16 Country Assistance Strategy (CAS). of Lake Tanganyika (see figure 2.10). The CAS identified mining as an opportunity to 51. The geographical concentra- further diversify the Burundian economy and tion of potential large-scale mining may yield significant revenues for the country. exacerbate intercommunal differences in the distribution of costs and negative 49. Today the mining sector consists of externalities. Large-scale mining is likely to be some 20,000 artisanal miners and contributes more geographically concentrated than small- minimally to GDP. Available production and scale mining, given the need for larger deposits export data may not accurately reflect minerals to make the activity viable. This means that actually produced in and exported from certain communes and communities will bear Burundi. The World Bank’s Oil, Gas and Mining the brunt of the negative environmental impacts Unit (SEGOM) is finalizing a study of Burundi’s of large-scale extraction. Currently, a large artisanal and small-scale mining sector, which, number of communes bear the costs of small among other things, aims to verify mineral scale unregulated mining activities. Exploitation export and production data.63 Data for gold activities may require additional investments in the construction and maintenance of infrastructure such as roads, schools, or hospitals, 61  The World Bank. (2011). Republic of Burundi-Country and additional costs in education, health, and Economic Memorandum (CEM): The Challenge of Achieving Stable and Shared Growth. Washington, D.C.: The World Bank, Report No. other social services that the producing area 51880-BI. 62  African Development Bank. (2009). An Infrastructure Action Plan for Burundi: Accelerating Regional Integration. Artisanal Mining Study, forthcoming”). 63  See Exploitations Minières Artisanales du Burundi sur la 64  Ibid. Période 2009-2013: Rapport en relation avec l’Etude Diagnostique de l’Initiative pour les Industries Extractives au Burundi (“Burundi 65  Ibid. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 55 would not otherwise have borne. For Figure 2.9 Active Mines example, the transport of minerals may accelerate wear and tear of roads, while an influx of workers attracted to the area may require expanded provision of services. The costs of environmental damage caused by extraction should also be considered. Other considerations may include the costs of resettlement of communities affected by extraction activities, the costs of additional strain on water sources, and the costs of general disruption of livelihoods. 52. For the high-impact activity of large-scale mining to gain social acceptance, it may be necessary that mining firms offer and finance needed local services. Often companies voluntarily enter into agreements with local communities in areas where extraction is taking place to avoid conflict and obtain a “social license” to operate. Such agreements may involve services such as building roads or schools or health centers. A number of countries require companies to enter into agreements with local communities for community development or to Source: World Bank, SEGOM, Burundi Artisanal Mining Study, Forthcoming. submit community development plans as a condition to receiving an exploitation license.66 Burundi might consider also be considered. a legal mandate for companies to contribute to development in communities affected by 53. Communes adjacent to mine sites mining activities (local content). Such laws may also be affected. The main site of a mine should be designed to ensure that companies will normally be located in one commune, but do not undertake wasteful projects that are not if it is an open-pit mine it could overlap com- in line with communities’ development needs munes. If the subsoil deposit extends across and goals. One possibility is requiring that several communes, then those without the companies’ community development projects main site will see “their” resources extracted be aligned with communes’ development plans. without any benefit, particularly if community Systems for monitoring the implementation of development obligations are not designed community development agreements should to cover all affected communes. Also, while mining may be carried out in one commune, the refining of the mineral may take place in 66  See, for example, the principal mining laws of Central African another commune, while transportation for Republic, Equatorial Guinea, Ethiopia, Guinea, Nigeria, Sierra export purposes may take the metal through Leone, and South Sudan. 56 Figure 2.10 Mining Exploration Activities, as Indicated by Claim Area Source : World Bank, SEGOM, Burundi Artisanal Mining Study, Forthcoming. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 57 several communes (whether by road, pipeline, from appearing across communes as revenues or, potentially, train). from mineral extraction increase. While the 2013 mining code refers to this possibility, the 54. Given the potential impacts of mining actual revenue distribution mechanisms remain and the potential magnitude of mining to be specified. revenues, it may be important to clarify revenue-sharing arrangement outlined in 55. An integrated approach to the mining code. As previously noted, the 2011 decentralization and mining could be CEM estimates tax revenues from industrial- implemented. There is an opportunity to scale exploitation of Musongati alone at $50 integrate both decentralization and mining million to $233 million (2 to 10 percent of 2011 through a mechanism that: (i) compensates GDP), depending on nickel prices. Artisanal communes with mining related activities for mining revenues for Burundi could amount to environmental degradation, (ii) sets aside at least $8 million in 2014. Revenue sharing, in some resource revenues for future generations, addition to community development agreements and (iii) shares some of the resource revenues with companies, may help to compensate between producing and nonproducing affected communes for the costs of mining communes. There are examples of revenue- activities. Revenue sharing across provinces sharing arrangements (i) with producing areas and communes more broadly—even those not only and (ii) with producing and nonproducing directly affected by mining—could serve as an areas alike (see boxes 2.3 and 2.4 for details). important additional source of budget financing for communes and prevent wide fiscal disparities 56. Burundi may consider sharing mining Box 2.3 Scenario 1: Mining Revenues Shared Across Producing Areas Under a system that shares revenues with producing areas, revenues from natural resources are collected centrally, but then a portion is shared with producing areas, either based on the contribution of that area to the total revenues collected (by derivation) or based on a formula or percentage. In many cases, the revenues are shared both at the state or provincial level and at the municipal or district level. In Ghana 80 percent of resource revenues are retained by the central government, while 20 percent is shared with producing areas: 10 percent goes to a Mineral Development Fund for redistribution to mining communities and the other 10 percent is distributed to the Office of Administration of Stool Land, which then distributes the funds among producing districts, and traditional councils and customary land title holders (“stools”) in producing areas (Morgandi 2008; Revenue Watch Institute 2013). Indonesia distributes its resource revenues to producing areas and areas affected by resource exploitation only. Whereas oil and gas revenues are largely retained at the central government level, most mining revenues are retained at provincial level. Resource Revenue Distribution in Indonesia Mining Oil Gas Central Government 20% 84.5% 69.5% Producing Provinces 16% 3.1% 6.1% Producing Districts 32% 6.2% 12.2% Districts Adjacent to Producing Districts 32% 6.2% 12.2% Source: Morgandi 2008; Agustina and others 2012. Note: The central government general revenue grant to provinces is reduced as provincial revenues increase. 58 Box 2.4 Scenario 2: Mining Revenues Shared Across Both Producing and Nonproducing Areas Under this system, revenues are generally collected by the central government, and both producing and nonproducing areas receive a share of resource revenues. In these cases, however, producing areas often receive a larger share than nonproducing areas. Generally, it appears that countries in which resource revenues represent a large share of the budget may be more likely to redistribute some of the revenues to all regions (Morgandi 2008). The 1999 constitution of Nigeria stipulates that 13 percent of resource revenues must be allocated by derivation. After “first-line charges,” including cash calls to finance the national oil company, 13 percent of oil revenues are paid to producing regions by derivation. The rest is divided among the federal, state, and municipal levels (local government areas), based on a formula that takes into account factors such as population size, revenue-raising effort, and social development efforts (Ahmad and Mottu 2003; Morgandi 2008). Distribution of Nigeria’s Oil Revenues (%) Central government 45,83 All 36 states (based on a formula) 23,25 All municipalities (based on a formula) 17,92 9 producing states (by derivation) 13,00 Total 100,00 Source: Morgandi 2008. Note: Under Venezuela’s hydrocarbon law 25 percent of royalties and petroleum taxes must be transferred to state and local governments (Revenue Watch Institute 2013). revenues with producing communes only, or Cibitoke, Kayanza, Bubanza, Ngozi, Kirundo, sharing across all communes. An estimated Muyinga, Ruyigi, and Gitega, containing a total minimum of $8 million in gold revenues in a of 61 communes. If revenues were to be shared given year to the Burundian government is used not only with producing communes but also as an estimate of total mineral revenues in the with those that experience an impact from illustrative simulations discussed below. mining, communes in other provinces would need to be included in this list. It is advisable 57. In one possible scheme, revenues are that a transparent procedure be developed shared only with producing communes. The to identify all potentially affected communes, political feasibility of such scheme needs to given the potential for conflict over revenues. be carefully considered. Capacity constraints For example, some communes in a producing at the local level and budget constraints at province may not be affected, while communes the central level may make a model like Ghana’s in nonproducing communes experience impacts. (or Indonesia’s)—in which 80 percent of The overall political feasibility of such scheme revenues are retained at the central level— needs to be carefully considered by the national most practical for Burundi. In other words, authorities, and a more equitable scheme (see 20 percent of revenues would be allocated to above) might be preferred given the Burundi producing communes (see table 2.5).67 Currently, context. provinces with producing communes include 58. In the second scenario, revenues are shared across both producing and 67  Of course, the sharing ratio could evolve as the central nonproducing communes. As indicated in government transfers more responsibilities to communes. table 2.6, a simplified version of the Venezuela Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 59 Table 2.5 A Simulation of Revenue Sharing with Producing Areas Only Entity % Amount ($US) Central government 80 6 400 000 Communes with direct/indirect 20 1 600 000 producing activities All eligible communes (61 communes) Assumed equal here but share could vary with status 26 230 of commune: producing (10/20), impacted (6/20), neighboring (4/20) Source: Authors’ own compilation. Table 2.6 Simulating Revenue Sharing with Producing and Nonproducing Areas Entity % Amount ($) Central government 70 5 600 000 All communes 10 1 800 000 Communes with direct/indirect 20 1 600 000 producing activities All communes (129) Could be population based or equal per commune. Here the communes 9 302 of Bujumbura are treated as ordinary communes Communes with direct/indirect Assumed equal here but share could vary with status of commune: 26 230 producing activities producing (10/20), impacted (6/20), neighboring (4/20) Source : Authors’ own compilation and Nigeria model might also be imagined, it by creating locally taxable economic activity in which, again, the majority of revenues are that would not occur otherwise? If local tax retained at the central government level (70 capacity is indeed created, then how should it be percent), and the balance is shared equally accounted for in a transfer formula? If there are between all (10 percent) and producing costs created by small-scale mining—such as a communes (20 percent). Note that under either need for more police, water, or garbage removal scenario, the division between amounts going services—how should these be handled? Also, to the central government, and communes projections based on industrial-scale mining should be determined based on a variety of should be taken into consideration. As mineral factors, including the capacity and division of revenues increase, the capacity of provinces responsibility for the provision of services and and communes to handle large mineral revenue infrastructure. flows has important policy design implications. Public investment capacity, as well as capacity 59. To determine the “right” revenue- to handle mineral revenue volatility, needs to be sharing mechanism, if any, a number of considered and managed. factors should be considered. The immediate impact of small-scale and artisanal mining may 2.3 Future Fiscal Decentralization be a pertinent issue to consider when planning Reforms in Burundi: Possible mineral revenue sharing. How does ongoing small-scale mining activity impact the local tax Paths and Recommendations potential? Does it reduce it by taking land out 60. There are two possible paths in the of agricultural production, or does it enhance near term for communes in Burundi. The 60 first trajectory for fiscal decentralization in At the same time, there is a need to address Burundi is that of the status quo. In this scenario inequalities across communes and their need nothing (or little) changes in terms of the for more resources, in particular given the number of communes and their legal status and provision of land services is now a responsibility responsibilities, while their tax powers remain of communes (see chapter 3). largely unchanged. In the second scenario, there is a significant reform of the intergovernmental 63. Several reforms proposed above (fiscal) structure. The responsibilities of require more central-communal inter- communes—both rural and in Bujumbura— actions. Reforms in the collection of communal and their number are settled by the passage of taxes involving more activity of or interactions one or more laws in parliament. For instance, with the OBR, the introduction of a transfer Burundi could opt for a Rwanda-inspired reform system or the sharing of natural resource wealth that reduces the number of communes while will all require more interaction and more considerably increasing their service delivery integration between the central government responsibilities. and the communes. Thus it may be appropriate to examine how to establish more formal 61. In the short-term (up to the 2015 consultative mechanisms between the central elections) the most likely path is the government and communes. Such a body could maintenance of the status quo. If reforms for example be set up along the line of a bipartite were to be undertaken, their likely shape is consultative committee. One may also want to uncertain. Since 2010, when a few amendments add one communal representative to the seven to the communal law were introduced, there member board of the OBR should it now collect has been little progress toward effective taxes for communes. decentralization in Burundi. The law on commune-level responsibilities has yet to be 64. To address the particular needs of adopted since parliamentarians deplore the lack Burundi’s communes and manage fiscal of appropriate transfers, human and financial, decentralization reforms, the following to communes. The laws on communal finances seven policy actions are recommended: and on changes to the status of Bujumbura’s 1. Clarify the functional responsibilities communes have also not been adopted. As of communes in the long term. It is not discussed in chapter 1, various factors are possible to decide on the proper size of relevant, such as the division of ministerial portfolios between two vice presidents and communes—and thus on their number the anticipated elections of 2015. These and or financing—until their responsibilities other factors suggest that no major reforms are clearly ascertained. Adding respon- are imminent; meanwhile, policy in this area sibilities—as envisaged in the transfer remains uncertain. of responsibilities law currently under study—is one option, but their financial 62. For both paths, the introduction of implications need to be carefully considered a recurrent expenditure transfer scheme (see # 3 below). Not adding responsibilities is appropriate. The appropriate degree of is another, and equally reasonable, option. deconcentration and devolution in Burundi But it is advisable that, either way, the is not clear. Burundi is a small country, and responsibilities are clarified in the short thus the span of control by the central govern- term (2014–16)—preferably before the ment can easily cover it. It uses a common elections of 2015—either by: (i) adopting language and is not divided geographically in parliament and promulgating the law into linguistic or ethnic groups. Thus many of now under study for implementation in the typical reasons for devolution do not hold. 2016, or (ii) stating explicitly that existing Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 61 responsibilities will be unchanged during creating an urban community is a medium- the medium and long term. term choice (three or more years). 2. Settle the number of communes—and 4. Revise existing tax parameters to take thus their boundaries—for the long term, inflation into account and establish a more and allow intercommunal collaborative effective approach to local taxation, in arrangements to maximize economies collaboration with OBR. The tax potential of scale. Unless accompanied by more of communes may also be evaluated, after far-reaching administrative reforms, it setting the new parameters. At the same is further unlikely that the consolidation time, efforts are needed to strengthen of communes—under existing respon- local tax administration system, improve sibilities—would create more fiscal space tax compliance and strengthen the link for local services through costs savings in between improved revenue collections and local administrations, and may actually raise better services for the citizens. This could additional costs for residents of communes. be facilitated by an exchange of information Moreover, a comprehensive infrastructure between OBR and communes regarding the plan is needed to ensure complementarities lists of taxpayers, as a first collaborative step between the delivery of important services— to facilitate collection of local taxes related for example, the road network and health to economic activities. (A similar example services. A more viable alternative for the of collaboration between the central tax short team (one to two years) would be to agency and the local ones can be found in take advantage of the new decree adopted Rwanda, where the last reform of the District by Government in October 2013 and start tax law introduced the requirement that the implementation of intercommunal RRA provide information on one category schemes of various dimensions according of taxpayers—VAT taxpayers—to districts). to the level of economy of scales associated The changes in the tax parameters can be with various services. introduced in the short term—by or before 3. Settle the status of the communes 2015—by adopting and promulgating the of Bujumbura, taking into account relevant law. The partnership with OBR their location in a large metropolitan can also be implemented in the short term area. There are three options at this through policy directives. A revaluation of time: making these full communes, with communes’ tax potential could be carried a status similar to those of communes out in the medium term, also to take in the rest of the country, keeping their advantage of an expanded land register (on status unchanged, or designating them as this, see chapter 3). boroughs (arrondissements). Meanwhile, 5. Introduce a scheme for funding current the boundaries of Bujumbura’s Mairie need expenditures that complements the to be examined to see if they should be existing capital funding transfer scheme expanded to account for the transformation (FONIC). This requires examining the of neighboring communes into suburbs. balance between communes’ own revenue One may want to examine the creation of a potential and their expenditure needs. It is Bujumbura urban community grouping both important that the maintenance of capital is Bujumbura Mairie and some communes of not neglected while new capital is funded. Bujumbura rural. Any decision on the status The first priorities are to increase own- of Bujumbura communes can be taken in source revenues and introduce a general the short term (one to two years). Changing transfer scheme. Also, the existing capital the boundaries of Bujumbura’s Mairie or transfer scheme (FONIC) can be improved 62 in the short term. Before these items are provide the authorities with an effective implemented, the financial capacity of steering tool, while also contributing to communes will be in flux and thus their strengthen transparency and accountability repayment capacity hard to establish. Then, in the use of public resources. a lending mechanism could be implemented 7. Undertake a systematic investigation in the medium term (fewer than five years), of the one-off, recurrent costs, and once these other issues are clarified. A service delivery costs of communes’ transfer scheme for current expenditures current and new responsibilities, to could be introduced in the short term, properly understand the long-term fiscal through the 2015 budget and separate from sustainability of the decentralization the law changing the tax parameters. A reform process. This follow-up analytical medium-term review of this scheme should work can focus on (i) costing the existing be carried out in 2017 for implementation staffing norms, institutional arrangements, in 2018. and prevailing unit costs for key services; and 6. Introduce a monitoring mechanism (ii) estimating options for consideration in permitting to better capture the terms of factors/variables that Government geographic distributions of subnational should consider in order to ensure fiscal expenditures in the central budget, sustainability, should new service delivery to improve budget transparency and responsibilities be devolved. It might be effectiveness of public spending. There advisable to start with sectors that have a is currently no mechanism in the national major contribution to local service delivery budget for systematically identifying the (including health, education, agriculture). distribution of public expenditures, both in The key issue at stake is the proper use the case of deconcentrated and decentralized and remuneration of human resources spending. Hence, while it is possible to learn in a decentralized setting. In this regard, the overall budget ceiling for a given sector, it a more extensive analysis of local public is difficult to obtain its disaggregation at the sector expenditures would be helpful in subnational level. A mechanism permitting determining the optimal contribution the monitoring of the subnational allocation of communes to public service delivery of expenditure in the central budget would outcomes. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 63 Chapter 3 Decentralized Land Administration: Challenges and Opportunities Summary of the Main and others 2013) and results from interviews and fieldwork visits to SFCs in Ngozi province. Arguments and Key Findings The main policy innovation discussed is the Moving from an analysis of macro-level introduction of a more relevant, accessible, and institutions to the local service delivery cost-effective way of registering rights to rural responsibilities of communes, this chapter land than the prevailing land titling system. This innovation builds on the broad presence reviews the experiences of decentralized and authority of communes and introduces the land administration services, and discusses certificate as a legal alternative to the title (a the challenges and successes of the certificate can later be converted into a title, if government’s ongoing efforts to scale up desired). To this end, this chapter will: access to these land services across all the communes in Burundi. To this end, this chapter • Identify the main obstacles to land builds on the results of the land sector review tenure security and their implications for undertaken in 2013 for the World Bank’s Projet conflict and violence in Burundi. This will de Productivité et de Développement des Marchés set the stage for a discussion of the rationale Agricoles (PRODEMA Report 2014) as well as to promote land governance reforms and a the cost-benefit analysis of selected communal review of the key innovations introduced in departments of land registration (the Service the post-conflict period. Foncier Communal or SFC) undertaken with • Present the results, strengths, and financial and technical support from the Food weaknesses of the new land certificate and Agriculture Organization (World Bank/FAO and demonstrate that the combination of a 2014), and other recent studies (Habwintahe decentralized system of land administration 64 and a new document to record rights (the decentralized land services, combined certificate foncier) have the potential to with strategies to ensure the financial improve land tenure security in Burundi sustainability of SFCs and the efficiency and minimize conflicts over land and gains of communes, as well as to contribute investments. to efforts to mobilize revenues through • Discuss the outcomes of pilot operations local taxation, land-use planning, and better in establishing SFCs, with a focus on information. results and pending challenges. The aim 3.1 Land Scarcity and Tenure is to allow for land tenure registration—at scale—that is accessible, less costly, uses Insecurity as Drivers of Conflict a faster procedure, and is embedded in a in Burundi communal process to ensure transparency 1. A mountainous, landlocked country, and to contribute to dispute resolution/ Burundi has one of the highest rural prevention. In particular, the analysis population densities in Africa (310 people/ suggests that an impact evaluation exercise square kilometers [km2] on average); 90 percent is urgently needed to rigorously assess of the total population live in rural areas. There potential household benefits. To date, there is significant variation in population density is limited quantitative evidence of how across regions (see figure 3.1), especially when communal-level registration of land rights and issuance of land certificates affect (i) the excluding Bujumbura-Mairie.68 The amount long-term prevention of disputes and the of arable land is estimated at 35.8 percent, efficacy of conflict resolution mechanisms and cultivated land at 15.6 percent in 2011 (a across social groups, (ii) investments, (iii) marked increase from 14.0 percent in 2000; land markets, and (iv) transformation of World Development Indicators [WDI]). Farming rural livelihoods. is the main livelihood for most of Burundi’s rural households, making land the main economic • Define the sequence of reform solutions asset for the majority of the poor. needed to improve land tenure security and ensure the sustainable and efficient 2. In Burundi small landholders scale up of current reform efforts. These dominate the agricultural sector. In recent questions are at the forefront of the national years most farmers have experienced both land policy debate. The government is keen to a reduction in crop area and productivity;69 obtain technical support from the Bank and the average farm is around 0.4 to 0.5 hectares those donors active in the sector, as it aims (ha; 0.1 ha per person in 2011; WDI) and is to decide the most effective and sustainable only shrinking as a result of subdivision law mechanisms to strengthen land governance (PRODEMA Report 2014). Meanwhile, the waves in rural areas, as well as the actions needed of violent conflict undermined the crop rotation to successfully prepare for the scale up of bananas and livestock, as the loss of livestock of ongoing pilots to all rural communes. Specifically, the long-term sustainability of SFCs (in terms of ensuring the quality 68  Bujumbura, the only city of significant size, has fewer than 1 and integrity of records, sustaining the million inhabitants, making it relatively small for a country whose requisite human resources and technical population currently stands at nearly 10 million inhabitants. The country second city, Gitega, has fewer than 50,000 inhabitants. skills, and financing) has to be addressed. Scaling up requires investments in capacity- 69  Government of Burundi. Second Strategic Framework for Growth and the Fight against Poverty (Cadre Stratégique de building initiatives and improvements Croissance et de Lutte contre la Pauvreté II (CSLP-II), p. 16–17. in accountability and transparency in 2011. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 65 Moreover, settlements carry high costs for those Figure 3.1 The Population Density of Communes, by Quintile engaged in land disputes. Commune officials are frequently requested by the involved parties to act as mediators in land conflicts.72 Improved land administration can reduce the burden on scarce human resources by mitigating conflicts and speeding up dispute resolution. 4. Conflicts and disputes over land are widespread. Land disputes may be linked to political and ethnic divides, but the majority involves members of the same family or long- time neighbors. Conflicts may be divided into two categories: (i) structural conflicts, prompted simply by scarcity and demand, and (ii) conflicts due to political crises and 5th (highest) associated displacement and violence. So-called structural disputes are usually over boundaries, 4th 3rd 2nd 1st (lowest) Bujumbara Mairie transactions (of family land,73 multiple sales of the same plot, and so on), rights to land, and Source : ISTEEBU 2013 inheritance. As shown in table 3.1, disputes over “inherited plots” are the most frequent (58 percent) and the most difficult to resolve. More affected soil fertility management strategies. than 500,000 refugees who left in the 1970s have The growing pressure on land70 also reduced the returned to Burundi over the past several years possibility of using fallow to restore soil fertility, and are seeking to regain ownership of their while the spread of annual grain crops, such as land, now used by others (PRODEMA Report maize, on the commonly steep slopes increased 2014). But disputes that are due to the political soil erosion. Bare slopes are prone to erosion, crisis involving returnees are relatively few in which can in turn exacerbate flooding in valleys number in most areas, except in paysannats and plains, as happened in February 2014.71 schemes; 20.6 percent) and in urban areas (see table 3.2). 3. The demand for land (for farming, urban expansion, investments, 5. The majority of pending cases in environmental protection, and so on) Burundi courts are related to land conflicts. continues to rise. A combination of increasing It is estimated that about 72 percent of civil demand, growing land scarcity, and ambiguous court cases are linked to land issues; meanwhile, land rights undermines stability at the local level. land disputes are often behind criminal cases.74 70  Burundi’s population density has changed significantly in 72  The mayor (administrateur communal) in Ngozi estimated recent years, going from 144 inhabitants/km² in 1979 to 310 that she spent two days a week, on average, in the mediation of inhabitants/km² in 2008. This tremendous increase is due to the conflicts often caused by land disputes (interview, June 2013). high population growth, spurred by high fertility rates. 73  Land sales can be a source of conflict, especially when one 71  Forest management on hill tops, multiannual crops individual makes the sale without the consent of the spouse and (bananas, coffee, and tea), combined with terracing and other family. forms of erosion protection, require that farmers have tenure security and the capacity to invest, supported by (enforced) land- 74  Bigirimana, Gilbert. “Étude sur la situation des conflits fonciers use regulations for fragile areas. au Burundi.” (In Ngozi province, interviews were conducted with 66 Table 3.1 The Frequency of Land Disputes, Ranked by Household Type and Method of Acquisition Marshland Rural Urban paysannats Households near % (marais) Households Households Households Protected Areas Households N = 2257 N = 1047 N = 385 N = 306 N = 299 N = 220 Inherited plots 57,9 % 53,4 % N = 306 34,2 % 36,9 % Purchased plots 25,6 % 30,8 % 69,8 % 27,4 % 20,9 % Plots attributed from public 3,7 % 3,8 % 21,4 % 11,0 % 20,9 % domains Plots claimed by returning 6,4 % 5,4 % 3,2 % 20,6 % 9,7 % refugees Other 6,1 % 6,6 % 0,4 % 3,9 % 10,8 % Source: PRODEMA Report 2014: 77. Note: Ranking provided by participants in focus group. Percentages indicate how frequently risks were ranked as either first or second in order of importance. The mandatory engagement of the traditional Burundi to promote joint titling, which requires conflict mediators, the bashingantahe, before the registration of both the husband’s and wife’s referral to a court was abolished in 2005 even names on the land certificate—even though this though in practice they are still being consulted. is legally possible. Few women own land; within To date, the only way for Burundian courts to customary laws women are denied rights to land resolve land conflicts is to rule in favor of one ownership through inheritance. Widows do not of the parties involved. But difficulties with inherit land from their husbands but are mere corroborating evidence, coupled with suspicions caretakers of their sons’ inheritance. Family of judicial bias make such resolutions particularly law does not acknowledge equal rights, and the difficult to achieve. The current legal framework parliament has several times rejected proposals was identified as a major obstacle to conflict to change the law in this respect (most recently resolution during consultations conducted in 2013). According to Burundian law, women in 2011 by the Ministry of Justice. Moreover, will only inherit their parents’ land when there the judicial system lacks a formal mediation are no sons. Meanwhile, under customary mechanism (ICG 2013). laws brothers have an obligation to support their sisters in times of hardship by providing 6. Systematic discrimination against access to land (igiseke). Women may also farm women exacerbates tensions over land part of their husband’s land, but without rights ownership. Gender equity receives surprisingly over it, and may not be consulted when these limited attention in land policy and legislation plots are sold, leased out or used as collateral (ICG 2013). Unlike countries such as Rwanda (Habwintahe and others 2013). and Ethiopia (see box 3.1), there is little effort in 7. The Batwa represent a category of citizens without land. They are an ethnic 240 individuals, analyzing more than 1,600 court judgments as minority group that composes about 1–2 well as collected opinions from about 230 administrative, judicial percent of the total population, lacks secure and customary) in “Les Terres de la discorde: la reforme fonciere au Burundi (Field of Bitterness I: Land Reform in Burundi” ICG rights to land. This is exacerbated by their weak Report, 2014. social status in Burundian society (even though Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 67 they are acknowledged as the first inhabitants of the Batwa are acknowledged problems, and the region). Batwa have forest-based livelihoods governments have on occasion allocated public and are affected by centuries of deforestation land to some groups of Batwa (PRODEMA Report aimed at creating more farmland. The total 2014: 47–50), although the documentation forested land left in Burundi is estimated at proving these allocations may be lacking or 6.6 percent (WDI 2011). The limited rights of incomplete (Habwintahe and others 2013). Box 3.1 Gender Equity and Land Ownership: Learning from Ethiopia and Rwanda In Ethiopia gender equity in land rights was improved when, during land redistribution, in the 1990s unmarried adult women were allocated land. Women’s and men’s land titling rights was fully recognized under the 1997 Federal Proclamation. Through the Land Certification Program that started in the late 1990s and lasted until the mid-2000s, efforts were initiated to strengthen women’s rights to land by formally recognizing married women’s rights to land, which was now registered jointly with spouses. Under this program, women received land certificates either jointly or singly, with certificates that carried the names and photos of both husband and wife (Deininger, Ali, and Alemu 2011a; Bezabih and Holden 2010; Securing Land in Africa, box 3.4). This initiative required active support from decision makers, active facilitation, and the representation of women on the committees accompanying the registration and rights recognition processes (Teklu 2005). Rwanda addressed gender discrimination on land rights through a comprehensive land sector reform, backed by extensive legal and institutional policies—initiated at the end of the 1994 genocide—in support of a nationwide land certificates registration program. Subsequent amendments to the national land policy in 2004 coupled with the Organic Land Law in 2005 further elaborated the governing land laws and established several new clauses, notably: (i) the equal rights of daughters and sons to inherit their parents’ properties, (ii) the protection of women’s property rights when in legally registered marriages, and (iii) the mandated consent of both spouses in case of sale or mortgage of the property by one party (Government of Rwanda 2005). An evaluation of the pilot program in preparation for its expansion showed, however, that most marriages were not formally registered, and as a result, women were losing out. This was corrected in the rollout of the program, when informal marriages were included in the appropriate provision. Recently, a study analyzing the short-term impact of the land registration programs demonstrated that married women’s access to land had improved. The recording of inheritance rights has become more gender neutral, and most important, investment in and the upkeep of soil preservation methods has increased, specifically among households headed by women. The land registration program in Rwanda also demonstrates that the systematic registration of all rural land parcels is feasible if well prepared (with a legal and institutional framework in place before rollout). But although contested plots will not be registered, any latent conflicts—such as those involving returning refugees—may go unmentioned at the time of registration. As of now, there are 11,809 parcels recorded as under dispute, representing only 0.1 percent of all land registered in Rwanda—a figure that seems far too low to be accurate. Another issue is keeping the registry current; registered transactions number 10,535 out of a total of 10.6 million, or only 0.1 percent. Finally, the government’s use of the land registration database to direct land use (farmland consolidation and villagization) has generated controversy. Sources: Ali, Deininger, and Goldstein 2011; Byamugisha (2013: box 3.4). 68 3.2 Background: Types of Tenure Droits de la Personne Humaine et du Genre) also under the first vice presidency. The mandate for and the Institutional Setting the CNTB was renewed in 2009, and in 2010 8. The institutions in charge of land it was placed directly under the Office of the governance in Burundi fall under the President of the Republic.76 Its judiciary powers presidency and two vice presidencies. Burundi were increased substantially in 2013.77 has a National cadastre Department with offices 9. The main types of tenure are private in Bujumbura, Gitega, and Ngozi. This semi- lands (terres des personnes privées physiques autonomous department is under the tutelle of ou morales) and public (state land). Most the Ministry of Water, Environment, Land Use smallholders farm private lands(terres privées and Urbanism; (Service national du cadastre— non-titrées) under a status that is recognized by administration personnalisée sous la tutelle de the government although not formally registered, MEEATU—Ministère de l’Eau, de l’Environnement, or deriving from customary rights. These private de l’Aménagement du Territoire et de l’Urbanisme) landholdings evolved from customary holdings; as well as a land titling registry under the tutelle user rights may exist but are not recorded, and of the Ministry of Justice (Direction des Titres are mainly held by women on the land of their Fonciers sous la tutelle du Ministère de Justice). husband, father, or brothers.78 According to the The General Directorate of Land Use (Direction law, ownership rights may be acquired after Générale de l’Aménagement du Territoire, et la having been used for more than 30 years without Gestion du Patrimoine DGAT) in the MEEATU is this use being challenged. But this legal clause responsible for the management of rural public in the land law is controversial: most refugees land (Terres domaniales en milieu rural), while the who left in the 1970s did not return within the General Directorate of Urban Planning (Direction 30-year period.79 The PRODEMA study found Générale de l’Urbanisme, DGU—MEEATU) that one-third of households interviewed use manages urban public land. The MEEATU plots that are leased from other private holders, is represented at the provincial level, and is are leasing public land (such as marais) or are also the tutellle for the National Land Policy in paysannats and holding annual renewable Program (Unité de Coordination du Program leaseholds (PRODEMA Report 2014). National Foncier, PNF)75, which is responsible 10. Public land was created when colonial for coordinating donor engagement and authorities expropriated all nonpermanently coordinating and monitoring (suivi operationnel) farmed land (forested, fallow, and land policy reform, such as decentralized land grazing), which remained state land after administration. An interministerial Steering independence. Public or state land can be Commitee for Land Sector Reform (Comité de owned by the central government or communes pilotage de la réforme du secteur foncier) also (terres du domaine de l’Etat—Les terres du includes the Ministry of Justice, under the first vice presidency as well as a National Land Commission. The National Commission on Land 76  Projet de loi y relatif a été adopté par l’Assemblée Nationale and other Properties (Commission Nationale ce 24 décembre 2010. des Terres et autres Biens, CNTB) was created in 77  Loi N°1/31 du 31 décembre 2013 portant révision de la loi 2006, to serve three years under the Ministry of N°1/11 du 4 janvier 2011 portant missions, National Solidarity, Human Rights, and Women’s Rights (Ministère de la Solidarité Nationale, des 78  Igiseke is the term for user rights in local language. 79  About 500,000 people (or 5 percent of the population) fled the country around 1972, and many did not return to Burundi until after 2002 (over 30 years later, meaning that formally they 75  Ordonance ministérielle No:1218/du 12-8-2010. could lose claims to their land). Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 69 domaine des Communes et des autres personnes the initial government practice of rewarding morales de droit public). Public land consists rights to those who clear plots and make them of two types: (i) the Domaine de l’Etat Public— productive has been gradually abandoned. protected forest areas, parks, and marshlands The marais and bas-fonds are also the focus of (marais), which cannot be privatized; and (ii) agricultural investment projects and donor the Domaine de l’Etat Privé, which is state land interventions. In the case of investment projects, that can be allocated by authorities to private the risk is that plots’ current occupants will be parties. This latter category includes fertile displaced without recognition of their rights or lands such as the marshlands (bas-fonds) and any type of compensation (PRODEMA Report the agricultural schemes (paysannats) in the 2014: 50–52). As a result, tenure insecurity fertile plains, where cash crops (cotton, rice) is high in these areas (Ouédraogo and others were cultivated during colonial times, as well 2007; 2010), and the new land law (2011) has as the palm oil plantations. There are only a not resolved the issue of many plots’ ambiguous few public lands and they are unregistered; status. In addition, the marais are public lands neither the government nor the commune has that legally belong to the central government, a complete overview of all its lands which may yet some communes also claim ownership over be encroached on by smallholders (PRODEMA the marais and collect rents (PRODEMA Report Report 2014: 38). The lack of accurate maps 2014). hinders the management of public lands. But the key governance issue surrounds the procedures 12. The registration of land rights is not for official attribution of public lands to private common practice. Of the urban households parties, a process that is not always transparent, interviewed for the PRODEMA study, only respecting procedures or serving the public 40 percent reported having a title. Other interest (PRODEMA Report 2014). State land documentation included a purchase contract— is becoming scarce. For instance, inventories not necessarily officially validated (29.4 in a few communes near Gitega found that percent)—an affidavit (acte de notoriété; 16.7 only 8 percent80 of total land area is public. An percent), and an attribution note for former inventory of all public lands is planned with public land (4.3 percent) to sustain their the support of the European Commission and claim (PRODEMA Report 2014: 83–84—draft Gesellschaft für Internationale Zusammenarbeit version).81 In rural areas titles are very rare; (German Agency for International Cooperation, but the trend is to record transactions at least informally (petits papiers). The PRODEMA GIZ). (2014) study, using a small sample, found 11. The marshlands are the last that 58 percent of buyers did not register the agriculture frontier in Burundi and cover transaction formally, 29 percent registered about 120,000 ha. Parts of the marshlands with the authorités collinaires, and 15 percent (marais) are protected areas according to the registered at the communal level, confirming code de l’eau, but the distinction between the transaction via the acte de notoriété. protected marais and cultivable bas-fonds can 13. Tenure insecurity stems from multiple be difficult to ascertain (Habwintahe and others risks, both real and assumed. According to 2013). Despite the risk of flooding and other the PRODEMA study (2014), a primary source difficulties surrounding their cultivation, these of perceived tenure insecurity is government fertile lands have been increasingly sought after since the 1980s. As demand increases, 81  Survey covering 159 households and 211 parcels—which is 67 percent of all households in the sample. The others did not 80 Interview, PNF, June 2013. reply to the question. 70 Table 3.2 Perceived Risks to Land Rights, Ranked by Household Type and Source Marshland Rural Urban “paysannats” Households near % (marais) Households Households Households Protected Areas Households N = 2257 N = 1047 N = 385 N = 306 N = 299 N = 220 Court case by a family 25,6 % 15,8 % 25,0 % 5,7 % 7,5 % member Court case by a neighbor 14,3 % 8,3 % 11,0 % 4,1 % 4,3 % Court case from a 3,7 % 0,8 % 5,0 % 10,7 % 3,2 % returning refugee Court case from other 0,3 % 0,0 % 2,0 % 7,4 % 0,0 % residents Questioning of rights by a 13,7 % 18 % 16 % 24,6 % 16,1 % third party Expropriation by the 10,4% 9,0 % 9,0 % 10,7 % 26,9 % state Expropriation by the 9,1 % 22,6 % 8,0 % 23,8 % 2,2 % CNTB Other 14,0 % 15,8 % 18,0 % 8,2 % 11,8 % Source: PRODEMA Report 2014. Note: Ranking provided by participants in focus group. Percentages indicate how frequently risks were ranked as either first or second in order of importance. intervention. Study respondents reported fear documenting transactions (2014: 86).82 of (i) expropriation or eviction by the state, as in cases of land allocation to companies like 3.3 Promoting Land Governance the Imbo Regional Development Corporation Reforms: Rationale and Policy (Société Régionale de Développement de l’Imbo, SRDI) or—in Gihanga and Mutimbuzi—for Objectives sugarcane, palm oil, and the creation of military camps, (ii) withdrawal of paysannats’ rights to 14. With the Arusha Peace and land sought after by investors, (iii) rezoning of Reconciliation Agreement signed in 2000, land use (that is, the creation of protected areas land restitution and land governance reform such as marais), and (iv) land confiscation by emerged as key priorities in efforts to the CNTB or other state authorities. In addition, address the root causes of conflict. The Arusha standard expropriation procedures were not agreement proposed a set of measures connected respected, with the result that compensation to (i) the restitution of land to citizens affected by was delayed or not paid at all. A second source the violent conflicts, and (ii) the reform of land of tenure insecurity is found at the community governance. Given the large number of refugees and family level. For example, property rights may be challenged by well-off neighbors, forcing smallholders to defend their rights 82  In the wake of expanding land markets, semi-formal in court at great costs. The PRODEMA study initiatives to record land transactions emerged (petits papiers) or reported that, of the households interviewed, buyers of land also applied for an affidavit (acte de notoriété) to the municipality, for which a percentage of the transaction value had 33 percent expressed interest in the possibility to be paid (10 percent, reduced to 3 percent by most communes of documenting their rights, and 20 percent in following a declaration by the president of the republic). Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 71 consultations - both at the central and provincial Box 3.2 Land Governance: A Definition levels - which have received ODA financing and technical support.84 The land code (2011) lays Land governance is defined as the process by out the procedure for registering property rights which decisions are made regarding rights over using either a title or a certificate. land, the access to and use of land and natural resources, the manner in which those decisions 16. Current land titling procedures have are implemented, and the way that conflicting their origins in colonial times and were interests are reconciled. Characteristics continued after independence. The land title of land governance are: (i) the legal and (titre de propriété) is issued by the Registrar of institutional framework for rights recognition Land Titles (Conservateur des titres fonciers) of and enforcement; (ii) land-use planning, land the Directorate of Deeds (Direction des titres management, and taxation; (iii) management of public land; (iv) public provision of land fonciers, DTF) within the Ministry of Justice. It information (registries, cadasters, and so cannot be challenged in courts unless the title on); and (v) dispute resolution and conflict issued is a product of fraudulent acts85. Almost management. all land titles concern urban lands (residential, Source: Palmer and others 2009; Deininger and others business); holders of large tracts of rural land 2011. also have titles. The DTF estimates the total number of titles in Burundi at 62,000, of which 58,000 are issued in Bujumbura and most others and internally displaced people in Burundi, it is in the cities of Gitega (854 titles in 2013) and not surprising that land issues are a key element Ngozi. The DTF estimates that only 1 percent of in the Arusha accord and a focus of postconflict rural land is titled (PRODEMA Report 2014: 75– donor and nongovernmental organization 76). Titling via the classic land administration (NGO) programs, such as mediation and conflict system tends to be a long and expensive process, resolution. With the establishment of the CNTB particularly due to requirements for precise in 200683 a procedure for land restitution to surveying processes, as well as the lack of human refugees and displaced people was developed, and material resources of the administration. while the land reform and the revision of the This procedure is unaffordable for the majority 1986 Land Code took longer to materialize and of the population—even in urban areas, let is only just beginning (ICG 2013). alone in rural areas (Gapin 2011).86 Accordingly, 15. Several donors have partnered with titling is neither appropriate nor affordable for the government on the land governance reform process, culminating in the approval of the Land Policy Letter (2010) and Land 84  Les principaux financements ont été assurés par la Code (2011). The policy dialogue centered coopération suisse, l’Union européenne, USAID, les Pays-Bas et le Fonds international de développement agricole (FIDA). on developing a land policy, the revision of the 1986 land code, and developing innovations 85  But the registry is prone to irregularities. in land administration to facilitate securing 86  Before submitting their demand to one of these offices rural land rights as a strategy to reduce land- (Bujumbra, Gitega, Ngozi), applicants must hire a geodetic surveyor from Bujumbura, report the precise land limits to the related conflicts. In this context, the adoption of national cadastral services, and demarcate the land boundaries the letter on land policy and land code reform with stones made of imported concrete. The majority of people in the countryside, accustomed to marking land boundaries at the was the result of several years of participative most with perennial border plants, are totally unfamiliar with these procedures. In addition, for most of them, the price they would have to pay for the concrete is already higher than the market price of their land plots, not to mention the cost of the surveyor 83  With financial support from the UN Peace Building fund. and the taxes for cadastral services (Kohlhagen 2010: 4). 72 smallholders, as the costs are high compared to and implemented in many countries (Deininger the value of the land. The number of parcels to and others 2003; Byamugisha 2013). be registered is at least 8–9 million.87 18. The central government has 17. Improving capacity of the state’s established a second, alternative way registration authorities could help improve to record land rights: the land certificate tenure security in urban areas and for the (certificat foncier). The central government has private sector but not in rural areas in the decided to establish a second, alternative way short to medium term. Investments are taking for recording land rights in rural areas only, place to improve capacity and strengthen the land which will serve the demands for boundary administration system (archives, digitalization, and rights registration (through the certificate) training, and so on), which is important for and will be accessible through communes.90 urban areas where informality may concern The innovations consist of two key elements: 60 percent of the properties. The capacity will (i) establishing land administration services remain insufficient to serve rural areas where at the commune level, delivered by communes; most land conflicts are found and which is the and (ii) introducing an additional document— focus of most policy concerns. A study conducted for rural lands only—to register interest. The in 2003 pointed to the urgent need to address Land Policy Letter (LPL 2010) and Land Code rural land governance issues to prevent future (LC 2011) introduced the establishment of outbreaks of violence (bombe foncière) (ICG decentralized land administration services, by 2003). Several international organizations then creating communal land services (the Service commissioned research to identify key land Foncier Communal, SFC) for the registration of issues and to establish priorities for action.88 existing land rights, land transactions, and other The problems identified included: widespread, mutations. long-lasting land conflicts—intensified by the reintegration of refugees seeking the return of 19. Land administration has been their properties—the weak capacity of local decentralized to help reduce land disputes, mediation and dispute resolution mechanisms, and improve tenure security. The new inequality in access to land, the weak oversight of certificates for rural land offer almost the same public authorities, perceived land grabbing, and security to full land titles, but at significantly an overall lack of agrarian policy. One priority less cost to rural landholders. The certificate for action identified was that an alternative to is accepted as evidence in courts and carries the titling system was required.89 Making the more weight than informal documentation and registration of rural land rights feasible and witnesses (Habwintahe and others 2013).91 The affordable for the poor has been identified as a key strategy for reducing land-related conflict. Burundi is not unique, as the need for reform of 90  Article 313 code foncier (Loi n°1/13 du 9 août 2011 portant révision du code foncier) states : Le droit de propriété foncière land registration systems is widely acknowledged peut être établi : soit par un titre foncier établi par le Conservateur des Titres Fonciers ; soit par un certificat foncier établi par le Service foncier communal reconnaissant une appropriation régulière du sol se traduisant par une emprise personnelle ou 87  At least 1.6 million rural households with on average two collective, permanente et durable, selon les usages du moment et plots of land. du lieu et selon la vocation du terrain. (Land ownership may be established: either by a land title issued by the Registrar of Land 88  Notably, several detailed field studies have been conducted Titles; either by a land certificate issued by the Communal Land by mostly NGOs and UN agencies (RCN Justice & Démocratie 2004; Service recognizing acquisition of land personally or collectively Ntampaka and Mathieu 2009). in a permanent and sustainable way, according to the custom of the time and place and in the use of the land.) 89 CED-caritas 2005; Global Rights; Kamungi and others 2005; Van Leeuwen 2004; Ntampaka 2005. 91  In addition, the process for converting a certificate into Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 73 Box 3.3 Decentralizing Land Management: The Experience of Madagascar The Madagascar National Land Program (Programme National Foncier) was formally initiated by the Government of Madagascar in 2005—supported by donors—to modernize land governance though the decentralization of land administration. The program design included the introduction of land certificates (certificats fonciers), the establishment of communal land offices to oversee local-level land administration (a newly introduced service provided by communes), alongside the modernization and computerization of existing land and surveying services, the renewal of land laws and regulations, as well as the training of decentralized land management personnel (Byamugisha 2013: 62). This new procedure for the local recognition of untitled private property rights has dramatically reduced the time and cost for individuals to obtain full legal documentation proving their ownership. The land reform process changes the concept of public land and the principle of presumption of state ownership of land to enable the formal recognition of what is de facto private land, even although it is not titled. The reform process included the introduction of land certificates as well as the establishment of a “local government land office” as a legally competent service provided by elected local governments that deal with terre privée—non titrée—that is, private landholdings without a land title. The legal value of the land certificates is almost identical to that of land titles. Most transactions permissible with land titles are also possible with land certificates, such as sale, lease, partition, and inheritance. The decentralization of land management has greatly increased the number of actors involved in the national task of registering land rights. Since 2006, 546 of 1,550 local governments have established their own land office, which employs at least one land officer. Each local government has set up a Rights Recognition Committee (Commissions de Reconnaissance Local). On average there are 10 such committees in each local government with at least 3 members, or about 15,000 people, in the 546 communes, who are involved from time to time in vetting land rights (World Bank 2014: 10). The average cost of a land certificate has been drastically reduced to around $14, with an average processing time of about seven months. (Byamugisha 2013: 62). Since 2005 local government land offices have registered more than 170,000 applications and issued 96,000 land certificates (Byamugisha 2013; World Bank 2014: 10).This is compared to the previous system, which produced 2,000–3,000 land titles per year, in a process that involved 24 steps, and took 6–10 years on average, at an estimated $500 per title. Madagascar started with an on-demand approach and is now experimenting with the systematic first-time registration of rights to improve access to certificates by reducing their cost. two systems of rights registration are not in Inspired by experiences in Madagascar (see competition but in fact complement each other box 3.3), the issuing of certificates by local and incremental (PRODEMA Report 2014). governments follows a clear process that is Transactions permissible with land titles are codified in the Land Code of 2011; special also possible with land certificates, such as sale, decrees to further guide implementation are in lease, partition, and inheritance, and a certificate preparation (2014). can be converted into a title, if so desired. To mitigate overlapping claims, the first step of 20. Using the certificate to document an SFC is to make a georeferenced inventory of land rights is also expected to facilitate land all public and titled land in a given commune. transactions or markets (rental, purchase). Rural land markets are expanding, particularly in the provinces near Bujumbura, where prices are a title, if desired by the landholder, is laid out in a new decree relatively high compared to farmers’ revenues draft decree on the certification of Land Rights (Projet de Décret portant Certification des Droits Fonciers) that is to be approved by (Wennink, Lankhorst, and Irutingabo 2012). the council of ministers. According to the PRODEMA study, 14.9 percent 74 of the rural households interviewed had sold notoriété), issued by communes, that provide land. Those interviewees who were asked how a quasi-legal status to purchases (Leisz 1996), plots were acquired indicated that, on average, as mentioned earlier. The cost of an acte de 47.8 percent of plots were inherited and 41.2 notoriété was once 10 percent of the value of percent purchased (PRODEMA Report 2014: the transaction; this was reduced to 3 percent 80). Habwintahe and others (2013) report that by many communes following a presidential 5 percent of the plots that were documented declaration before the senate in 2008 (to on maps as part of a systematic registration facilitate access to returning refugees). The costs effort (but for which no certificate had yet been are generally paid for by the buyer of the land. issued) were sold in a period of about 12 months. The legal basis for an acte de notoriété differs Purchasers of land have demonstrated their from that of a tax on land and property transfers. readiness to pay for securing land transactions by Also in 2008, the senate issued a law abolishing applying for an affidavit (acte de notoriété) from taxes on sales of land and other property (Loi the commune, and are certainly most interested no1/31 du 4 novembre 2008 portant suppression in obtaining certificates. A study in a commune de la taxe de transaction des propriétées foncières near Gitega found that 39 percent of the plots et d’immeubles) (Habwintahe and others 2013: being registered were purchased (Gutwara 17). But some communes do not differentiate Neza [2010], as reported in Teyssier and others between the acte de notoriété and the tax; [2011]). Habwintahe and others (2013) counted practices vary across communes. In a study of 12,957 certificates in the registry in six SFCs, of five communes with a SFC, it was found that one which 86 percent concerned purchased plots no longer charged for an acte de notoriété, three (7 percent by women). The certificates will charged 3 percent, and the fifth continued to offer a less expensive approach to formalize charge 10 percent. In one commune, payment transactions. for an acte de notoriété to the commune was even made a condition for receiving a certificate 21. Land transactions are an important from the SFC (Habwintahe and others 2013: source of revenue for communes. Land sales 60–61). Even though the sale price tends to may be accompanied by affidavits (actes de be undervalued, the actes de notoriété are a Table 3.3 The Acquisition of Plots in Rural Areas, Ranked by Method and Type of Household Marshland “paysannats” Households near Acquisition Mode Rural Households (marais) Total/Average Households Protected Areas Households # % # % # % # % # % Inheritance 1 340 50,0 412 43,4 175 31,4 282 64,1 2 209 47,8 Purchase 1 161 43,4 364 38 3 260 46,7 122 27,7 1 907 41,2 Grant/gift 108 4,0 44 4,6 48 8,6 12 2,7 212 4,6 Attribution 39 1,5 110 11,6 55 9,9 21 7,4 225 4,9 Exchange 8 0,3 3 0,3 1 0,2 1 0,2 13 0,3 Abolishment of 6 0,2 2 0,2 0 0 0 0 8 0,2 « Ubugererwa » Judicial/CNTB order 5 0,2 0 0 8 1,4 1 0,2 14 0,3 Total 2 678 950 557 440 4 625 Source: PRODEMA Report 2014. Note: Ranking provided by participants in focus group. Percentages indicate how frequently risks were ranked as either first or second in order of importance. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 75 Box 3.4 Securing Land Rights in Burundi: Why Does It Matter? The policy objectives guiding investment in securing land rights in Burundi have a basis in global impact evaluations and other research (Deininger and others 2011; Binswanger and others 1995; Banerjee and Iyer 2005; Iyer 2010; Nugent and Robinson 2010; Naritomi and others 2009; Rajan and Ramcharan 2011): • The primary aim is to reduce and prevent conflict. The interest in formalizing rights grows as (i) competition over land increases, (ii) land conflicts become more entrenched and vicious, or (iii) fear of expropriation. Land security is a special need in a post-conflict setting, as refugees reclaim land with the help of (external) mediation and would benefit from formal evidence. • Where there is widespread insecurity of land rights, clarifying such rights through systematic adjudication and registration can be a cost-effective way to increase tenure security and improve the effectiveness of dispute resolution. In particular, the documentation of rights and transactions is considered instrumental to resolving conflict and enhancing tenure security, as the process of publicly recognizing and recording rights—as well as mapping boundaries—can be used to resolve disputes and prevent future conflicts, while records and maps should help speed up court cases and will also help ensure the acknowledgement of rights in expropriation and compensation procedures. • Another policy rationale behind rural land registration is to encourage investments in land by reducing sources of land tenure insecurity due to disputes over rights and boundaries. The potential advantages of registration include more medium- to long-term investments in land (improving soil quality, planting trees, mitigating erosion) and property. Formal documentation of land rights is less important when there is sufficient trust in local land institutions and customs’ capability to protect rights—and no fear of expropriation. • A growing body of literature points out the importance of facilitating rental markets (Baland and others 2007; Deininger and others 2008a). Clear, accepted, and accessible land records contribute to solving disputes and enhancing readiness to engage in land transactions, particularly the leasing out of land. It is a low-cost and flexible mechanisms to bring land to its most productive use can generate opportunities in the context of structural transformation of rural economies and increase welfare (de Janvry and others 2001). High transaction costs, caused by unclear rights or institutional inefficiencies, can reduce the incidence of such transactions, with far-reaching impacts in the long term (Libecap and Lueck 2011). Institutions that allow the leasing out of land at low costs and allay fears that those who transfer their rights will not be able to reclaim their land may help respond to this demand. It may also allow labor to move out of agriculture without foregoing the benefits of for example, the social safety net function of land ownership. 76 significant source of revenue for communes. site visit, commité collinaire de reconnaisance), According to our mini-survey of public finances recording land information, publicly displaying a in 16 communes (see chapter 2), the reported request, setting up contests, setting procedures revenues from the acte de notoriété varied for dealing with opposition to claims, and the from 2 to 80 million FBu for 2013. This was, on issuance of certificates are all described clearly. average, 4.8 percent of the total revenue (ranging In addition, the decree clarifies the procedures from 0.6 percent in Giteranyi to 24 percent in for updating records (mutations like transfer, Mutimbuzi [a commune near Bujumbura], while subdivision, or consolidation of plots). The one commune—Mbayi—reported 0 percent).92 decree also clarifies the responsibilities of the administrator, the SFCs, the line 3.4 Piloting Decentralized Land ministeries, and—at the local level—the mayor Administration: Results and (administrateur communal), the commitee de Challenges reconnaissance, as well as the chef collinaire. 22. The Land Policy Letter of 2010 23. Under the new land code (2011), (Decret No 100-72 du 26.04.2010), communes play a central role in the introduced a system for land tenure to administration of rural land and land facilitate and speed up first-time land transactions. Communes have acquired registration. By introducing an alternative the instruments for decentralized land approach to titling, the new process—available administration, and could use the information at the communal level—is more accessible and generated for planning purposes and taxation. less costly (given that surveying requirements Certificates are issued by the SFC, located in have changed). This process is embedded in a communal offices—an arrangement that brings local, public process to ensure transparency and service delivery closer to farmers. The SFC is to contribute to dispute resolution/prevention. established by the communal administrator and The Land Code of 2011 provided the legal council, and issues certificates for parcels (rights framework and formalized the land policy letter recognition), records land sale agreements and (2011 Loi no 1/13 du 9 aout 2011). Detailed other transactions, and manages the records procedures and accompanying manuals (on archived at the communal level. Currently, legal implications, staff recruitment, working revenues to cover SFC operational costs procedures, fee structures, and so on) were are generated through fees for registration, developed by the government in consultation transactions and the costs of certificates. SFCs’ with the implementers of pilot programs. mandate is limited to rural land that is privately Moreover, procedures for the establishment of owned but not yet titled (terre—privée—non an SFC are being codified in the Draft Decree titrée). SFCs have no mandate to issue certificates on Land Rights Certification (Projet de Décret in urban areas or register leases for public land portant Certification des Droits Fonciers) but are (such as marais) or government leaseholds not yet approved by government. Specific steps (such as paysannats) that have not yet been such as starting a registration request, verifying privatized.93 the request at the communal level (through a 24. The establishment of the SFC is 92  In the SFC pilot areas, administrators expressed concern that certificates would replace these acte de notoriété, which is not the case as a transaction is not involved. Administrators seem 93  The new land code introduces the possibility that farmers’ disinclined to make revenues from land transactions available annual leaseholds of public land in government schemes (the to cover the costs of SFC operations. Another obstacle is that the paysannats) may become private land. The nonprotected areas issuing of an acte de notoriété is not automatically accompanied of the marais (the bas-fonds) can also be privatized; however, the by an update to the registry, which must be changed. required decrees are not yet available. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 77 coordinated by a special government unit la Paix et les droits de l’Homme APDH/ICCO), and has been largely funded by donors. with some differences in approach and resource This innovation in service delivery is the base). An additional seven SFCs are planned for result of the work of several donor-supported 2014, while the remaining communes do not yet pilot programs working closely together with have such a service. One of the SFCs - set up with the Unité de coordination de program foncier Belgian Technical Cooperation (BTC) support - national, an interministerial entity operating no longer benefits from donor funding. under the tutelle of the MEEATU. The aim of 25. The first activity of a new SFC is the the pilot programs was to design and test land inventory and mapping of public lands and administration service-delivery systems for the titled lands, including lands already in use or registration of land rights that are trustworthy, set aside for residential areas (lotissements and transparent, accessible, effective and legal. The terres loties). The communal land map (plan first pilot programs started in 2007 mainly with foncier communal) is therefore an essential support of the European Union (EU) and the instrument for an SFC to prevent any overlap Swiss Agency for Development and Cooperation as it presents all public, private, and titled lands (SDC) in six communes. Currently, there are (also see para 19). It enables the SFC to check 28 communes (of which 3 are starting up) if the request is not in contradiction with other with an SFC, supported by 5 different donors claims and to properly map the certificated plot. (SDC, EU, International Fund for Agricultural This task has become much easier for a SFC Development (IFAD), Programme Post-Conflit de since 2013, as detailed georeferenced maps have Development Rural (PPCDR), Association pour become available for each locality (EU funding). Box 3.5 Principles for First-Time Registration Systems • Land registration systems must be set up carefully to prevent negative outcomes. The literature (Colin and Ayouz 2006; Fenske 2010; Benjaminsen and Sjaastad 2002; Peters 2004; Berry 2009; Lavigne- Delville 2000; Meinzen-Dick and Mwangi 2008) documents instances where efforts for rural land registration either failed to achieve their stated objective or had the opposite effects of reducing access to land and increasing tenure insecurity because of: (i) neglect of the distributional impacts, (ii) unsuccessful efforts at wholesale “replacement” of existing rights, and/or (iii) a focus on high-cost technology that undermined broad access and affordability and could not be sustained. • Systematic documentation of existing rights on an area-wide basis in a way that links with land use planning will not only be more cost-effective, transparent, and in line with criteria for proper land use than a sporadic approach but also allow better functioning of markets and reduces risks of encroaching on existing rights or interacting with individuals or groups who are not true right holders. • Speed, outreach, and equity impact of systematic first-time registration can often be enhanced by starting with communal land boundaries first and focusing on cheap image-based approaches, to be supplemented by more precise surveys on owners’ demand (and at their cost), which will allow for achieving scale quickly.94 Service delivery by land administration institutions can also be improved, for example, by replacing analog paper-based methods with modern technology, using common base maps, sharing data among government departments, publicizing service-delivery standards, and 94  Even in South Africa, a country that has more than 500 qualified surveyors, high-precision surveys of the estimated 20 million informal parcels will be very time consuming, implying that a more expeditious approach will be needed. 78 institutionalizing customer feedback to increase the quality and accessibility of land information. • Ideally, registration should include a record of secondary rights, such as user rights—but this is often missing. Those holding so-called secondary rights (user rights, leaseholds, sharecropping, and so on) are at a particular risk, as are users of common lands who do not reside in the community (for example, pastoralists), or those with a weak social position (Batwa). Another risk is that group rights (family, clan) managed by customary authorities on behalf of all members, are registered in the name of the chief or clan head—although they are custodians only. • A careful first-time registration of rights, combined with boundary demarcation, contributes to tenure security and conflict prevention. It requires a cautious and high-quality process in public and in the presence of witnesses, with mechanisms to contest or redress within a reasonable time period (and no record of contested land). Fraudulent records undermine tenure security, contribute to weak governance, and prompt conflict. Key steps toward enhancing the transparency of the registration process are (i) informing all potential claimants about the processes and criteria used to decide between competing rights, (ii) requiring claimants to come forward, and (iii) adjudicating rights at one point in time and in public. These steps are essential to preventing the loss of rights in the process and the creation of new land disputes. • The registry should be trustworthy. Distrusted registries have a limited role to play in solving disputes or for due diligence in transactions and for credit operations. Robust procedures should be in place to ensure the reliability of the system, and it should be possible and also easy to cancel fraudulent titles and mutations. Another risk is the alteration of records. • The utility of a registry for conflict prevention and resolution, and for economic development, depends on its completeness and whether it is up to date. Holders of certificates should subsequently report and register changes such as transactions (sale, mortgage, leases, and so on), subdivision, and plot consolidation. This is a major challenge for registries and depends on the adherence of right holders, the costs of transactions and changes, and the ability of the registry to record, among other obstacles. If registries are not maintained, land registration systems return to informality. There are examples of such deregularization due to a registry could not be maintained, even where land titling had generated significant investment benefits (Galiani and Schargrodsky 2010; Galiani 2011; Barnes and Griffith-Charles 2007). Land registration structures and registry operation are not without costs. First-time registration of rural farm land is labor and resource intensive (many plots, distances) and a carefully designed process to prevent loss of rights in the process, and the resolution of disputes, also requires human and financial resources. The following section discusses key challenges There is little effort in Burundi to promote joint such as constraints on women’s access to land, titling, even though this is legally possible. Some resource mobilization for the systematic land women do hold land singly. In the systematic recording, as well as communal land tenure registration operation in Ngozi, approximately services including local taxation, the sustainability 10 percent of the 71,168 registered plots were of the SFC and their implementation to ensure a registered in the names of women. In urban national coverage. areas 11.6 percent of the plots, on average, Key challenge 1: Gender imbalances in land were registered by women. Among the plots registration remain recorded as a result of “on demand” registration in Ngozi (see para 27 below), only 3 percent 26. What is not owned cannot be were recorded in the name of women and registered; land registration, as such, cannot concerned mostly purchased land; this figure improve the land tenure security of women was 7 percent, on average, in five other SFCs. or of other vulnerable groups without land. An SFC starting up in Makamba province (in Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 79 the communes of Mabanda and Vugizo, with the “on demand” approach stood at 23,019 in the support of the NGO ZOA and the Embassy of 19 SFCs; In the six SFCs supported by the SDC, the Netherlands) will explorethe possibility of 10,731 parcels were registered “on demand” registering women’s secondary rights (igiseke) and 96,955 through the systematic registration on certificates (ZOA 2013: 14). This step could covering about 20 percent of the total commune prevent the weakening of such rights upon land area. However, 22 percent already had a printed registration. certificate issued requiring an application and payment by the land owner. Key challenge 2: On-demand registration is not an effective option in Burundi; communes 28. Evidence suggests that on-demand cannot finance the start-up costs of systematic registration is less effective than systematic registration with their own resources registration in reducing conflict and 27. In Burundi all pilot programs initially promoting investment. Full coverage of all offered registration only on demand. On- parcels—via systematic registration—can demand or sporadic registration implies that prevent conflicts from emerging, facilitates a landholder puts in a request for registration, dispute resolution, and also reduces negative which is put on a list by the SFC to be addressed distribution effects. Finally, a complete land in the future (either in order of arrival or grouped information system is a “public good” that can by zone to reduce travel time) of SFC staff. be used for planning and tax collection (internal Sporadic, on-demand, first-time registration revenue generation). benefits those who are prepared (and can 29. A well-designed process for sys- afford) to pay for such services, but may leave tematic registration, with broad community out the more vulnerable. Systematic registration involvement, contributes to dispute implies that all plots in a given area are surveyed resolution. In Burundi the systematic and recorded. When subsequent updating of registration of rights followed by mediation records and reporting of transactions are on and adjudication—with the involvement of demand, however, this requires that landowners local authorities, both formal and informal— commit to report changes and to pay for the developed in the SDC pilot resulted in a decrease update of records. In 2011 SDCs’ changed in the number of land disputes, both open its approach from providing on-demand to and latent. It even resulted in the withdrawal systematic registration when results showed of court cases. During a first-time systematic that the number of plots registered was very registration, many disputes—over boundaries, low compared to the total number of plots, that rights, or transactions—that can be mediated most of such plots had been purchased, and that come into the open. The approach used includes it would take decades to reach full coverage. A the possibility for mediation by a hillside study calculated how long it would take to certify recognition committee. The registration process all parcels in the commune, taking into account creates a sense of urgency among the parties SFC productivity, and found that it would take involved to resolve their differences, since rights from 11 to 240 years (World Bank/FAO 2014; over the parcel will not be recorded when there is see table 3.4). An approach to systematic a conflict or when there is opposition.95 In Ngozi, registration was designed and tested in 2011 for example, 25,400 parcels were measured and (opération groupée de reconnaissance, OGR), and applied in all communes supported by the SDC in 2012 in the province of Ngozi) with dramatic 95 In around 2,600 cases opposition were noted in five SFCs, effects on number of plots registered. In March although records are not complete (see Habwintahe and others 2014 the total number of plots registered using 2013) 80 Table 3.4 Estimating the Timing of On-Demand Registration No. Of Capacity No. Of years No. Of years No. Of parcels (if 1 of demand If CF demand necessary to Communes necessary to Population household = instruction hits 40% process 40% of process all parcels 2 parcels)* per year parcels Giheta 93 560 37 424 288 14 970 52 130 Gitega 123 811 49 524 206 19 810 96 240 Bugenyuzi 72 621 29 048 719 11 619 16 40 Kigamba 43 000 17 200 698 6 880 10 25 Nyabitsinda 58 004 23 202 1 338 9 281 7 17 Gashikanwa 72 621 29 048 2 648 11 619 4 11 Ngozi 127 000 50 800 765 20 320 27 66 Ruhuroro 75 429 30 172 579 12 069 21 52 Source: World Bank/FAO 2014. Note: *Assumption: one household is composed of five members; the average number of plots per household is 2. recorded, of which 1,681 (6.6 percent) were central government agrees that systematic contested, but opposition to issuing a certificate registration is the preferred option. was in the end waived for 1,496 cases (89 Key challenge 3: The business model proposed in percent), implying that an agreement was found the pilots for SFC funding may weaken its function with help of the local committee (Habwintahe as a public good, and cannot be supported by and others 2013). Most entrenched conflicts the small number of printed certificates being were related to inheritance. In the first round requested of OGR for all SFC in 2012, 71 percent of the conflicts were resolved; the following year the 31. There is full participation in systematic number of land cases in the courts dropped from registration campaigns, but the request 70 percent to 45 percent, and the time that local for printed certificates is low. The SDC pilot officials spent on mediation was considerably reported active participation in the systematic reduced (Ferrari 2014). Such effects can only registration program (OGR), with nobody opting be achieved when systematic registration is out, but a printed certificate was collected for only available to all. 19 percent of the parcels. This percentage was higher in peri-urban and urban communes, such 30. Although systematic registration is as Ngozi (38 percent; see table 3.5). An overall the preferred option, most SFCs cannot offer low level of certificate collection has also been this service, due to a lack of resources. Due to observed in Rwanda, Uganda, and Madagascar. economies of scales, systematic registration is The hypothesis is that those who have their cheaper per plot (where 50 percent will be paid parcels mapped and names registered perceive for by the landholder for issuing of the certificate; their rights as secure—even without an actual see World Bank/FAO 2014) and achieves full certificate—since they know that the certificate coverage much faster, since it concerns many can be requested (and paid for) whenever the plots in one round. The systematic registration need arises (for example, during the rental or therefore requires up-front funding for the sale of land, the use of the land as collateral, operation and a well-organized campaign. The or in case of a dispute). The limited request for Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 81 printed certificates, however, undermines the business model Figure 3.2 SFC Revenues over Time (2011–13) of the SFC, although data for the SFC in the city of Ngozi and other communes in the province show 18000000 16000000 that revenues are increasing 14000000 over the years, sometimes 12000000 2011 revenue significantly so (figure 3.2). 10000000 2012 revenue 8000000 32. The government and 2013 revenue 6000000 donors have chosen to recover 4000000 costs with fees to ensure 2000000 the financial sustainability 0 of SFC registries, but this ba a a zi ro ga z ar approach may undermine its o en ro an m Ng ng re ho ur ar Ta Ki contribution to the public good. M am Ru Ny The cost of a certificate depends Source : Swiss Development Cooperation Project 2014. on the methodology used for recording land rights and the costs for the establishment of the 33. In principle, Efficient working required offices, equipment, and staff—as well as procedures, combined with a growing volume the productivity of the staff. For all existing SFCs, of transactions, could turn an SFC into an donors paid start-up costs (building, equipment, additional source of revenue for communes. transport, and training) and covered first- An SFC can generate revenues when the costs for time registry operations. An SFC is expected to issuing certificates and recording transactions break even on the basis of fees paid for services. and mutations is lower than the fees that are Its running costs include salaries, materials, charged, and when more transactions are transport costs, and so on. The rationale of the registered—assuming that this can be taxed, current approach is that a low-cost registration as the request for the more expensive acte de coupled with high numbers of beneficiaries will notoriété is likely to drop when the change can be keep fees low while generating sufficient total noted on the certificate instead. But as of now all revenues to operate the SFC. The revenue sources such investments are covered by donors, with the for an SFC are fees for the various services set by the commune. For example, for the (EU-funded) intention that costs for maintaining the SFC will SFC working with Gutwara Neza, fees are voted be covered through fees. An analysis of the costs on by the communal as set up as it follows: 1000 per certificate under an “on-demand” approach Fbu at the time of certificate withdrawal; 2000 found much variation across SFCs (World Bank/ Fbu for plots between 0 and 30 acres; from 3500 FAO 2014). The differences are caused by the to 5000 Fbu for plots between 31 and 60 acres; procedures used, the type of equipment, and from 7000 Fbu to 10000 Fbu for plots between the type of transport (car, motor). This study 61 and 100; and 10,000 FBu for each additional also found large differences in the productivity acre. The registration of a mortgage or a lease of SFC staff members. Such differences in staff is 2,000 FBu per parcel. Finally, a copy of the productivity and even in the quality of service certificate costs 2,000 FBu. Habwintahe and delivery were also found by another study of others (2013) found significant fees variation seven SFCs. The issuance of certificates was exists across communes, as well as a tendency slower in some SFCs than others. Staff turnover to increase fees over the years to ensure SFC was found to be high (66 percent); staff members cost recovery. interviewed considered their responsibilities— 82 Table 3.5 Plots Registered and Certificates Issued by the SFC in Ngozi Province (SDC Pilot) Sous- Plots Average Plots for Which Certificates Commune Colline Collines Households Measures Plots/ Certificate was Paid (% of SDC (no.) (no.) (no.) Household Paid (no.) Plots) Ngozi 8 27 3 761 17 949 4,8 6 745 38 % Marangara 7 29 s.o. 11 574 2 197 19 % Ruhuroro 8 23 3 792 12 069 3,2 1 409 12 % Tangara 6 26 3 186 11 231 3,5 1 612 14 % Nyamurenza 3 13 2 805 9 956 3,5 598 6% Kiremba 6 15 2 418 8 389 3,5 743 9% Total 37 133 71 168 13 304 19 % Source: SDC pilot administrative data, June 2013. Table 3.6 Costs of First-Time Registration (On-Demand) for 8 SFCs Procedures + Cost of the Procedures + Procedure + operations + Applied CF procedure operations + Demand Communes operation work tools + Price (forms used) work tools /month motorcycle BIF $ BIF $ BIF $ BIF $ BIF $ Giheta—UE 6,737 4.3 23,353 15.1 27,115 17.5 30,877 19.9 24 3,000 1.9 Gitega—UE 6,737 4.3 29,923 19.3 35,172 22.7 40,420 26.1 17 3,000 1.9 Bugenyuzi—UE 2,280 1.5 16,227 10.5 16,593 10.7 16,960 10.9 60 3,000 1.9 Kigamba— 6,401 4.1 11,562 7.5 13,043 8.4 14,525 9.4 58 5,000 3.2 PDDRC Nyabitsinda— 6,448 4.2 8,541 5.5 9,659 6.2 10,777 7.0 112 7,500 4.8 ancienne CTB Gashikanwa— 1,895 1.2 5,765 3.7 5,864 3.8 5,964 3.8 221 6,000 3.9 SDC Ngozi—SDC 3,463 2.2 4,714 3.0 6,705 4.3 8,696 5.6 64 4,000 2.6 Ruhuroro-—SDC 3,463 2.2 6,283 4.1 7,205 4.6 8,127 5.2 48 6,000 3.9 Source : World Bank/FAO 2014. Note: BTC = Belgian Technical Cooperation; EU = European Union; PPCDR = Programme Post-conflit de Développement Rural; SDC = Swiss Development Cooperation. which involve a lot of field work—as relatively Key challenge 4: Limited results have been heavy compared to other communal functions. achieved by communes using SFCs to generate Other challenges to staff productivity relate to local revenues cash flows at the commune level, which affect 34. Communes could, in theory, increase the resources available for transport and the tax revenues using the information generated presence of local committees for on-demand by the surveying of parcels, recording of approaches (Habwintahe and others 2013). landholders, and from expansion of land markets. In other countries, information Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 83 on land and property collected as a result of fees. The 28 communal land services currently first-time registration is also the basis for operating are primarily funded by donors. taxation. In the Philippines, for example, local governments themselves are funding the costs 36. Political support for SFCs varies of tenure regularization (also using a type across national stakeholders. At the national of certificate), after which the information is government level there is some interest in land used to prepare a tax map (and are recovered reform—as previously noted, the land code was within a few years by increased tax collection). expedited and approved in 2011—but support for In Madagascar a pilot conducted in five local a more comprehensive land reform is probably government areas also showed that improved lower. Key political issues seem centered on the information used to collect taxes was a key role of the CNTB in the institutional landscape, motivator for local governments to increase the villagization, and the allocation of fertile public coverage of certificates, although some people, land to private investors, including marais and in their quest to evade taxation, also missed out paysannats. There is some interest in land tenure on enhancing tenure security. But transaction security: promises of land titles were used in taxes also undermine landholders’ willingness the 2010 political campaigns. But communes to register any changes formally, undermining are more interested in land reform than the the update process of the registry. Moreover, central government, though not all SFCs are fully Burundi has no land taxation system, and tax integrated in communes’ plans and budgets. on land transactions was abolished in 2008. It is reported that, locally, the population Property taxation in urban areas is possible, perceives some SFCs as donor projects and although the fee is low (see chapter 2) and may not a government service (Habwintahe and not compensate for the costs of tax collection. others 2013). Civil society organizations (CSOs) are supportive of the reform agenda and Key challenge 5: Sustainability requires leader- supported decentralized land administration ship in the land sector and more devolution of and SFCs from the start. Their main concern is resources to communes sustainability, quality of service delivery, and 35. Sustainability issues need to be ensuring transparency. addressed first before scaling up to make 37. Coordinating two systems is needed the SFC service available in all 116 rural for managing land information. The title and communes in Burundi. Scaling up requires land certificates systems for property rights that all citizens in rural areas have access to an registration are complementary, not competing. SFC and that SFCs have the capacity to service all They should continue to coexist for at least a inhabitants within a reasonable period. Although decade. Although the conversion of certificates the costs per parcel for systematic registration into land titles is possible, it is probably not are relatively low and can be expected to go a priority for most homeowners in Burundi. down as experience is gained. The total costs of Land information management and sharing establishing an SFC and complete registries by mechanisms will avoid any overlapping of rights a systematic registration cannot be carried out (Report PRODEMA 2014). by communes with their own resources and are likely to require noncommunal funding. (This 38. Line ministries are interested in land issue has been addressed in chapter 2.) The key reform, although engagement is not constant. issue for sustainability is to move away from a Multiple line ministries play an important near-total reliance on donor funding and toward role in the establishment of a sustainable an increased central government transfer to land administration system (institutionally, communes as well as local resource mobilization technically, and financially) and maintaining through taxation—in addition to landholders’ quality standards, land information, training 84 for land administrators as well as the correct insufficiently coordinated. For example, the cost application of procedure and integrity of the per certificate can probably be reduced even system to ensure efficient and appropriate more by streamlining procedures and enhancing service delivery. Initially not all staff in the efficiency, but this requires regular monitoring line ministries and departments in charge of and testing and comparing of alternatives. land administration (titres fonciers, cadastre, aménagement) welcomed the development 40. Impact evaluation (IE) is required to of the SFCs and certificates. Reasons included assess the outcomes of first-time registration. professional differences over the minimum IE is needed to assess outcomes and causal precision standards for surveying, and doubts relations, such as around conflict prevention, about the capacity of communes. Another factor changes in costs due to land-related disputes and was competition between decentralized agencies conflicts, changes in investments, sustainable and, perhaps, efforts to protect a “monopoly” land use and land markets, and their impacts on around land titling. In Burundi much care has people’s well-being (food security, livelihoods, been taken from the start to associate the line economic development, empowerment of ministries and also attend to their constraints vulnerable groups, and stability). IE research may around titling (archiving, digitalization, and so be instrumental in justifying the mobilization of on) with project resources for the modernization additional donor funding. IE using international of the registry and cadastre included in the standards is lacking, and most research efforts support programs. Representatives from these on outcomes seem limited to focus groups and ministries and departments are also members rapid appraisals. But the systematic registration of the steering committee for the pilot programs. work in Ngozi (2012 round) has sufficient Given concerns over the financial sustainability scale and can be evaluated, with a baseline of the current approach (full-cost recovery by reconstructed using the sampling approach the SFC alone), a draft interministerial decree on developed by the World Bank for Rwanda; and land reform (currently under discussion within would take about 6 months to complete (Ali, government) proposes96 to fund SFC staff from Deininger, and Goldstein 2011). In addition, central government resources. household surveys need to integrate a module with question on land tenure, which has not yet Key challenge 6: Scaling up requires data on occurred. performance, outcomes, and impacts 39. Plans for scaling up SFC in all 3.5 The Way Forward: Actions communes require the availability of reliable Needed to Scale Up Land and regular statistics and administrative Governance Reforms in Burundi data for monitoring progress and impact. Currently, data on monitoring is weak or 41. Since 2005 Burundi has made unavailable with respect to outputs, costs, remarkable progress in making rights quality, and so on, and land data are scattered, registration a prospect for rural landowners, not digitalized and not aggregated at the central by developing a new product (certificat), level. Supervision seems to be irregular while delivered by local governments, and which progress assessment and learning seems to are currently part of the legal and institutional depend on donor initiatives for self-assessment framework for land governance in Burundi. and external evaluations, which also seem The pilot programs designed a land governance system that is trusted by the local population and can be implemented by communes. The 96  The second vice president is knowledgeable on the matter, SFC staff has shown that issuing certificates is since there is an SFC in his commune. technically possible and that, when a systematic Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 85 registration approach is used, cases of conflicts the registry. are reduced (those that can be resolved more easily). These achievements are the result 43. Decentralized land administration is of considerable investments and sustained a key element in improving land governance, support of key development partners working but other issues remain, to be addressed in close collaboration with the PNF as well as by a comprehensive land reform program. CSOs. The result is a land administration service Issues such as the accessibility, effectiveness, that is both accessible and cost-efficient for the and transparency of the judiciary system; the transparent and efficient management of public majority of farmers. lands and procedure for privatization of these 42. Land administration services lands; rural land use and urban land use planning provided by the SFC should be permanent; (for example, to prevent soil degradation and sustainability is therefore essential. Land flooding); women’s land rights issues; and certificates document rights, and are thus a key reducing urban informality which requires aspect of an institutional system and rule of law. are also important land governance issues. Currently all SFCs are adopted by donors, which The implementation of the Land Governance bring approaches of their own and also decide Assessment Framework (LGAF) is therefore when funding stops. The sustainability of SFCs is urgent for a more comprehensive assessment discussed regularly in Burundi, with a focus on and will provide both a benchmark and basis for finances and the aim to achieve self-sufficiency a monitoring system and performance indicators through full cost-recovery at the SFC level. But for tracking progress on land governance an SFC is providing public goods (in addition to reform. The LGAF can help to establish support private goods), such as contributing to conflict and prioritization for the institutional and prevention and speeding dispute resolution, financial audit for the land reform and a providing a basis for facilitating land markets participatory process to plan land reform policy (including rentals), collecting land information as recommended by the PRODEMA study. that can be used for planning and taxation, and 3.5.1 Short-Term Priorities for Reform so on. Moreover, the service should be available (1–3 Years) to all, and not only those with the capacity to pay. Land registration will have the highest 44. Adopt the pending decrees on value as a public good when local government certification and decentralized land areas are fully covered and all transactions are administration. The proposed draft decree regularly reported to the SFC, ensuring that on the certification of Land Rights (Projet de public information on landholdings remains Décret portant Certification des Droits Fonciers) current. The full registration of transactions and lays out the steps for the certificate application subdivisions is particularly important for the process, for the OGR (systematic registration), quality of the registry. How transactions are taxed and for the conversion of a certificate into a influences landholders’ willingness to formally title. Through these decrees, the results of the report transactions.Therefore, systematic approach emerging from the pilot programs registration of land is superior to on-demand will be codified. This decree’s adoption has registration—but the total up-front costs are been postponed by the council of ministers, due relatively high, because so many more parcels to concerns over financial sustainability.. The are included, although the amount per parcel proposal is that these costs will be included in the is much lower. Central government transfers proposed draft decree and paid through central to cover basic costs (such as the salaries of SFC government transfers, while the communes staff) are required. Taxation is another aspect— remain responsible for the SFC, which is in line but high fees or transaction taxes, resulting in with the conclusions of this ESW. The use of nonregistration of transactions, can undermine FONIC as a fund for SFC start-up costs should 86 be considered to make it possible for local as a basis. As this report indicates, costs can governments to request the establishment of an be brought down and capacity be increased SFC even when there is no donor directly active by better streamlining procedures, improving in the area. efficiency, and making more effective use of human resources, according to the work load. 45. Develop a national land reform Cutting down costs, however, should not be at program. This land reform program is the expense of building local understanding and expected to guide the implementation of the engagement, transparency, and the system’s 2010 land policy letter as well as the land capacity for addressing pending conflicts. code (2011). In particular, it should lead to Moreover, an impact evaluation on land tenure the institutionalization of decentralized land is desirable in the short term and can be management at both the central and local levels, undertaken relatively quickly with technical and enhance sustainability by integrating the assistance from the World Bank. SFC in the operating structure of communes, and linking it with local taxation. Moreover, the 48. Reinforce monitoring and evaluation program should include an assessment of the (M&E) of performance and impact that can resources, operational procedures and timeline evolve in a multistakeholder land observatory. needed to achieve national coverage of SFC and Monitoring is also important as an early warning massive recording of land rights. mechanism for problems that may arise, given the sensitivity of land governance. Plans for 46. Development of standard operating scaling up require the availability also of solid procedures and guidelines for the supervision monitoring data on process, output, costs, and of SFC. Burundi has now established a model quality. Regular monitoring of impact is also for SFC based on the donor supported pilot important as a mechanism for early warning of initiatives, which is adopted by the MEEATU. To emerging problems, given the sensitivity of land ensure quality of SFC, performance standards governance. and operating guidelines have to be developed, 49. Use FONIC grant and donor resources which is a shared responsibility of the Unite de to address set-up costs of communal land coordination of the Programme National Foncier, registration services, while the recurrent local authorities, CSOs and other institutions costs can be addressed by the proposed currently supporting the SFC initiative. The PNF recurrent grant (see chapter 2). The 28 could coordinate this process and also support communal land services currently operating the start-up of new SFC and design capacity are primarily funded by donors. Unless fiscal- building mechanisms. Ministries in charge of administrative aspects of decentralized tutelle also play a key role in supporting and governance are clarified, the SFCs will continue providing guidance to SFC, as well as maintaining to depend on development partners to finance land information databases. their operations. A recent study shows that the 47. Test various approaches to increase set-up costs of an SFC range between FBu 20.6 productivity and reduce costs of SFC million and 59.3 million; a reasonably equipped installation and operation. National coverage office averages FBu 40.9 million.97 Such set-up requires that the process is as efficient as costs could potentially be covered by the annual possible without losing the quality of the process FONIC grant; to this end, it would be targeted in and trust in the system. Financial assessments on costs of starting up an SFC and the process of first-time registration and maintenance of 97  Analyse financière de la gestion foncière décentralisée Expertise complémentaire à la Revue du Secteur Foncier au the registry need to be refined and discussed. Burundi Rapport Version du 11 février 2014 Zo Ravelomanantsoa, The 2014 FAO/World Bank report can be used table 5, p 6–7. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 87 much the same way as past grants were targeted reported that certificates are handed over during to primary schools—with national priorities a sale to the buyer, but without awareness that overriding communal investment priorities for this transaction also needs to be reported at the one year. The recurrent costs of an SFC range SFC level (Habwintahe and others 2013). from FBu 2.8 million to FBu 10.3 million in a demand-driven environment.98 This implies a 3.5.2 Medium-Term Priorities for Reform per-certificate cost of between FBu 6,000 and (3–5 Years) FBu 17,000.99 The study estimates the costs of 52. Strengthen the institutional frame- a certificate in a supply-driven environment at work, including communal engagement, FBu 17,000.100 Thus the recurrent costs of an relations with decentralized services, and SFC are most likely around FBu 15 million a year, national-level support for the land rights gross of any revenue generated. The proposed registration process. Setting performance recurrent grant covers this additional cost. standards and supervising their achievement, 50. Develop an effective taxation strategy (to avoid erroneous certificates) are impor- for rural land, in addition to urban land.101 tant for maintaining a reliable system that is The abolishment of the taxation of transactions trusted. needs to be lifted, and a local government tax 53. Establish permanent structures needs to be set. The revenues from transfer for capacity building and training for taxes102 would compensate the income from decentralized land administration. Improving acte de notoriete for local governments, but they local tenure security on a massive scale requires should not be so high as to discourage the formal support from line ministries and the full registration of changes, thus undermining the capacity of communes, such as SFC agents and registry. The option that local governments the local committees that are fully engaged in rural areas of Burundi will invest in land in first-time registration. A dedicated cadre registration to enhance tax collection is unlikely of career civil servants in the decentralized at the moment, as rural land is exempted from human resources/staff structure is needed tax collection, while they already capture taxes to provide Burundi with the skills required on land transactions. for maintaining land administration services. 51. Promote awareness campaigns to Decentralized land administration staff should ensure the registration of all land transactions. also be trained in land administration along The value of a land registry depends on whether with general commune staff and administrators, it remains current and up to date. Keeping the and so on. The Land Coordination Unit (Unité registry current may also require information de Coordination Foncier) is developing training and awareness campaigns. For example, it is modules for SFC staff for the National Training Center for Local Actors of Decentralization (Centre National de Formation des Acteurs Locaux de la decentralization), but funding for 98  Ibid, table 6. implementation has not yet been secured. There 99  Ibid, table 10. are plans, too, for a university program focused 100  Ibid, table 11. on land tenure. 101  Another source of revenue is the rental of public lands 54. Establish an integrated land by communes. Some communes charge land rent for public land (redevances marais). information system. The digitalization of records for certificates is not considered a 102  A possibility would be to introduce a tax of FBu 2,000 per plot in 2016, going up FBu 100 per year over the following 20 priority in the short term. The focus of the years. national land information system, which would 88 bring together all land records and spatial monitoring unit, disaggregated by gender. Other information, will be first on spatial information departments and the courts will also report to (and inventory public land, locating titled this land information system on property titles, land and land with certificates, and so on). In transfers, mortgages, land-related disputes, the medium term the information produced and taxation. These data can be used by a land by SFCs will need to be made available in a observatory to track progress, to inform policy digital form and reported regularly to a central dialogue, and to improve accountability. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 89 Chapter 4 Promoting Social Accountability through Decentralization: The Voice of Citizens Important note for the reader: The team wishes to note that the findings illustrated in this chapter are based on various focus groups held in 16 communes across 8 provinces in Burundi. The fieldwork activities were undertakend by Burundian consultants, in close coordination with the Bank team. The assessments of state-citizens accountability relationships directly reflect the point of view of more than 300 local residents and elected communal officials who participated in the qualitative surveys. They are intended for illustrative purposes only, and do not claim to sytematically capture national patterns. Summary of Main Arguments the dissemination of the communal budget, and social service delivery at the local level. and Key Findings 2. To collect the opinions of Burundian 1. The objective of this chapter is to investigate the context of social citizens, a series of 16 focus group discussions accountability, including openings for were conducted in eight communes across social action and the nature of state-society the country. In each of these communes relations in Burundi today. Embracing the idea discussions were held with two separate that “context matters,” this chapter discusses groups, one of citizens with no connection to the opinions of Burundian citizens on key issues public office, and the other of local officials. such as the use of tax revenues, tax collection, the The citizen group included members of local formulation of communal development plans, development committees, local associations, 90 and representatives of vulnerable groups, the recent support of the Bank (through the including youth and women. The group of local Community and Social Development Project, officials included communal administrators and PRADECS) and the donor community at large, technical staff (advisors to the administrator), communal administrations have become members of the communal council, elected hill increasingly able to manage direct responsibility council members (conseil de collines), and local over funds for small development projects. This officials of deconcentrated services such as included the formulation of the first communal education, health, water, and justice. development plans; the identification of priority subprojects through participatory community 3. The macro framework employed to meetings; as well as support in several key areas, guide this analysis and present key findings such as managing the procurement process, is the “Conceptual Framework of Constituent monitoring subprojects, and guaranteeing Elements of Social Accountability.” This the maintenance of newly built communal framework was developed as part of the infrastructure. By facilitating open dialogue 2014 World Bank Flagship Report on Social and the inclusion of vulnerable groups, these Accountability (World Bank 2014b). It presents interventions made a significant contribution five constituent elements for a contextual analysis toward social cohesion and conflict mitigation. of social accountability: (i) citizen action, (ii) state action, (iii) information, (iv) mobilization, 6. Despite this recent progress, social and (v) the citizen-state engagement interface. accountability in Burundi remains weak, with limited opportunities for citizens and 4. The main finding that emerges communities to engage in the decision- from research on the social contract at making process and hold local authorities the local level in Burundi is that the state accountable. Citizen engagement in local has more opportunities to promote social decision making and participation in local accountability than do citizens. In other words, development remains limited, undermining the supply side rather than the demand side of the potential benefits of decentralization social accountability provides more potential reform as a vehicle to improve accountability entry points at this time. In employing the and strengthen local governance. Given that conceptual framework of constituent elements, it in Burundi the majority of citizens access became apparent during focus group discussions their basic services at the communal level, it is that there are critical obstacles to citizen essential to better understand what works well action: namely, a lack of efficacy among citizens and what does not, and how social accountability (meaning the self-perception that citizens can interventions could be promoted to improve the bring about change), a lack of collective action, quality of service delivery and strengthen local and limited intrinsic motivation (as in a limited governance. sense of moral responsibility to address a given problem). The state, too, has limitations—the 7. A key message that emerged most obvious being a substantial rise in political from the focus group research was that tension in the run-up to national elections in social accountability may be immediately June 2015. strengthened by making tax collection frameworks and communal budgets more 5. Recent political reforms, meanwhile, transparent and systematic. Based on have created some opportunities for discussions held during the focus groups, local communes to play an important role in officials are open and willing to communicate building social cohesion at the local level, more effectively with citizens to increase in particular the direct election of communal communal tax revenues. On the flip side, citizens administrators and local councils. Due to are eager for greater transparency around local Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 91 resources and to better understand the use of the nonprofit and private sectors. Through tax revenues by local officials. various dimensions—such as transparency, participation, and accountability—social 8. Although the nature of the state- accountability empowers citizens by giving society relationship in Burundi has them a greater voice in governance and policy- significantly improved since the government making processes, enhances transparency, embarked on an ambitious decentralization and strengthens the environment for civic process in 2005, political realities and the engagement. Such processes, in turn, inform and legacy of conflict persist. After decades of strengthen social accountability interventions. conflict, the severity of violence observed across Burundi has been markedly reduced. 10. According to the literature, the success But sources of the conflict (such as exclusion— or failure of social accountability interventions is social, economic, and political) remain intact. highly dependent on context. Experiences from Even as social cohesion is being rebuilt across previous social accountability interventions local communities, the domination of a neo- have shown successes as well as failures where patrimonial system of governance is blocking results were not only negative for the poor but state reform, constraining improvements in also to overall development goals. As a result, social service delivery at the local level. This evidence found that: “context shapes the form chapter explores the local context in Burundi and and effectiveness of social accountability, but how the endogenous processes of state reform often in unpredictable and complex ways” are affecting social accountability, providing (World Bank 2013). Moreover, evidence found recommendations on a way forward. that some contexts are more enabling of social accountability than others and the context will 4.1 Social Accountability influence (although not necessarily determine) and Local Governance: Their the form social accountability will take and Connection and Importance the probability of achieving its objectives. Comprehensive analyses of operational 9. How social accountability is defined contexts are therefore essential before social varies around the world, as do interventions accountability interventions. meant to improve it. Although there is no universal consensus on the definition of social 11. Although it is clear that context is accountability or the range of actions that fall critical in shaping, making, and breaking within its realm, several key characteristics may social accountability interventions—and be discussed. Social accountability typically their outcomes—it is still difficult to involves a process of engagement, dialogue, and determine how. Knowledge of how context negotiation between citizens and the state, with shapes the outcomes of interventions remains the goal of influencing the broader development limited. A World Bank report, Mapping Context agenda. It is also a process by which citizens in Social Accountability, provides an analytical hold the state to account for commitments made framework across six contextual domains—civil to them as part of a governance compact.103 society, political society, interelite relations, Moreover, it applies to contractual situations state-society relations, intrasociety relations, where demand interacts with supply, creating and global dimensions. These domains feed accountability relationships, including in into the Conceptual Framework of Constituent Elements of Social Accountability developed by the World Bank Social Accountability team for 103  World Bank, “Flagship Study on Social Accountability,” Draft the Flagship Report on Social Accountability Concept Note 2013. (World Bank 2014b). 92 12. Existing literature on the relationship can also help these societies avoid and withstand between social accountability and context conflict and rebuild in a postconflict scenario.104 points specifically to the importance of Overcoming obstacles to citizen engagement and citizen engagement. Findings from the the possible limitations of state responsiveness literature suggest that the willingness of citizens are areas to be further explored, especially in to pursue social accountability goals and to fragile contexts where institutional capacity challenge the state is a key factor, often shaped and state legitimacy are weak.105 Where roles by citizens’ historical experiences of state- and responsibilities are not clearly delineated citizen relations. Although evidence highlights across institutions and influence of authorities the importance of state capacity in social is unclear, bolstering social accountability is an accountability outcomes, there is no evidence especially daunting task. that social accountability cannot be achieved even in low-capacity environments (Gaventa 14. Given Burundi’s status as an FCS, and Barrett 2010). On the one hand, evidence the critical assumption that low levels of suggests that the level of state democratization accountability endanger peace is a salient does influence the emergence and effectiveness aspect to explore. If social, economic, and of social accountability. McGee and Gaventa political exclusion are sources of conflict, (2011: 12, as quoted in World Bank SDV 2013) then social cohesion is a worthy goal. To this state that: “in a regime lacking the essential end, how can social cohesion be furthered by freedoms of association, voice or media, citizen- accountability and, in turn, positively impact the led TAIs (transparency and accountability nature of the relationship between the citizen initiatives) do not have the same prospects for and the state? success as in societies where these conditions exist.” On the other hand, evidence suggests 4.2 Analytical Framework, Data, that state democratization alone does not and Methodology decide state-citizen relations. Specifically, Crook 15. This report aims to fill the evidence and Booth (2011) found that: “the presence of formal democratic institutions and frameworks and knowledge gap on social accountability. reveals only part of the picture in many contexts The concept of social accountability is currently as it is informal institutions and the underlying under review by both policy makers and political settlement that explains what happens academics. The possibility that greater social and why.” In addition, Stasavage (2005) found accountability fosters more state-society that “different forms of social contract or interactions, promoting stronger and more open developmental accountability can emerge states, is an exciting promise that has yet to be within weakly democratic or semi-authoritarian bolstered by concrete data. To further research regimes as in the case of primary education in in this important area, the World Bank Social Uganda.” In sum, and as previously noted, how Accountability Team has developed a conceptual context influences social accountability remains framework to assist practitioners working in unclear. the field. This framework was designed to be used as a tool to assess which elements of social 13. Citizen engagement in fragile and accountability have the most potential influence conflict states (FCSs) is in particular need of support. Given the weak sociopolitical environment of FCSs, social accountability 104  For example, see World Bank (2012), which suggests that interventions to support stronger state- fractured social relationships and citizens’ relationship with the citizen relationships need to be explored. state underpin situations of fragility and conflict. This will not only result in improved service 105  See World Bank (2012) for an interesting literature review delivery in areas where it is most needed, but on social accountability in fragile states. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 93 within a given context. transparency, accuracy, and so on) to be used as a tool by both the state and citizens. 16. Understanding social accountability through a conceptual framework promises 17. Using the framework—and under- to further key research. The framework— standing the importance of the constituent Conceptual Framework of Constituent Elements elements of social accountability—is relevant of Social Accountability—illustrated in figure to Burundi’s particular context. The World 4.1, outlines five key elements of social Bank Social Accountability Team developed accountability: citizen action, state action, this macro framework to support practitioners information, mobilization and the citizen-state in analyzing contextual determinants of social engagement interface. The critical element accountability effectiveness, and to assess is how citizen action and state action can be enabling and disabling contextual factors. The fostered to influence the interface element, and key elements may be seen as entry points for the overarching role of information (access, social accountability interventions. Further, Figure 4.1 The Key Elements of Social Accountability: A Conceptual Framework STATE ACTION INFORMATION CITIZEN-STATE INTERFACE CIVIC MOBILIZATION CITIZEN ACTION Source : World Bank SDV 2014. 94 each constituent element is accompanied by a diagnostic checklist to guide contextual analysis Table 4.1 List of Communes Where Focus Group Discussions Were Convened in the field. The framework was specifically employed in Burundi to research the roles of Commune Province state action and citizen action through focus Mutimbuzi Bujumbura group discussions. Information was assessed as a cross-cutting element affecting the state and Rutegama Muramvya citizens alike, and pointing to the need to bolster Makebuko Gitega the citizen-state engagement interface. Gisozi Mwaro 18. Research in Burundi may help clarify Mabayi Cibitoke the link between state fragility and social Gihanga Bubanza accountability (or, more precisely, its lack). In Busoni Kirundo Burundi research was motivated by the question, “What is the nature of the social contract at the Butaganzwa Ruyigi local level in Burundi?” This question specifically Source: Authors’ fieldwork data collection 2013. refers to the mutual expectations of entitlements, roles, and responsibilities between the state and citizens in Burundi. It is hoped that answers country, (ii) to align with the same communes gained through field work will help clarify how visited by colleagues who completed chapter 2 state-society relations and societal dynamics on fiscal decentralization (and thus reinforce influence social accountability. The 2011 points of comparison), (iii) to assess the urban- World Development Report highlighted the link rural split among communes, and (iv) to identify between fragility and a lack of accountability, differences between communes with lesser and stating that, “periods when states or institutions greater resources. The selected communes are lack capacity, accountability, or legitimacy to listed in table 4.1. mediate relations between citizen groups and 21. To better understand the local between citizens and the state, making them context, information was gathered from both vulnerable to violence.” The argument that a lack local citizens and communal authorities. In of accountability is a fundamental characteristic each of the eight communes two different focus of fragility necessitates specific attention on group discussions were held—one with a group how accountability can be improved to foster of individuals comprised of commune officials, state-society relations and promote stability. and one with a separate group comprised of 19. The research completed for this citizens. Two different focus group guides were chapter includes three aspects of original prepared. Each focus group lasted approximately data collection and qualitative analysis. The two hours, and all of the questions from primary research includes the completion of the guide were covered in each of the eight 16 focus group discussions in 8 communes, in- communes, using a semi-structured discussion depth data collection from 4 of these communes, approach. One international researcher and and further qualitative analysis for 2 of these one Burundian researcher completed each communes. This was conducted using the macro of the 16 focus group discussions. Each focus conceptual framework explained above. group discussion collected the names of those interviewed, and no less than 250 people took 20. A series of 16 focus group discussions part in these discussions across the 16 groups. To were conducted in 8 communes across the confirm and further understand the information country. The eight communes were selected (i) received through the focus group discussions, to represent a geographical spread across the further data were collected on four of the eight Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 95 at the local level in Burundi is that the state Table 4.2 List of Communes Where Additional has more opportunities to bolster social Data Was Collected Beyond the Focus Groups accountability than do citizens. In other words, the supply side, rather than the demand side, of Commune ProvinceS social accountability provides more potential Mutimbuzi Bujumbura entry points at this time. In employing the Rutegama Muramvya Conceptual Framework of Constituent Elements Makebuko Gitega of Social Accountability, it became apparent during focus group discussions with citizens that Gisozi Mwaro critical obstacles within the citizen action cluster exist, namely, lack of efficacy among citizens communes (table 4.2).106 (meaning the perception among citizens that they can bring about change), a lack of collective 22. Data collection paralleled the topics action, and limited intrinsic motivation (as in a in the focus groups guide and focused limited sense of moral responsibility to address a principally on availability of social services, given problem). Within the state action element budget data and local taxes. Communal there are of course clusters that are limited, budgets (including donor support for specific but these are more operational. The limitations investments), tax revenues, and availability within the citizen action constituent element are of social services and infrastructure by major more profound in Burundi and as such activities sectors (health, education, water) were employing the use of a social accountability tool reviewed, using communal records alongside in a given community will gain limited impact communal officials. on creating change. The conclusion provided by Fox (2014) on the importance of maintaining a 23. Two communes among the eight, “sandwich strategy,” meaning the coordination Rutegama and Makebuko, were further of both the supply and demand sides, is a analyzed to be used as examples for the more useful approach to promoting a theory of process chain on communal investments. change. Specifically, to review citizen participation within the context of social accountability, four social 25. Four areas of interaction between infrastructure projects financed by the World citizens and the state were continually Bank’s CDD project, PRADECS, were reviewed in emphasized during focus group discussions. terms of citizen participation in decision making These were: (i) the role of participation in and project monitoring. The data were collected society, (ii) the preparation of communal through qualitative, semi-structured interviews development plans (CDPs), (iii) budget with citizens who had participated in the social preparation and fiscal transfers through Fonds infrastructure projects. National d’Investissement Communal (FONIC, National Investment Funds for Communes), 4.3 Empirical Analysis: Key and (iv) tax collection. Taken together, patterns emerge that lead to the conclusion that Findings communes in Burundi are not only a space for 24. The main finding that emerges from development, but rather a domain for control by research into the nature of the social contract central state authorities. This control extends not only over territory, but also the population. This conclusion corroborates the findings from 106  Full tables with all data collected for the four communes are the political economy analysis emerging from available from the authors upon requests. this study. How, then, can social accountability 96 be better understood in Burundi to augment, most often linked to extreme poverty. Several in the first instance civic engagement and respondents spoke of their own distancing participation, and subsequently influence the from the local administration, with one woman state-society relationship to effect meaningful summarizing the sentiments of all the focus change in local communities? The next four groups: “These questions you are asking—on sections articulate the views expressed during the development of the budget, tax collection, the focus group discussions in eight communes prioritizing the communal development plan— across Burundi, along with evidence from data these questions and issues are for ‘them.’ ‘They’ collection from the four communes among the are the ones to make decisions and tell us. It is eight. not my place to discuss these issues and make decisions.” Others echoed the same sentiment, 4.3.1 Obstacles to Participation at the arguing that the time of the most poor and Local Level vulnerable is wasted in endless local meetings 26. The political participation of citizens that do not provide any immediate benefit and in Burundi is characterized as nominal only take away from valuable work in the fields rather than substantive. There are multiple and managing the household.107 opportunities for participation at the local 28. For some local officials, citizen level. A plethora of acronyms account for local participation is considered a means to development committees at the village and gather information at the communal level commune level, and school committees. There is without necessarily engaging citizens in also a prominent culture of social participation decision making. The vertical political culture in Burundi. Local communities demonstrate in Burundi was confirmed through feedback in strong social capital in organizing important the focus groups. Citizens reported the existence events such as weddings, graduations, and of different meetings, mostly around security- funerals. The mobilization of friends and related themes, where local officials provided families into planning committees for these information to citizens. The only meetings more events, with detailed calendars, responsibilities, substantively focused on development issues and the necessity to mobilize resources shows were those financed by donors to develop the capacity of Burundian citizens to participate, the new series of CDPs. These participatory plan, and execute. Groups of citizens within local meetings were held in every hill (Colline) within communities in Burundi (extended families and networks of neighbors) account for a high degree a commune that received support from a donor of social participation, but have very low levels to develop a new five-year plan. of civic participation (civic engagement). Why 29. Although local authorities have then is participation characterized as nominal, been organizing numerous meetings at with little openings for a transformative process the communal level, citizens still lack to take root?—for two reasons: exclusion (often opportunities to actively engage and discuss self-imposed) and stigmatization. local needs. The government organizes multiple 27. Social exclusion in Burundi is meetings at the hill (Colline) and communal widespread, and many respondents conveyed level throughout Burundi. Hill chiefs (Chefs a sense of being automatically excluded de Colline) are expected to organize a weekly from civic life. In all of the eight communes security meeting, and then ad-hoc meetings on surveyed, both the groups of local officials and development issues. The majority of citizens of citizens raised the same issue: achieving more meaningful participation is blocked by a pervasive problem of social exclusion in Burundi, 107  Quote from focus group, Commune of Rutegama. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 97 reported that these meetings are being used 4.3.2 Preparation of Communal as a vehicle to spread political messages.108 Development Plans Additionally, meetings to discuss relevant issues, such as security concerns or a new initiative in 31. The government-initiated decentrali- the commune, are dominated by local elites. zation reform process may be further One respondent stated, “There are no formal supported through community-driven invitations or notices for meetings. Everyone development. In 2005 the Government of just knows who is supposed to go and who is Burundi (GoB) embarked on a decentralization not invited.” He continued: “The two mandated program with the stated goals of strengthening meetings for social accountability are controlled social cohesion, improving local governance, by the elite from our community and are held and promoting access to basic infrastructure simply because it is a criterion now for the and service delivery. The World Bank responded performance indicators of the commune.”109 by supporting a request by the GoB to initiate a large-scale, community-driven development 30. A long tradition of the “tyranny initiative that responded to this notion of of participation” is culturally rooted, building social cohesion at the local level through particularly among the rural population. PRADECS. In 2007 the GoB, supported by the Demand-side governance in Burundi is very PRADECS team, launched the first CDPs in all weak. During focus group discussions it emerged 126 communes across the country for a period that even where some communal officials of five years. In 2012–13 second-generation attempted more interaction with citizens CDPs were developed in Burundi for another (for example, in the Commune of Rutegama), five-year period. obstacles remained—namely, self-exclusion (described above), poverty (as manifested 32. Focus group discussions revealed by lower education levels), and a dominant the success of the CDP planning process in perception among citizens that the state is not significantly augmenting participation of genuinely interested in open dialogue (this citizens in the identification of priority needs last point was most strongly voiced regarding in local communities. The process for the politics and the central level, while participants development of the CDPs demands community cited more interaction with communal meetings at the hill (Colline) and commune level, administration officials). There is a proverb in and sector input by local officials working on Kirundi that exemplifies a cultural norm—Nta health, education, agriculture, water, and so on. muntu vitumira mu rubanza— “everyone knows All focus group participants confirmed the open exactly who is invited and who is not to any given process undertaken to hold public consultations forum.” Therefore, a major obstacle to fostering in the CDP process, which was, however, largely greater social accountability in Burundi is not driven by donors. External facilitators were only institutional constraints, but the existence engaged to work in communes alongside local of cultural norms limiting entry points for more officials to hold meetings with community transformative engagement of citizens into groups. In one focus group (Makebuko), citizens civic affairs. This reality is compounded by the reiterated that because of the Gesellschaft deterioration of trust during cycles of violence für Internationale Zusammenarbeit (German ongoing since the 1960s. Agency for International Cooperation, GIZ) they received regular information on where and when meetings would be held.110 It would seem that 108  The rural communes in Burundi are largely dominated the mechanics of holding participatory meetings, by the ruling party, including the eight communes for the focus groups. Pockets of opposition areas in more urban settings would have a different perspective, perhaps, on security. 109  Quote from focus group, Commune of Gihanga. 110 Focus group discussion, Commune of Makebuko. 98 identifying and including vulnerable members when commune officials require informal of the community, and providing feedback contributions. Information on the realization along the process fell mostly to external actors of any element of the CDP usually takes the form engaged on donor-funded projects. This process of a request for a voluntary financial or material did, however, mobilize communal planning contribution. “We find out about the building teams that prepared a local prioritization and of a new school or a new road when someone planning document for the first time in Burundi. from the commune comes to our house to ask The development of new CDPs was largely for money as a contribution and also wants donor funded, with the process managed and volunteer labor and materials like sand and coordinated by the Ministry of Communal rocks for the construction.”112 The process for Development. developing the CDPs therefore lacks downward accountability and does not empower the 33. The process of developing population. The mechanical participation at the CDPs encouraged participation in the front end legitimizes the process as participatory, identification of priorities upwards, but but in the end the final document and the actual little or no information trickled back down selection of investments financed are removed the chain. Citizens confirmed their engagement from the citizens in planning meetings, but not one person in all eight focus group discussions with citizens 4.3.3 Budget Preparation and Execution reported having seen the final version of the CDP document. There seems to be no restitution 35. There are no practical links between of the content of the CDPs back to citizens. One the CDPs and the communal budgets. The communal administrator stated, “We can only development of five-year CDPs do not relate inform citizens of specific investment projects to any budget-planning exercise. The CDPs are when the financing is identified. We cannot essentially a list of all the needs in the commune, tell them the full plan as this would be political in a prioritized way. The needs in Burundi are suicide—our population would consider the overwhelming, creating a disconnect between dissemination of the plan as a promise that I the CDPs and communal budgets. This disconnect know we certainly cannot keep.”111 The CDPs is further compounded by the fact that all CDPs are positively viewed by citizens as well as seem unrealistic—there is no possibility that a communal administration officials; however, commune would ever be able to achieve, within it is evident that due to limited resources the the five-year period, the realization of the plan. dissemination of the plan is rarely achieved. For example, the recent CDP for 2013–17 for the Although the dissemination of the CDP would Commune of Rutegama requires $9.5 million for be useful in terms of information sharing, the all identified investments—that is, close to $2 major constraint of a lack of realism within the million per year to achieve the plan, when the plan, leaves little possibility of it being achieved. annual budget for the Commune of Rutegama Therefore, for the time being, the dissemination is currently only $182,000. Even when external of the existing plans to augment communication funding - by either FONIC or donors – is taken with citizens is not a benefit. into account, the resouces gap between the plan and the annual budget is still significant.113 34. Furthermore, many citizens stated Therefore, the CDP cannot yet serve the that they receive specific information on the financing of a priority in the CDP only 112  Quote from focus group discussion, Commune of Mutimbuzi. 111  Quote from one communal Administrator during a focus 113  Data collected from the communal accountant, Commune group discussion with communal administration officials. of Rutegama. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 99 Box 4.1 GoB Performance Evaluation of Communes In Burundi the Ministry of Decentralization and Communal Development has established a performance evaluation of communes on the basis that communal authorities have transparency and accountability obligations under the Communal Law. The performance evaluation process aims to enhance the quality of governance and management of each commune. Conducted on an annual basis, this practice is based on a number of objectively verifiable and easily measurable criteria. With 20 criteria subdivided into 3 assessment areas—(i) administrative governance of the communes (10 criteria), (ii) management of budgetary resources of the communes (6 criteria), and (iii) social inclusion and gender (4 criteria)—all criteria are weighted at 100 points. Thus, all the communes of Burundi, after evaluation, are classified according to their scores. Based on the evaluation results, an additional financial allocation above the equal 50 million per commune is made by the National Communal Investment Fund (FONIC) to communes deemed successful, meaning those which earned 60 or more points out of 100. Source: Government of Burundi, Manual of Performance Evaluation 2010. purposes of a budget planning document and information on expenditures requires too much rather reflects a “shopping list” of all commune transparency for local officials to share this activities that ought to be implemented. information.114 This contradicts the goals of the new decentralization process noted above. Also, 36. The communal budget has been although every commune visited had the budget publicly disseminated in all communes since physically posted, some had sheets missing or 2011, but this is either not known or cannot torn down because of wind and rain, and others be accessed by the public. In all communes were posted up too high on the wall to be read. visited we confirmed by visual inspection that Thus, meaningful communication of the budget the communal budget is indeed posted publicly to citizens is lacking (see figure 4.2). at the communal administration office for the public to review, as a requirement in the 4.3.4 Central Transfers: The Case communal performance indicators (see box of FONIC 4.1). The dissemination of this information 38. Local administration officials have a is, however, not understood as the budget is strongly negative perception of FONIC due to written in French and most local citizens do not low levels of decision making on the use of read French. Also, the format is complicated and funds. Local citizens hold limited knowledge of covers multiple sheets of paper with different the work of FONIC, and other than knowing that line items and amounts—there is no simple it exists and finances the construction of schools, budget summary posted in an easy-to-read there was little citizen feedback regarding FONIC. format in Kirundi for citizens to review. But this was made up for by the overwhelming 37. The budget information posted is views of local officials. Globally, a strong negative of the annual budget, but there is no public perception emerged on three issues: (i) a total posting of actual expenditures. The communes lack of decision making in the use of funds in Burundi do not provide information on locally, (ii) partial payments and late payments actual expenditures and yearly reconciliations to the public. One respondent during a focus group discussion stated that providing 114  Feedback from citizen during the focus group in the Commune of Gihanga 100 and each commune will be Figure 4.2 Public Posting of Communal Budget responsible in managing its own budget. 4.3.5 Tax Collection 39. There is no uniform way that tax collection is carried out at the communal level in Burundi. In each commune visited the team queried local officials on the method used for local tax collection, and it emerged that each commune develops its own system. One commune cited the requirement to follow the communal guidelines outlined in the Administrative and Financial Guide developed by the Ministry of the Interior, though in practice this was not Source : Fieldwork visit to the Commune of Makebuko undertaken by the authors, September 2013. applied.115 Most communes employed, until recently, the elected hills chiefs (chefs collinaires) to oversee (which created problems in implementation tax collection. The hill chiefs worked in a team and completion of projects), and (iii) materials with the other members of the hill council to delivered in kind (as opposed to cash) from collect taxes, providing the “quittance” (receipt) Bujumbura rather than purchasing materials to citizens who paid, and depositing the funds locally. These three issues solidified the view with the communal accountant. Ten percent of of local administration officials that they have the amount collected is kept by the tax collectors little input into the use of funds transferred as payment. In the last year, there has been a shift from the central to the local level in Burundi. by communes to either hire all tax collectors Several communal administrators voiced the on contract, or use hill chiefs and contracted concern that they are not able to finance their tax collectors jointly. Only one commune out priorities, but must follow a standardized plan of the eight visited used solely contractual tax for the construction of primary schools. From collectors. these views, two important issues emerge. First, there is little to no citizen participation 40. A lack of incentives leads to in the investments financed by FONIC, or the widespread leakage. All focus groups, both monitoring and oversight of funds received. citizens and public officials, spoke about the Second, there is no scope for either citizens or high level of leakage of tax funds collected. local officials to engage in the decision-making Two main reasons were advanced: first, the process of which investments to finance with remuneration is so small for tax collectors, support from the central level. It is expected that the recent FONIC reform (2014) will help transform this negative perception. This reform 115  Data collected from the four communes available upon is to provide funds to the communal accounts request. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 101 that the temptation to steal funds Figure 4.3 Rutegama Commune Revenues, is overwhelming. For example, Actual and Projected (2010–13) in the Commune of Rutegama a tax collector is paid after five 140000000 receipt books (quittanciers) are 120000000 submitted. The payment for the tax collector is 10 percent (FBu 100000000 5,000), with a value of FBu 50,000 80000000 collected on each book. It can take 60000000 Projected Revenues almost two months to collect all Actual Revenues 40000000 five receipt books, meaning that a 20000000 tax collector is unable to maintain a stable income, and when paid the 0 Revenues Revenues Revenues Revenues amount is equivalent to only $17. 2010 2011 2012 2013 The second reason put forward is the personal web of hillside connections that places sharing Source: Fieldwork data collected by authors, September 2013. goods higher over payment in cash. This informal system is deeply embedded over the use of tax funds collected, citizens in local communities in Burundi. Several focus are more willing to contribute with nonfiscal groups spoke of the traditional system of contributions for local investment projects. sharing banana beer and brochettes with the Traditionally, when the commune embarks on hill chief as a form of tax payment. One woman an investment project linked to the CDPs, all stated, “When the hill chief, who is my neighbor, households that will benefit from the project collected the taxes he always told me in advance contribute in cash. There is a formalized structure when I needed to prepare my money. He also around this process where local representatives let me pay him in-kind. Now that the commune are elected to collect the cash, and account for has hired tax collectors that I do not know, they its use in the project. Many citizens stated that show up at any moment demanding payment in because they can visibly see the development of cash.”116 the project, combined with the knowledge of the amount that was collected locally, they are more 41. Deep mistrust over the use of tax willing to contribute. revenues is a disincentive for citizens to pay. An overwhelming view emerged from all citizens’ 43. The willingness of citizens to pay a focus groups—they do not understand the nonfiscal contribution points to an entry method of tax collection, they do not understand point for social accountability to augment where the money goes or its use, and there is no the capacity of communes to generate more public accounting of funds collected. None of the revenue through tax collection. We analyzed communes visited publicly post the amount of two communes more in depth. One commune, taxes collected, and there is a strong sentiment Rutegama, had embarked on a process of and perception of high levels of corruption fostering greater social accountability with the around tax collection. explicit goal of increasing tax revenues. The new communal administrator elected in 2010 set 42. Because of a deep sense of mistrust about local reform by (i) revising the mechanism for tax collection, hiring both elected hillside officials and contracted tax collectors; (ii) 116 Quote from focus group, Commune of Gisozi.   revising the system for funds to be deposited 102 transparency. The communal administration, Figure 4.4 Makebuko Commune Revenues, Actual and Projected (2010–13) since 2010, has allowed local CSOs to monitor tax collection and review the use of funds toward the communal budget. Citizens in Rutegama 70000000 have access to the same information, but are 60000000 largely unable to read the documents in French 50000000 and decipher the long budget format. The local Projected 40000000 Revenues association acts as a bridge between citizens and the administration, explaining to citizens during 30000000 Actual Revenues church services and community meetings the 20000000 content of the documents (communal budgets, 10000000 revenues collected and method used, use of 0 taxes within the commune, and so on). Revenues Revenues Revenues Revenues 2010 2011 2012 2013 45. Political discords undermine social accountability mechanisms within the commune. A second commune reviewed in depth Source: Fieldwork data collected by authors, September 2013. did not achieve the same results. Makebuko was reviewed because of a known problem between communal officials and the former administrator in the communal account and oversight of and allegations of corruption. Review of tax these funds; (iii) hiring a local civil society revenues (figure 4.4) collected in the same organization (CSO) to work in partnership with period from 2010 indicates that Makebuko the commune to explain to citizens the purpose achieved much less progress. Although revenues of paying taxes and their use; and (iv) inviting are mostly increasing (aside from 2012), there local CSOs to meetings to discuss the budget and is a larger discrepancy between the projected tax collection. The evolution of tax collection in and actual revenues. the Commune of Rutegama is outlined in figure 4.3. The interesting element here is that one of 46. Focus group discussions varied across the best-performing communes is collecting more revenues than Figure 4.5: Comparison of Revenue Collection Efforts in expected—they are outpacing their Makubeko and Rutegama (2010–13) projections, which they have been 140000000 increasing year after year since a new communal administrator took 120000000 over. 100000000 Revenue collection 80000000 44. Increasing transparency may also increase communal 60000000 revenue collection. During 40000000 the focus group discussion in 20000000 the Commune of Rutegama, the 0 2010 2011 2012 2013 representative of a local CSO Makebuko 27769145 35942700 41219100 56493790 was consulted on the process of Rutegama 44592956 67286760 95459688 117538216 budgeting and tax collection. The most salient aspect of success in Source: Fieldwork data collected by authors, September 2013. Rutegama may be summarized as Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 103 communes, with some being tenser than by corruption allegations and a rotation of others. One citizen left the focus group meeting administrators. Figure 4.5 shows the results over stating, “You are asking me about use of public time, indicating a significant “performance gap” funds and the administrator is right next door. I between Makubeko and Rutegama in collecting don’t want to be in prison tomorrow.” In terms of revenues. The two communes are geographically revenue mobilization, one commune (Rutegama) close, they both have administrations from the demonstrated a willingness to engage with ruling party, while the commune of Makebuko has civil society actors and citizens, and focused its a larger population, and there are no significant efforts on raising communal revenues, while the commercial activities in either commune to other commune (Makebuko) has been plagued account for the gap in revenues. Box 4.2 Participatory Budgeting (PB) in South Kivu, DRC: Supporting Decentralization and Empowering Citizens to Participate in the Budgetary Process through ICT In the Democratic Republic of Congo, the first cycle of the Participatory Budgeting (PB) initiative took place in the province of South Kivu (population: 4 million) in 2011, and targeted eight subprovincial decentralized entities including the capital (population: 250,000). The project helped address a major bottleneck in the country’s recent decentralization agenda, a high priority reform enshrined in the DRC constitution given the dysfunctional transfer of funds between different territorial entities; as well as the promotion of good governance—a top priority action of the DRC 2008–11 Country Assistance Strategy (CAS)—through effective decentralization (to bring the government closer to its citizens) and local governance. Through the World Bank Institute (WBI) diagnostics and studies in 2010, a series of challenges that needed to be addressed were identified to improve governance and service delivery at the local level, such as asymmetry of information between stakeholders and low levels of understanding of budget procedures and engagement with civil society. These diagnostics indicated that PB as a social accountability mechanism could be a promising entry point for a potentially broader governance reform in the country. The province of South Kivu was considered the most promising area, as there was an interest on the part of the communities to participate in a PB process. The findings in the diagnostics also pointed to the possible leverage effect the use of ICT could have in terms of enhancing the process through engagement, information distribution, monitoring, and follow up. The project mobilized a reform coalition of civil society, private sector, IT experts, and both the provincial and local governments to work on using ICT to engage citizens on local budgets. As the mobile phone penetration is relatively high in the country (estimated to reach 47 percent in 2013, with 55 percent of the population residing in areas currently covered by mobile networks, including most rural areas in the eastern province of South Kivu), the precondition for successfully applying an ICT component in the project existed. Utilizing mobile phones would therefore contribute to creating a more inclusive (women, minorities), transparent, and participatory decision-making process on budget allocations and monitoring of public spending. Mobile phones were integrated in the process in the following ways: (i) citizens living in a given area were informed via SMS about the next meeting, (ii) citizens on a trial basis could vote via SMS on which public works they wanted to be addressed, (iii) citizens were informed on the outcome of the vote via SMS, and (iv) citizens could provide feedback on the quality of projects under implementation, or those already implemented projects by SMS. For the first time, the PB process allowed citizens to determine how public funds should be spent while establishing their priorities. Moreover, an equal “first” was the decision by the provincial government to transfer funds to the local governments. Although this is required by law, it has been 104 ignored in the past. As a result of the process, the ministry of budget has also institutionalized the PB process in South Kivu, mandating all local governments to start following the PB principles while drafting their budgets. Also the quality of the budget-drafting process has improved. Rather than ending up being a wish list, budgets are now formulated in a more realistic manner with expenditures better aligned with the revenue side. With the increased transfer of funds from the provincial level to the local level, local governments began allocating part of their budgets to investments, rather than to funding salaries and equipment only. Some communities have gone from having no budget for investments to having a 40 percent budget allocation. With more budget resources available, they are also having more resources to deliver services to the poor. In Luhindja the process has resulted in the repairing of 54 classrooms and a bridge, in Bagira a health center has been established and the sewage system has been repaired, and in Ibanda water fountains and restrooms are being built. One of the more encouraging results connected to the project is a fairly substantive revenue increase at the local level. More citizens seem to be willing to comply with their tax obligations, as they associate the improved service delivery with government spending. There is evidence that since the process started, local tax revenues have increased up to 20 times. Another important achievement of the project was that, strong country ownership was built leading to the institutionalization of the PB process. Following an executive Government Decree, a law was passed by the Provincial Parliament in 2012 to institutionalize the practice in the province. In South Kivu the PB process is now locally sustained with local actors taking strong ownership and extending subsequent cycles to the entire province on their own. Other provinces in the country are following suit and starting to replicate and adopt PB processes. Source: World Bank, WBI 2012. 47. The communal administrator in 4.3.6 Monitoring Project Execution: Rutegama acknowledged that they are A Case Study from the Communes starting a reform process by hiring new tax of Makebuko and Rutegama—The collectors, creating a new payment system, Experience of the World Banks’ CDD and revising the mechanism for monitoring Operation in Burundi of tax collection. The data collected from the Commune of Rutegama confirms the 48. The World Bank–financed community incremental, but steadily increasing, revenue and social development project (PRADECS) collection. This finding is consistent with similar was developed following the civil conflict and experiences in other fragile setting contexts, in the midst of a major political transition where innovative reform initiatives to involve in Burundi. The decentralization process the citizens in the decision-making process act initiated by the government in 2005 aimed as a key driver to improve revenue collection. In to empower communities to improve local the South Kivu region of the Democratic Republic governance and social cohesion, and to improve of Congo (DRC), for example, the establishment access to socioeconomic infrastructure. The of the participatory budgeting (PB) through five-year, $40 million project was designed as a information and communications technology community driven development (CDD) project (ICT) mechanisms—mainly mobile phones— with a two-pronged approach to support the was instrumental in significantly increasing new national decentralization strategy, and to local tax revenues up to 20 times in 2012 (box ensure improved and more equitable access to 4.2). local socioeconomic infrastructure and service Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 105 delivery in rural areas where over 90 percent of 51. The research team completed focus the Burundian population live. group discussions in eight communes across Burundi. Two of these communes, Rutegama 49. The overall objective of the project and Makebuko, were chosen for further in- was to establish and operate a decentralized, depth follow-up to qualitatively assess the role transparent, and participatory financing of citizen participation in the selection and mechanism that empowered local govern- implementation of PRADECS subprojects. In ments and communities to ensure improved each commune, two subprojects were reviewed. and more equitable local service delivery. The This analysis is therefore based on the review of project consisted of two major components: the: (i) capacity building of both local officials at the commune level, as well as training of 1. Commune of Rutegama, construction of officials within the central government working a new primary school on decentralization; and (ii) the financing of demand-driven subprojects within most 2. Commune of Rutegama, construction of rural communes of the country. The bulk of a communal college (akin to secondary the financing, $30 million, went toward the school) realization of 1,067 subprojects. 3. Commune of Makebuko, procurement of 50. PRADECS closed in December 2012, school furniture to various schools as both the Bank and the GoB concluded that 4. Commune of Makebuko, construction of the project was successfully implemented. For a new health center the Bank, there was recognition that PRADECS benefitted from government ownership and 52. The objective of this analysis was to leadership. Furthermore, through PRADECS, concretely tie similar questions that were the communal administrations were able, for posed during focus group discussions to the first time, to take direct responsibility concrete projects in local communities. over funds for subprojects. This included the This then provides an illustration to further development of the first CDPs, the identification analyze the key findings of this chapter. The of priority subprojects through participatory main questions asked to local citizens who were community meetings, management of the somehow involved in the subproject were: (i) procurement process, monitoring of the who identified the subproject, (ii) how were you subprojects, and finally guaranteeing the informed of the subproject, (iii) what was your maintenance of the newly built communal contribution toward the subproject, (iv) were you infrastructure. But a major drawback was that involved with or informed of the procurement although the government demonstrated strong process, (v) as a citizen what was your role in performance with PRADECS, there was still no subproject monitoring, and (vi) how was the operational mechanism for financial transfers maintenance of the subproject organized. The from the central government to the communes. detailed matrix showing feedback on each of There were limited transfers from FONIC, these questions by project is available, but, but the perception of local administrations is globally, a trend emerged from all four projects that these transfers lack both a transparency that parallels feedback from the focus groups and a participatory element. As PRADECS was discussions. These findings include: positively evaluated, primary research for this chapter included the review of four PRADECS 53. The identification and selection of subprojects to illustrate the role of participation subprojects financed through PRADECS was and monitoring by citizens. Box 4.3 captures this affected by elite capture—and this extended analysis. into project implementation to include 106 Box 4.3 Balancing Short-Term Needs with Long-Term Goals: Findings on Community Driven Development (CDD) Projects from the 2013 IEG Report Community-driven development (CDD) projects supported by the World Bank have evolved over time from the initial aim of providing a safety net toward multiple objectives relevant to the FCS context. CDD projects emerged initially in the form of social funds during the 1990s, as a safety net to offset the adverse impacts of structural adjustment loans. When the social development, agriculture and rural development, water, and urban sectors adopted CDD modalities, their scope expanded considerably. Within the FCS they have had three different objectives involving various forms and degrees of collective action, the emphasis varying across different projects: providing public goods and services to local communities, building local institutions, and enhancing downward accountability to citizens. In FCS, when implemented well, they have proved to be efficient vehicles for expanding the reach of the state and building state legitimacy. Several country directors have argued that CDD projects are quick-disbursing projects that are crucial to maintaining a good disbursement ratio and, for that reason, have been reluctant to allow their design to evolve over time. CDD projects have grown by number and commitment volume much faster in FCS than in IDA countries that were not FCS. In several conflict-affected states, these projects have been the principal means of state engagement with distant communities. They have served to establish the presence of the state, and besides their short-term contribution to local economies, they have contributed to state legitimacy. These projects cover a considerable range of demand-driven projects across many sectors. Most of them finance local infrastructure subprojects or service delivery subprojects; although some of them, like the Nepal Poverty Alleviation Fund (PAF) include a mix of subproject financing and micro- lending to local communities. They have three common features: they are typically demand driven; involve some form of community organization and community participation, and are administered by a special entity established by the government outside the structure of line ministries. Afghanistan, Nepal, and the Republic of Yemen have well-established, large-scale CDD programs functioning efficiently for more than a decade and reaching out to distant rural communities. Haiti too has relied on CDD to provide services to urban slum dwellers. One noteworthy feature has been their outreach to women and their ability to involve women in decision making, as evidenced by impact evaluations from all three countries. But the reliance on project financing remains the Achilles heel of CDD programs, and without institutional evolution the second and third objectives of building sustainable local institutions and promoting downward accountability may not be achieved. Despite its wide use, CDD projects remain controversial. One aspect of controversy is whether CDD projects undermine local government by creating parallel structures. The other aspect is the extent to which these projects lend themselves to elite capture. A recent policy research report concludes, “Only when projects explicitly link community-based organizations with markets, or provide skills training, do they tend to improve group cohesiveness and collective action beyond the life of the project” (Mansuri and Rao 2012). The report finds that community participation programs are most effective when they are systematically linked to state institutions: “Most successful programs tend to be those implemented by local governments that have some discretion and are downwardly accountable . . . Local participation appears to increase, rather than diminish, the need for functional and strong institutions at the center.” Mansuri and Rao conclude that local participation does not work when it is merely the ad hoc creation of a project. CDD programs involve institution building at two levels—the community level and the program level. The FCS case studies show that it is relatively easy to form ad hoc, informal community organizations when they are the conduit for receiving development grants. Their sustainability depends on the sustainability of the program and continuity in the stream of benefits. As in the case of microfinance, Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 107 the support structure at the program level needs to be sustainable to sustain community involvement. Alternatively, the program can be linked to local government structures with regular fiscal transfers from the annual budget, as is the case in Bangladesh and Indonesia. Such a link between CDD programs and decentralized local government structures has not been established in any of the FCS. On the contrary, they appear to compete for the local governance space. Nor has any alternate institutional structure been established for these programs in FCS, even though viable apex organizations—microfinance in Afghanistan and Bangladesh and the PAF in Pakistan—have been established in fragile situations. The consequence is that CDD programs have not joined up with local governments, nor do they have any alternative system of ensuring institutional sustainability; as a result, they remain completely donor dependent. The National Solidarity Program in Afghanistan is the largest CDD program in FCS. It has been very successful in sponsoring the CDPs across the entire country, financing local subprojects, promoting local governance, empowering women, and, in turn, enhancing the writ and legitimacy of the Afghan state. Its financing model, however, is unsustainable and depends on large injections of donor financing for one-off grants to local communities. In FY2014, more than a decade after its inception, the program is reported to be undertaking a set of studies to develop a more sustainable model, but with the 2014 transition fast approaching there is very little time left to test out and institutionalize an alternate model. Programs should start planning for institutional sustainability much earlier. The experience of Afghanistan’s CDD program is similar to that in many other FCS. The programs play an important role in addressing fragility through short-term, intermittent measures to support local communities, when they are originally designed. But their very success as a safety net becomes an impediment to their evolution and institutional sustainability. It would be useful to take stock of this genre of projects to help adapt and institutionalize them within country systems and country budgets over time. The Bank has not instituted alternate mechanisms to ensure their viability beyond the life of the projects supporting them. As a result, CDD projects have been found to be unsustainable. Source : World Bank IEG Report 2013. control over implementation by contractors, contributing to access to social services within with little role for citizens to monitor the the commune. The provision of school furniture execution of works. All four subprojects, based was, however, negatively viewed as, essentially, on citizen feedback, were selected by a small being controlled solely by the administrator with group of local officials. In each of the four, the all aspects of procurement hidden, even from communal administrator was the main person other officials working within the commune. to coordinate the selection of the subproject. Indeed, PRADECS was established to principally 54. Akin to feedback from focus group work with the administrators, through discussions, participation by citizens in the participatory meetings between officials and subprojects was only in the form of receiving community members to open up dialogue and top-down information as well as requests decision-making responsibilities. One of the four for contribution toward the project both subprojects, the provision of school furniture financially and with physical labor and in Makebuko, elicited controversial feedback materials. The construction of the primary from citizens. The other three subprojects were school in Rutegama provides a useful example. all similarly acknowledged: they included only The citizens interviewed stated that they and a limited participatory element in decision the majority of the community were made aware making, but were viewed positively for of the financing for the primary school by the 108 hillside chiefs. This information came through a 56. Qualitative interviews with citizens notification of the project, along with a request to concretely discuss specific subprojects for each household to contribute financially (to exposed a view held by citizens that make up the 3 percent community contribution demanding more accountability of local required under all PRADECS projects) and to help officials would appear as “rebellious” in complete the levelling work and deliver rubble the eyes of those in power. To maintain what to the construction site. The dissemination of citizens referred to as “social peace” the view the school furniture subproject in Makebuko was not to question too extensively the actions is even more serious (with respect to a lack of of local officials. More concretely, during participatory process) in that beneficiaries and interviews it was exposed that among all four school officials were informed of the project at subprojects there were significant deficits in the time of delivery of the furniture. There is the way that local officials included citizens; speculation by both citizens interviewed as well however, when the projects ultimately provided as other communal officials that the supplier improved services there was a level of acceptance paid the 1 percent communal contribution and and status quo was maintained of little social a 3 percent community contribution to obtain accountability. Anger and strong views emerged the order; thus the communal accounting around the one subproject in Makebuko (for the service was never involved in any aspect of this provision of school furniture) which the citizens subproject. concluded was fictitious, in the sense that the administrator controlled the entire process and 55. Although the procurement process delivered furniture that a few months afterwards and the monitoring of the execution of was broken.117 the subprojects was largely managed by the same individuals who selected the 57. In conclusion, the views that emerged projects, along with the communal engineer from these interviews, combined with and contracted firm, local management feedback from the focus group discussions, committees were allowed limited input. captures a picture in Burundi of extremely The decision making around the selection of low levels of social accountability, an absence the firm was not known by any citizens across in demand for greater social accountability all four subprojects; however, particularly in in rural areas, and limited attempts by local the Commune of Rutegama, the established officials to foster stronger relations with citizens management committees were able to modestly using social accountability tools. Referring back influence subproject monitoring and execution. to the macro framework used in this analysis, These management committees included many the constituent element of social accountability officials from the communal administration, but that is noticeably the weakest is the citizen also hillside chiefs as well as representatives action element. Culturally, the view cited in from communal development committees—the the paragraph above by citizens regarding an latter group being more representative of the acceptance of the status quo to maintain “social local community. The management committees peace” with the local elites will be difficult to were all formulated by the Communal overcome. Although the country is currently Administrator, but in Rutegama they were more experiencing political volatility in the run-up broad-based and representative. Concretely, to national elections in 2015, the state action during the construction of the communal college (secondary school), the committee discovered that workers were stealing bags of cement at the construction site. The committee 117 A communal official from Makebuko stated during an then took action and monitored execution of the interview that more than 60 percent of the furniture was broken and damaged a few months following delivery, and even on subproject more closely. delivery it was obvious that it was poor quality. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 109 constituent element provides more tangible actual budget expenditures from the previous and immediate openings to foster social year. accountability, specifically around a desire by local officials to increase tax revenues to gain 4.4.2 Increase Awareness-Raising Efforts further autonomy from the center. The following Among Both Local Citizens and Elected section provides specific recommendations for Officials the road ahead. 60. Evidence suggests that information alone is rarely sufficient to improve 4.4 Recommendations accountability outcomes (O’Meally 2013). 58. The recommendations for increasing In focusing on a multipronged approach to social accountability in Burundi are geared accountability reform, it will be essential to toward state action emerging from the political supplement any information/communication society domain, and the role of information. strategy with robust awareness raising efforts Building citizen demand is an important to support citizens in an ongoing and sustained element in fostering social accountability, but engagement with local authorities. Moreover, within the Burundi context it appears that this supporting the awareness of locally elected must be done in tandem with encouraging state officials is as important. From interviews with developmental responsiveness. The argument locally elected officials, communal councilors presented above is that a major obstacle is the justified their lack of transparency by the fact lack of demand by citizens (lack of collective that they were elected by the population. Since action and disengagement among citizens), every decision they take is on behalf of the signaling that a more relevant entry point for population, no further account to the population Burundi is first within the political society should be necessary (communal data collected domain. in Makebuko, p. 24).118 4.4.1 Support a Comprehensive 61. Given the existing role of CSOs at the Communication Strategy communal level in disseminating information, specific recommendations are to support 59. Information is the single greatest CSOs to (i) increase citizen participation through entry point for fostering greater social more proactive awareness-raising strategy, (ii) accountability in Burundi. All citizen focus actively support citizens in contributing toward groups spoke about the lack of information communal projects and activities, and actively regarding communal affairs. The situation demand accountability to local authorities via presents itself as either information is posted, but it is neither well communicated nor understood, their contributions, and (iii) actively support or there is a complete absence of information citizens in the transformation of traditional and communal affairs continue under a veil of political relations. secrecy. Specific recommendations include: (i) 4.4.3 Empower Civil Society Actors to increase local citizens awareness of meetings, Bridge the Gap between Local Citizens the venues, and rights to participate; (ii) simplify and Public Authorities the communal budget format, to include closing and opening balances of not more than one 62. Greater engagement with civil society page; (iii) translate the budget into Kirundi and actors at the local level has the potential disseminate it to the hillsides to be read aloud to bridge the gap between citizens and during local development meetings; (iv) post, at least monthly, the tax revenues collected by the commune; and (v) post in a simplified format in 118 Full data collected for the Commune of Makebuko available Kirundi, at the beginning of the fiscal year, the upon request. 110 administration officials, and foster a culture group discussions point to the readiness of of greater social accountability. The lack most communal officials, particularly around of demand-side accountability in Burundi is the necessity to engage more with citizens striking at the local level. The capital and larger with respect for increased revenue collection. towns have a vibrant civil society scene, but the Communes visited where citizens provided majority of rural communities have an absence more positive feedback on performance of the of civil society actors that focus on governance commune (namely Rutegama and Gisozi for and accountability issues. Local CSOs in rural example) had already commenced with limited communities that focus largely on development reforms. These communes require support to objectives are technical in nature. Acknowledging standardize approaches and provide guidance the role played by GIZ in the CDP process by on specific tools that may be appropriate for disclosing regular information on meetings their use. But communes where leadership and their locations to citizens and identifying was lacking (recent turnover of communal mechanisms to include vulnerable members administrators, and more complaints by citizens of the community, it seems there is an entry around high levels of corruption in the focus point for communal administrations to engage group discussions) will have a more difficult with CSOs to increase awareness of citizens on time accepting reforms that demand greater civic matters. Working through local CSOs will transparency and open dialogue with CSOs act as a bridging mechanism to enhance a more and citizens. Based on the primary research positive social contract in Burundi. Returning undertaken for this chapter, communes in to the conceptual framework, to bolster the Burundi have already begun to emerge on constituent element of “information,” the two different tracks—those that are working average Burundian citizen requires assistance toward citizen participation and more social to understand information that has not been accountability, and those that are grappling widely discussed in public to date—such as the with good governance issues within their communal budget and tax revenues collected and internal administration. The latter is required to their use. Specific recommendations include: (i) commence on the first track. communes to develop a partnership with local 4.4.4 Harmonize Tax Collection across CSOs specifically around recommendations Communes cited above (which requires training and an awareness on both sides on which issues should 64. A transparent and harmonized be the focus of actions), (ii) CSOs to monitor method for tax collection, that acknowledges communal budget planning and tax collection, the prevalence of leakage of funds, should (iii) civil society to voice problems as they be adopted across communes. During focus arise and encourage citizens to learn the role group discussions all groups expressed an of local government and the responsibilities enormous problem around moving beyond a of citizens, (iv) CSOs to encourage citizens to local tradition of “in-kind” tax payments over fully participate in local meetings (particularly cash tax payments. Focus group discussions the twice yearly “accountability” meetings in with communal officials showed that communes each commune), and (v) CSOs to advocate for are ready to engage in reforms and work with vulnerable groups within society and raise citizens to increase revenues. The example concerns of these groups to local administration of the Commune of Rutegama demonstrates officials. concretely the possibility of employing a social accountability strategy to increase, in the 63. The above recommendations rely first instance, confidence among citizens to heavily on communal administrations that engage with the state, and subsequently to are prepared to reform and hold themselves increase local tax revenues to be used for accountable to citizens. Feedback from focus communal investments. Recommendations Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 111 include: (i) linked to the above category, launch an relating to the political, social and economic information campaign targeted toward citizens situation of the commune and its prospects.” and local civil society actors on the reason for paying taxes, and more importantly the use of 66. Although interviews with local tax revenues; (ii) to repeat from above, post on authorities reveal that they are well aware a monthly basis the tax revenues collected by of communal council meetings and their the commune; and (iii) move away from using activities, discussion groups with local elected officials, such as hillside chiefs, as tax citizens reveal the contrary. Some local citizens collectors, and hire and train contracted tax stated that: “heads of hillsides (chefs de collines) collectors. A reasonable salary scale will need are invited but rarely are hillside councilors to be instituted to mitigate against leakage of (conseils de collines) invited”; others think that funds. meeting dates of the communal council are not published or announced; and some are not sure 4.4.5 Support Citizen Engagement whether they have the right to attend communal and Participation within the Current council meetings. In practice, the bottom-up Decentralization Framework approach to the decentralization process does 65. A thorough review of the existing not hold true, and accountability continues to decentralization framework, and its rules run upward to the center. and procedures, provides opportunities for citizen engagement within the current context 67. Cognizant that the Burundian and process. For example, the communal law decentralization process aims to empower stipulates that: “Le conseil communal organise and engage citizens through its various au moins deux fois par an des rencontres ouvertes texts and procedures, and also that this aux conseils de collines [et….], pour les informer de framework is known by local authorities, manière transparente sur la situation politique, it is important to ensure that it is practiced sociale et économique prévalant dans la commune in the field. Therefore, it is recommended et ses perspectives” (GoB Manuel d’évaluation de that country partners: (i) be fully aware of la Performance des communes), which translates the various opportunities available for citizen (by author) as: “the communal council is to hold engagement within the current context, and (ii) at least twice a year meetings that are open to support local officials, including locally elected hillside councilors (conseils de colline) [and . . .] citizens, in abiding by the decentralization rules to inform them and be transparent on matters and procedures in place. 112 References Ali, D.; Deininger, K.; and Goldstein, M. 2011. Environmental and Gender Impacts of Land Tenure Regularization in Africa: Pilot Evidence from Rwanda. Policy Research Working Paper 5765 World Bank, Washington, DC. Association Burundaise des Elus Locaux (ABELO). 2009. L’État actuel de la Décentralisation et la Situ- ation des Communes au Burundi. Etude de Base. Octobre (ABELO 2009) Baltissen, G, Sentamba, E. 2010. Programme et Plan d’action opérationnel de renforcement des ca- pacités des acteurs en charge de la mise en oeuvre de la politique de la décentralisation et du développement communautaire (PRADECS). Banerjee, A.; and Lakshmi, I. 2005. “History, Institutions, and Economic Performance: The Legacy of Colonial Land Tenure Systems in India.” American Economic Review. Berry, S. 2009. “Property, Authority and Citizenship: Land Claims, Politics and the Dynamics of Social Division in West Africa.” Development and Change 40 (1): 23-45. Bezabih. M. and Holden, S. 2010. “The Role of Land Certification in Reducing Gender Gaps in Pro- ductivity in Rural Ethiopia: Environment for Development. EFD Discussion Paper, RFF Press, Washington, DC. November. Bigirimana, G. 2012. Étude sur la situation des conflits fonciers au Burundi. RCN Justice & Démocratie. Septembre Bigirimana, G.; Thinon, P.; Sindayigaya, S. and Seberege, P-C. 2014. PRODEMA: Revue du Secteur Fon- cier au Burundi. (PRODEMA Report). Unpublished draft. January Boeckmann, S. 2012. Social Accountability in Fragile States: A synthesis of Findings from the Available Literature. Social Accountability and Demand for Good Governance Group, The World Bank. March Boulenger, S., Gauthier, I., & Vaillancourt, F. 2012 Déconcentration, délégation et dévolution : avan- tages, inconvénients et mise en place. Aout (Vaillancourt et al. 2012) Brachet, J., & Wolpe, H. 2005. Conflict-Sensitive Development Assistance: The Case of Burundi. Social Development Papers, The World Bank. June Brosio, G., & Singh, R. (forthcoming) Revenue Sharing of Natural Resources: A review of International Practices. Background Paper, the World Bank 2014. Bukenya, B., Hickey, S., & King, S. 2012. Understanding the role of context in shaping social account- ability interventions: towards an evidence-based approach. Report commissioned by the World Bank’s Social Accountability and Demand for Good Governance Team. June Byamugisha, F. 2013. Securing Africa’s Land for Shared Prosperity: A Program to Scale up Reforms and Investments. World Bank, Washington DC. Colin, J. P. and M. Ayouz. 2006. “The Development of a Land Market? Insights from Cote d’Ivoire.” Land Economics 82 (3): 404-23. Coopération Suisse. 2012. Document de Capitalisation méthodologique des expériences pilotes en Province Ngozi. Programme d’appui à la gestion foncière au Burundi. Avril Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 113 Deininger, K. 2008. “A Strategy for Improving Land Administration in Inda.”Land Policy and Adminis- tration Note 33, Agriculture and Rural Development, World Bank, February. Deininger, K., D. A. Ali, and T. Alemu. 2011a. “Impacts of Land Certification on Tenure Security, In- vestment, and Land Market Participation: Evidence from Ethiopia.” Land Economics 87 (2): 312–34. Fenske, J. 2010. “L’Etranger: Status, Property Rights, and Investment Incentives in Cote d’Ivoire.” Land Economics 86 (4): 621-644 Ferrari, F. 2014. « L’Appui de la DDC aux Reformes Foncières en cours au Burundi et en RD Congo : Mise en œuvre de mécanismes de reconnaissance et de sécurisation des droits fonciers cou- tumiers » Working Paper presented at the Annual World Bank Conference on Land and Pov- erty, Washington DC. March 24-27 Galiani, S. 2011. “The dynamics of land titling regularization and market development.” Working Paper. Helsinki: United Nations University. Galpin, C., Niyonkuru, R-C., Habwintahe, JM., & Nkurunziza, A. Ingoro Yamotongo out le Service fon- cier communal : Province de Gitega. Gasabanya, Z. 2012. Bilan de mise en œuvre de la politique nationale de la decentralisation et du plan d’actions triennal de la decentralisation 2009—2011. Perspectives d’avenir pour la periode 2012—2014. Mai Gaynor N. 2011. Opportunities for Citizen Participation in Local Governance in Burundi “A Research Report prepared in collaboration with the Governance and Human Rights Programme of Tró- caire-Burundi” Gaynor N. 2013. The Tyranny of participation revisited: international support to local governance in Burundi. Oxford University Press and Community Development Journal. Guidance Note. 2005. Formalizing Informal Artisanal Mining Activity. A review and comparative analy- sis of mining codes and policy approaches towards Artisanal and Small-scale Mining. CASM/ BRGM Research Project. November Habwintahe, J-M. et Nkurunziza, A. 2011. Diagnostic Socio-foncier dans la commune Bugenyuzi de la province Karuzi, Bujumbura: Programme « Gutwara Neza ». Heinkele, J. 2011. Programme d’appui à la décentralisation et à la lutte contre la pauvreté au Burundi. Rapport préparé pour la GIZ Septembre Hilhorst, T. 2012. Innovating Land Governance in Rural Burundi to Reduce Conflict and Promote Food Security. Royal Tropical Institute. January Hilhorst, T. 2011. Burundi MJSP 2012-2015 “Feedback on law and regulation: Access to land”. Royal Tropical Institute. November Hood, R.; Husband, D. & Yu, F. 2002. Recurrent Expenditure Requirements of Capital Projects Estima- tion for Budget Purposes. World Bank Working Paper 2938 Huggins, C. 2004. Preventing Conflict through Improved Policies on Land Tenure, Natural Resource Rights, and Migration in the Great Lakes Region. African Centre for Technology Studies. Eco- Conflicts Volume 3. January International Crisis Group. 2013. “Comprendre les conflits dans l’Est du Congo (I): la plaine de la Ruzizi » Rapport No. 206 Juillet International Crisis Group. 2014. “Fields of Bitterness (I): Land Reform in Burundi” Report No. 213 February International Crisis Group. 2014.“Fields of Bitterness (II):Restitutuion and Reconciliation in Burundi” Report No. 214 February 114 Iyer, L. 2010. “Direct Versus Indirect Colonical Rule in India: Long-Term Consequences.” Review of Economics and Statistics, 92(4), 693-713. Kamungi, P. M., J. S. Oketch & C. Huggins. 2005. Land access and the return and resettlement of IDPs and refugees in Burundi, in C Huggins and J Glover eds., From the ground up - landrights, conflict and peace in Sub Saharan Africa: Pretoria, ACTS -Institute for security studies, p. 195- 268. Kohlhagen, D. 2007. Le tribunal face au terrain : Les problèmes d’exécution des jugements au Mugamba, dans une perspective juridique et anthropologique. Rapport de RCN Justice & Démocratie. Kohlhagen, D. 2009. Statistiques judiciaires burundaises: Rendement, délais et typologie des litiges dans les tribunaux de résidence. Rapport de RCN Justice & Démocratie. Kohlhagen, D. 2010. Land reform in Burundi: Waiting for Change after Twenty years of Fruitless De- bate. Kohlhagen, D. 2010. Vers un nouveau code foncier au Burundi? L’Afrique des Grands Lacs, Annuaire 2009—2010, p. 1. Kohlhagen, D. 2011. In quest of legitimacy: Changes in land law and legal reform in Burundi. In Natu- ral Resources and local livelihoods in the Great Lakes Region of Africa: A political economy perspective. Palgrave Macmillian. Kohlhagen, D. 2012. Burundi: Land policy making in a conflict-prone country. Summary paper for the Annual World Bank Conference on Land and Poverty, Washington DC, April 23-26 Larizza, M. & Glynn B. 2013. The Institution-Building Context in Sierra Leonne. In Barma, N.H; Huy- bens, E. and Vinuela, L. (Eds) Institutions Taking Root: Building State Capacity in Challenging Contexts. Washington D.C. World Bank Group Latouche, D. 2012. Programme d’appui à la reconstruction et au développement Communal dans la province de Rutana (PARDC-R) Burundi. Projet UNCDF May Lavigne Delville, P. (2000) Harmonising Formal Law and Customary Land Rights in French-Speaking West Africa, Paris: Groupe de Recherche et d’Echanges Technologiques. Leisz, S. 1996. “Burundi Country Profile”, in: J. W. Bruce (ed.), Country profiles of land tenure. Africa, 1996, Madison: Land Tenure Center, University of Wisconsin-Madison. Libecap, G. D. and D. Lueck. 2011. “The Demarcation of Land and the Role of Coordinating Property Institutions.” Journal of Political Economy, 119 (3), 426-467. Martinez-Vasquez, J., & Vaillancourt, F. Forthcoming. Obstacles to Decentralization: Lessons from the Developing World Draft. Maurer, N. and L. Iyer. 2008. “The Cost of Property Rights: Establishing Institutions on the Philippine Frontier under American Rule, 1898-1918.” Working Paper 14298, Cambridge, MA. Mbonabuca, T. 2010. La réforme administrative et fiscale de l’institution communale au Burundi à la croisée des chemins : opportunités et défis. Ministère de l’Intérieure. Gouvernement du Bu- rundi. Aout Meloche, J.-P. 2014. Gouvernance urbaine, mégavilles et décentralisation sous-locale : un tour d’horizon des Amériques. Mimeo. Mosley, P., 2014. Two Africas? Why Africa’s ‘Growth Miracle’ is Barely Reducing Poverty. Brooks World Poverty Institute Working Paper No. 191 Naritomi, J., R. R. Soarez & J. J. Assuncao. 2009. “Instiutional Development and Colonial Heritage with- in Brazil.” Working Paper, Rio de Janeiro. National Democratic Institute (NDI). 2010. Atelier de Concertation et de réflexion sur la gouvernance participative locale. Novembre Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 115 Ndikumana, L. 2005. Distributional conflict, the state, and peace building in Burundi. University of Massachusetts, Amherst Nihangaza. C. 2012. Etudes sur la viabilité institutionnelle et financière des communes rurales au Bu- rundi. Juin Nimubona, J. 2013. Revue institutionnelle et de gouvernance (IGR) sur la decentralization: Economie Politique de la décentralisation au Burundi. Rapport provisoire préparé pour la Banque Mon- diale Nimubona, J. Forthcoming. Revue institutionnelle et de Gouvernance (IGR) sur la décentralisation « Economie politique de la décentralisation au Burundi » Rapport provisoire Nkurunziza, J., & Ngaruko, F. An Economic Interpretation of Conflict in Burundi. Journal of African Economies, Volume 9, Number 3, pp. 370-409 Ntampaka, C., & Mansion, A. 2005. Etude sur la problématique foncière au Burundi. GRET, CCFD, ACORD. (general overview issue, policy , donor responses , recommandations) O’Meally, S. 2013. Mapping Context for Social Accountability. Social Development Department, the World Bank O’Neill, K. 2003. Decentralization as an Electoral Strategy. Comparative Political Studies 36 : 1068-91 O’Neill, K. 2005. Decentralizing the State : Elections, Parties and Local Power in the Andes. New York : Cambridge University Press. Observatoire de l’Action Gouvernemental (OAG). 2007. Analyse du processus de décentralisation au Burundi (Cas des Conseils communaux et des structures locales de développement) Octobre (OAG 2007) Observatoire de l’Action Gouvernemental (OAG). 2010. « Evaluation du processus de décentralisation au Burundi » Décembre Observatoire de l’Action Gouvernementale (O.A.G). 2010. « Evaluation du processus de décentralisa- tion au Burundi » Décembre Ould Kebd, Y., & Sindayihebura, C. 2011. Etat des lieux et besoins en renforcement des capacités des acteurs en décentralisation et développement communautaire. Ministère du développement communal. Gouvernement du Burundi. Octobre PEM Consult. 2007. Mission d’étude sur la problématique foncière et les solutions alternatives face aux défis de la réintégration et réinsertion des sinistrés au Burundi. UNOPS-PNUD-CNTB, Bu- jumbura. Perks, R. Forthcoming. The Inclusion of Artisanal and Small-scale mining in National Legislation: Case studies from Sub-Saharan Africa. World Bank Peters, P. 2004. “Inequality and social conflict over land in Africa” Journal of Agrarian Change 4 (3): 269-314 Programme d’Appui à la Bonne Gouvernance (PABG). Projet de décret portant règlement général de la comptabilité communale. Programme d’appui à la gestion foncière au Burundi. 2012. Document de capitalisation méthod- ologique des expériences pilotes en province Ngozi. April 2012 (SDC- Bujumbura) (overview approach and implementation, including role of local government, mandate, legal, human re- sources, and not that much on taxation) République du Burundi. 2005a. Loi no. 1/02 du 25 Janvier 2010 portant Révision de la Loi no. 1/016 du 20 Avril 2005 portant Organisation de l’Administration Communale. 116 République du Burundi. 2005b. Loi no. 1/016 du 20 Avril 2005 portant Organisation de l’Administration Communale. République du Burundi. 2007a. Lettre Nationale Politique sur la Décentralisation et le Développe- ment Communal. République du Burundi. 2007b. Programme d’actions Prioritaires de mise en œuvre du CSLP 2007- 2010. Mai 2007 République du Burundi. 2009a. Politique Nationale sur la Décentralisation et le Développement Com- munal République du Burundi. 2009b. Ministère de la Décentralisation et du Développement Communal. « Document de Politique Nationale de Décentralisation du Burundi » Mai République du Burundi. 2010a. « Le Senat et la politique de décentralisation au Burundi » Aout République du Burundi. 2010b. Loi No.1/02 du 25 Janvier 2010 Portant Révision de la loi No.1/016 du 20 Avril 2005 Portant Organisation de l’Administration Communale. (Loi Communale 2010) République du Burundi. 2011a. Décret no. 100/95 du 25 Mars 2011 portant Missions, organisation et fonctionnement du Ministère de l’Eau, de l’Environnement, de l’Aménagement du Territoire et de l’Urbanisme. République du Burundi. 2011b. Loi No 1/13 du 9 août 2011 portant révision du Code Foncier du Burundi. République of Burundi. 2011c. Second Strategic Framework for Growth and the Fight against Poverty (Cadre Stratégique de Croissance et de Lutte contre la Pauvreté II (CSLP-II), p. 16–17). République du Burundi. 2011d. Projet de Loi Portant Code Minier du Burundi République du Burundi. 2012a. Cadre Stratégique de Croissance et de Lutte contre la Pauvreté CSLP II. République du Burundi. 2013a. « Note sur le Secteur de la Bonne Gouvernance »Ministère a la présidence charge de la Bonne Gouvernance et de la Privatisation. Octobre République du Burundi. 2013b. Ministère de la Solidarité Nationale, des droits de la Personne Hu- maine et du Genre. Note analytique sur le secteur du relèvement communautaire. Octobre République du Burundi. Ministère du Développement Communal « Note sur la décentralisation au Burundi : La commune porte d’entrée du développement et de la démocratie » République du Burundi. 2013c. Projet de Loi relative au transfert de compétence de l’État aux com- munes. Gouvernment of Rwanda. 2005. Organic Law Establishing the Land Tenure System in Rwanda. (accessed 13 October 2010). Thinon, P. 2012. Rapport de la mission d’appui de Pascal Thinon au Programme Foncier de la Coo- pération Suisse au Burundi. Avril-Mai Thinon, P., Rochegude, A., & Hilhorst T. 2012. Rencontres foncières, Bujumbura 28-30 mars 2011. Coo- pération suisse (Bujumbura, Burundi) & Cardère éditeur (Lirac, France), 288p. Uhuru, F. 2004. Le Régime foncier dans la région des grands lacs : Le cas du Burundi Aujourd’hui. Jan- vier United States Agency for International Development. Burundi Policy Reform: Report for Validation of Land Code Revisions. United States Agency for International Development. 2006. Enabling an Inclusive Private Sector in Burundi: Opportunities and Constraints. May Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 117 United States Agency for International Development. 2007. Réformes de la Filière Café au Burundi: Perspectives d’Avenir pour la Participation, la Prospérité et la Paix. Février Uvin, P. 2008. Local Governance After War: Some reflections on donor behavior in Burundi. The Fletcher Journal of Human Security, volume xxiii Vaillancourt, F. 2013. La Gouvernance Locale en Afrique Francophone : sept notes sur des enjeux traités par le réseau FAGLAF. Project report Centre Interuniversitaire de recherche en analyse des organisations (CIRANO) Février (Vaillancourt 2013) Van Leeuwen, M. 2014. Land Disputes and Local Conflict Resolution Mechanisms in Burundi. 2nd Aegis Euroconference in African Studies, African Alternatives: Initiative and Creativity beyond Cur- rent Constraints, Leiden, The Netherlands. July Vandeginste, S. 2010. Power-sharing as a fragile safety valve in times of electoral turmoil: the costs and benefits of Burundi’s 2010 elections. The Journal of Modern African Studies, 49 pp.315-335 Vervisch, T. and Titeca, K. Bridging community associations in post-conflict Burundi: the difficult merg- ing of social capital endowments and new ‘institutional setting’. The Journal of Modern Afri- can Studies, 48, pp 485-51 World Bank. 2007. Burkina Faso: Making Decentralization work. September World Bank. 2009. Burundi Mining Sector: Mining as a potential source of diversification and growth. CEM Background Paper World Bank. 2011. Republic of Burundi - Country Economic Memorandum (CEM): the challenge of achieving stable and shared growth. Washington, DC: World Bank. http://documents.world- bank.org/curated/en/2011/03/14730161/republic-burundi-country-economic-memoran- dum-cem-challenge-achieving-stable-shared-growth World Bank. 2011. World Development Indicators (WDI) < http://data.worldbank.org/data-catalog/ world-development-indicators/wdi-2011> World Bank. 2012. Demand-side Governance and Strengthening of Civil Society. Unpublished draft note, Washington DC: World Bank Institute. World Bank. 2012. Participatory Budgeting (PB) in South Kivu, DRC: Support to decentralization and empowering citizens to participate in the budgetary process through ICT. Draft Project Concept Note. World Bank Institute World Bank. 2013. Social Accountability Flagship Study Concept Note Draft. Social Development De- partment May World Bank. Forthcoming 2014 SDV Flagship Report on Social Accountability, World Bank Social Ac- countability Department, Washington DC World Bank. Forthcoming 2014. Social Accountability Flagship Report, Social Development Depart- ment Bank. Forthcoming 2014. Madagascar Land Policy Reform: Perspectives and Prospects. Policy Re- search Report, World Bank, Washington, DC. World Bank/Food and Agriculture Organization (FAO). 2014. Analyse financiere de la gestion fon- ciere decentralisee. Study Report. ZOA International Relief. 2013. « Rapport foncier communal dans les communes de Mabanda et Vu- gizo dans la Province de Makamba du Burundi ». Novembre. 118 ANNEXES Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 119 Annex 1 Communal Merger: Supplementary Information Data A survey was carried out in August 2013 for the 16 communes listed in table A5.1. This is a reasoned sample with one commune in each province with diversity in population, area, topography, and urbanization. Availability of information (administrative and financial) was also taken into account. Table A1.1 Characteristics of Communes Surveyed in August 2013 Province Commune Population Area km² Bujumbura Mutimbuzi 69 037 181,1 Bubanza Gihanga 53 222 287,3 Cibitpoke Mabayi 66 531 347,5 Kayanza Matongo 70 652 167,8 Ngozi Ngozi 64 321 121,3 Kirundo Busoni 146 188 420,9 Muyinga Giteranyi 152 276 397,8 Cankuzo Mishiha 49 941 381,2 Ruyigi Butaganzwa 64 063 341,3 Rutana Rutana 52 445 255,3 Makamba Mabanda 44 792 294,9 Karuzi Buhiga 70 392 275,4 Gitega Gitega 156 096 315,4 Muramvya Rutegama 39 896 83,5 Mwaro Gisozi 27 953 118,7 Bururi Matana 44 059 207,4 Source: Rapport Définitif De Collecte Des Données En Rapport Avec Les Frais De Fonctionnement des Communes et la Fiscalité Communale par BUGAFIRO Pascal et MBARUBUKEYE Séverin et données du Recensement 2008. Le travail de cueillette de données a été réalisé de façon efficace par BUGAFIRO Pascal et MBARUBUKEYE Séverin. The main objectives of the survey were to collect information on the time and expenses spent to acquire a communal service or fulfill a communal legal requirement and on the operation costs of communes (fixed and variable). To achieve the first objective, a sample of 50 bills by commune was drawn and for each the usual time and distance between the communal office and the residence of the payer with it ascertained from communal officials. Figures A1.1 and A1.2 present Figure A1.1 Distance Walked, 16 Communes, information on distance and Burundi 2013 time walked to reach communal (Communes Ranked by Area) headquarters. The Annual # of Interactions 30.0 24.4 between Residents of 25.0 21.5 18.6 Communes and Communal 20.0 16.3 14.9 18.0 14.5 17.7 13.6 Headquarters 15.0 12.7 10.8 13.4 12.3 11.9 10.0 8.0 8.0 The average # of bills paid annually 5.0 per inhabitant is 12,5 per hundred 0.0 inhabitant with extreme values n i i a ut go M zi Ru a Bu a Gi ga ab a ga a M ayi te a Bu yi ni M a da z M goz am M ang ta teg an n Gi hih w included and 8,3 when we exclude the so bu n so hi m ton ta an nz ab ra at Gi g Bu Gi im is h te a Ru two extremes at both end s going from 15 to 11 values. Payment of bills is not the sole interaction with the commune. Source: Author’s compilation, using survey data. We thus assume 15 interactions par 100 inhabitants. Since the communal Figure A1.2 Time Walked, 16 Communes, population outside Bujumbura is Burundi, 2013 (Communes Ranked by Area) about 10, 5 million this yields 1 575 000 annual interactions. 300 The Cost in Time and FBu of 250 247 255 Interactions 200 168 158 177 152 145 142 150 135 127 134 131 119 For the 800 bills examined the average 108 83 85 return distance is 15 km and the 100 average time 150 minutes. We had in 50 work before this survey computed 0 a plausible distance of 14 km based ng i i Gi ga ab a ga a te a Bu yi a ut go M zi Ru a Bu a M a M yi ni da z M oz ta teg m an n M ang Gi hih w so bu n a so hi m ton ta an nz ab ga ra on the fact that119 a typical commune at Gi Bu Gi im is h te a Ru would be 225 km² and thus a square of 15 km X 15 km. If headquarters are Source : Authors’ compilation, using survey data. in the centre the maximum distance as the crow flies is 7,5 km and the average, assuming a uniform population density, is 3,75 km. Doubling this to account for the fact that humans do not fly yields 3, 75 * 2=7 km X 2= 14 km return trip. The time required seems a bit low given the topography of Burundi and typical human walking speed120 (5 km/h) but not implausible. Multiplying 150 minutes X 1 575 000 yields 236 250 000 119  La décentralisation financière au Burundi: quelques observations rapport préparé pour la Banque Mondiale Juin 2013. 120  See references cited in http://en.wikipedia.org/wiki/Walking. minutes or 3 937 500 hours or 491 818 days (8 hours day). We were informed that the best way of valuing time in the countryside was to use the wage of shepherds; this is FBu 1,000 per day + food and housing also worth FBu 1,000 per day. Thus rounding we estimate at FBu 1,000 million the value of time allocated to travelling to the communal headquarters. Establish the Plausible Reduced # of Communes We have no official information on this; we assume a reduction of two thirds to 40 communes. Calculate the Reduction in Communal Costs Publics Resulting from the Reduction in the # of Communes The cost per inhabitant of bonuses paid by communes to mayors (administrateurs communaux) and their two key advisors (conseillers techniques) is FBu 31 per habitant, that FBu 325 500 000 for the 116 communes. Reducing the # of communes to 40 reduces these costs by FBu 215 Million. We must add the salaries paid by the central government to this. The mean annual salary of an advisor is FBu 1.5 million; thus the drop in the number of communes would reduce central costs by 228 million. The mayors cost 1,6 million each or a drop of FBu 122 million.121 So the plausible minimal reduction in salary costs is 565 million. Other costs could also be reduced if the number of communes was decreased. A communal development plan (Plan Communal de Développement Communal or PCDC costs FBu 27 Million to prepare in 2012122 and is valid for five year; thus an annual cost of about FBu 6 Million. A reduction to 40 communes reduces these costs by 456 Million (76 X 6) if the cost of a PCDC does not vary with the area of the commune. This is unlikely given the need for consulting populations. If we assume a 50 percent increase in the unit cost (from 6 to 9 in annual costs) with a doubling of the size of a commune then we have a 336 million reduction. Adding this yields a reduction of 931 million. The other employees of communes cost FBu 285 per inhabitant that is 3 billion for the 116 communes. If a reduction in the # of communes to 40 means they were dismissed, this would reduce costs by FBu 2 billion. This is not compatible with maintaining old communal headquarters as service outlets. And this does not take into account increasing service provision at the zonal level The Impact on the User of Communal Services With 40 communes, their population is about 250 000 and their area 640 km² a square 25 km X 25 km and a maximum straight line distance of 12, 5 km. The average distance becomes 28 km return doubling the time cost of accessing communal headquarters. This increases private costs by FBu 1,000 million. This assumes no drop in the number of interactions residents-communes. If there is a drop the increase in costs is less but other consequences follow such a drop in the number of registered births or in taxes paid. 121  This number is obtained by combining information from the Ministère de l’Intérieur and from the pay bureau of the Ministère des Finances. 122  Data provided by the World Bank office, Bujumbura. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 123 Annex 2 List of People Consulted during the Preparation of the Study To ensure effective dialogue with government authorities, and maximize the joint ownership of the study, the Bank Task team carried out six missions to Burundi. Specifically, the consultation process with government and representatives of the donor community started with two subsequent missions in early 2013 (January and March 2013, respectively), which helped to define the scope of the study and leverage donor financing in the context of the “Programmatic PER.” A third mission helped to continue discussions on the draft PCN (May 2013) and was followed by two rounds of data collection (August–September and November–December 2013, respectively), coordinated. Preliminary drafts of the analysis were shared during one subsequent mission (January 2014), to give the authorities the opportunity to react and provide feedback to the team. Moreover, a validation mission was held in June 2014 to share the draft report with government authorities and the donor community where government officials, development partners and civil society organizations provided several comments and feedbacks. While a final mission to present the final draft report incorporating all comments received has been shared with the authorities in the September 2014 mission. The list of people met and interviewed during the missions and their respective institutional affiliation is provided below. Table A2.1 Government of Burundi Name TITLE Organization Office of the Presidency and Vices Presidencies Ntahiraja, Térence Conseiller Présidence de la République Nkinahamira, Pascasie Conseillère Juridique 1ere Vice-Présidence Nshimirimana, Térence Conseiller 2eme Vice-Présidence Ministry of Communal Developement Giteruzi, Dieudonné Ministre Ministère du Développement Communal Sindakira, Lazare Secrétaire Permanent Ministère du Développement Communal Ntihirageza, Jean-Marie Directeur Général de la Ministère du Développement Communal Décentralisation Manirakiza, Léopold Directeur General Ministère du Développement Communal Kwizera, Priscille Agent Ministère du Développement Communal Njebarikanuye, Spès- Directeur Général Fonds national d’investissement Caritas communal (FONIC) Bigirindavyi, Thomas Département Comptabilité Fonds national d’investissement communal (FONIC) 124 Ntahomvukiye, Mathias Chef Comptable Fonds national d’investissement communal (FONIC) Manirakiza, Eric Directeur Administratif et Fonds national d’investissement Financier communal (FONIC) Ntavyohanyuma, Letitia Chargée de la Communication Fonds national d’investissement communal (FONIC) Ministry of Intérieur Nduwimana, Edouard Ministre Ministère de l’Intérieur Mbonabuca, Thérence Directeur Général de Ministère de l’Intérieur l’Administration du Territoire Sindayihebura, Firmin Directeur du Département des Ministère de l’Intérieur Finances communales Melchior Gahungu Inspecteur des Finances Ministère de l’Intérieur Ministry of National Solidarity, Human Rights and Gender Munyaneza, Aline Cadre Ministère de la Solidarité Nationale Ministry of Water, Environnement, de l’Aménagement du Territoire et de l’Urbanisme (MEEATU) Macumi, Damien Coordinateur National- Unité de Gestion du Programme National MEEATU Foncière (PNF) Ndikumana, Ernest Directeur Général MEEATU Aménagement du Territoire Ngendabakana, Frédéric Directeur Bureau Centralisation MEEATU Géomatique Mboneko, Félix Directeur Cadastre MEEATU Ministry of Finance, Planning and Economique Development Bigirimana, Immaculée Secrétaire Permanent Ministère des Finances et de la Planification du Développement Economique Mushariste, Désiré Coordonateur des Aides Ministère des Finances et de la Planification du Développement Economique Kwizera, Christian Coordonateur Adjoint des Aides Ministère des Finances et de la Planification du Développement Economique Office Burundais des Recettes Holmes, Kieran Commissaire Général Office Burundais des Recettes Nzeyimana, Thacien Directeur Office Burundais des Recettes Ndayzeye, Avit Chef de Service Affaires Office Burundais des Recettes Juridiques Tesfaye Wondim Directeur Informatique Office Burundais des Recettes Ministry of Public Services Nderagakura, Janvier Conseiller du DCAPR Ministère de la Fonction Publique Batungwanayo, Chef de Service Centre Inter- Balthazar Ministériel des Ressources « Ministère de la Fonction Publique Database of Civil Servants » Hakizimana Potame Secrétaire Exécutif du CERAP Ministère de la Fonction Publique Esperance Habonimana Directeur Général Ministère de la Fonction Publique Parliament of Burundi Honorable Nkurunziza, Sénateur Sénat du Burundi Innocent Kigeme, Ariane Conseillère Décentralisation Sénat du Burundi Honorable Nkurunziza, Député Assemblé Nationale du Burundi Aimé Nkunzimana, Tharcisse Chef de Service Décentralisation Sénat du Burundi Civil Society Organization Nzohabonimana, Alexis Président Association Burundaise des Elus Locaux (ABELO) Bikebako, Pontien Président ONG Twitezimbere Nihangaza, Charles Expert en Finances Publiques Academia Banderembako, Paul Vice-Recteur Université du Burundi Table A2.2 Development Partners NAME TITLE Organisation Claudio Tognola Directeur de Coopération Coopération Suisse Pitteloud, Alansar Directrice de Coopération Coopération Suisse Elisabeth Ndikumasabo, Directrice du Programme Coopération Suisse Annonciata Johannsen, Achim Chef de Coopération Ambassade d’Allemagne Proehl, Alexander Coordinateur du Programme GIZ ADLP Programme ADLP GIZ Bafutwabo, Léopold Programme ADLP GIZ Nzobambona, Isidore Représentant Local KFW Nindorera, Yves Chargé de Programme Ambassade de Belgique Hermouet, Murielle Assistante Technique PADILPC Coopération Technique Belge (CTB) Corbel, Hervé Assistant Technique Coopération Technique Belge (CTB) Tilli Blomhammar, Chargée de Programme Délégation de l’Union Européenne Charlotte 126 Charvet, Eric Responsable Programme d'Appui à la Bonne Gouvernance Okry, Marc Assistant Technique Programme d'Appui à la Bonne Gouvernance Vrijlandt, Anthe Manager Trade Mark East Africa Table A2.3 World Bank Staff Alain-Désiré Karibwami Senior Health Specialist André Teyssier Senior Land Administration Specialist Aurelien Serge Beko Economist Mamadou Ndione Senior Country Economist Nadia Belhaj Hassine Belghith Senior Economist Rachel Bernice Perks Mining Specialist Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 127 Annex 3 Validation Workshop Summary Report A. Allocutions d’ouverture Le 2 juin 2014, la Banque mondiale en collaboration avec le Gouvernement du Burundi et ses partenaires techniques et financiers a olrganisé un atelier de validation du rapport préliminaire de l’étude sur la décentralisation fiscale et la gouvernance locale à l’hôtel Royal Palace de Bujumbura. La cérémonie d’ouverture de l’atelier a été marquée par la présence du Représentant Résident de la Banque mondiale, plusieurs membres du gouvernement, dont le Ministre du Développement Communal, un député de l’Assemblée Nationale, de nombreux hauts cadres de l’administration publique burundaise, ainsi que la société civile. L’objectif principal de l’atelier était d’ouvrir le débat autour des résultats clés de l’étude ainsi que de recevoir des commentaires et observations constructives du gouvernement et de la société civile. Compte tenu de l’importance des questions traitées dans ce rapport, le ministre du Développement Communal a félicité l’équipe projet d’avoir fourni des analyses très pertinentes et de répondre aux besoins du Gouvernement Burundais dans la mise en œuvre de son programme de décentralisation tel qu’énuméré dans l’Accord de paix d’Arusha. Un groupe technique composé de quatre membres du gouvernement a été mis en place afin de donner une position unie et claire des commentaires et observations du gouvernement. Les discussions de l’atelier ont été organisées autour des chapitres clés du rapport notamment la décentralisation budgétaire qui a traité les questions de fiscalité et des politiques publiques liées à la décentralisation ; la gouvernance foncière décentralisée qui traite des questions de gestion et sécurité foncière, ainsi que le chapitre sur la redevabilité sociale. Un panel de 5 experts dont un représentant du gouvernement (le Directeur Général au Ministère du Développement Communal) ; un représentant du Parlement (un Député a l’Assemblée Nationale) ; un expert en finance publique (Conseiller fiscal agréé) ; un expert en politique foncière (le Coordonnateur national de l’Unité de Coordination du Programme National Foncier, PNF) et un représentant de la société civile (Président de l’, ABELO). 128 B. Commentaires et observations du Gouvernement Observations/Commentaires Soulevés Réponses de l’équipe Observations générales sur le rapport préliminaire Les observations générales sur le rapport sont les suivantes : En tant que rapport préliminaire • Le titre du document de rapport semble un peu flou le titre n’est pas encore définitif et comparativement au contenu ; l’équipe trouvera un titre reflétant mieux le contenu du rapport. • Il n’y a pas un lien dans la succession des chapitres ; • Le document nécessite un toilettage au niveau de la forme car certaines affirmations et chiffres ne sont pas correcte et dans L’équipe a réalisé que la traduction tous les chapitres ; n’a pas pu prendre en compte les termes techniques du rapport. Un • Certains paragraphes sont à reformuler ; toilettage se fera sous peu. • Il y a certains termes ou vocabulaires qui ne sont pas utilisés au Burundi ; • Les analyses effectuées par cette étude de la Banque Mondiale sont en général objectives, bien fouillées reflétant des observations et recommandations pertinentes mais quelques conclusions ou recommandations sont à reformuler. Chapitre 1 : Comprendre le contexte de la décentralisation au Burundi Le chapitre titré « comprendre le contexte de la décentralisation Une revue sera effectuée afin au Burundi » a été bien analysé mais il faut ajouter les récentes de s’assurer que tous les textes évolutions du cadre réglementaires par rapport à la décentralisation. règlementaires y compris les Il faut signaler que le Gouvernement vient d’adopter certains textes nouveaux ne sont pas omis. réglementaires. Chapitre 2: La décentralisation budgétaire- contexte, questions liées aux politiques publiques et possibilités de changement De manière générale et du point de vue d’un praticien Merci. de ce secteur de décentralisation fiscale et budgétaire les analyses effectuées dans le chapitre 2 qui traite de la « Décentralisation budgétaire, contexte, questions liées aux politiques publiques et possibilités de changement » sont réputées objectives, bien fouillées reflétant des observations et recommandations pertinentes. L’état des lieux de la Décentralisation budgétaire actuelle est celui réellement vécu aujourd’hui par les communes. Il est effectivement caractérisé par de faibles transferts de ressources de l’Etat aux communes. Chapitre 3 : Gouvernance foncière décentralisée : défis et opportunités De façon globale, les conclusions et observations du chapitre qui Bien noté. traite la gouvernance foncière décentralisée sont correctes mais L’équipe de la Banque mondiale dans le développement de ce chapitre les données et les analyses effectuera ses changements. restent à améliorer. A titre illustratif : Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 129 • l’Unité de Coordination du Programme national foncier, re- sponsable de la décentralisation de l’administration foncière » : cette expression paraît équivoque. On ne décentralise pas l’administration foncière classique (DG cadastre et DG titres Fon- ciers notamment) mais on crée des services fonciers au niveau des communes qui ont des rapports avec cette administration foncière classique mais n’en sont pas des entités décentralisées. Les procédures utilisées, les documents délivrés, les acteurs in- tervenants ne sont pas les mêmes. Donc il faut plutôt dire que « l’Unité de Coordination responsable de l’accompagnement du Gouvernement dans la mise en œuvre de la gestion fon- cière décentralisée » ; • Le nombre de parcelles à enregistrer au niveau national est entre 8 et 9 millions au lieu de 3-4 millions, • Il semble important de rappeler que les Opérations Groupées de Reconnaissance (OGR) n’ont (actuellement) pour résultat direct et automatique « que » la réalisation des reconnaissances collinaires et la production du PV de reconnaissance et non pas le Certificat Foncier, qui n’est lui produit que sur demande et après paiement par l’usager. Chapitre 4 : Promouvoir la redevabilité sociale par la décentralisation : la voix des citoyens Les analyses effectuées par cette étude de la Banque Mondiale L’équipe de la Banque mondiale sont, sur ce chapitre, peu objectives car certaines conclusions sont remercie le gouvernement tirées sur base d’un cas isolé. Les conclusions de ce chapitre sont à d’avoir relevé ces commentaires améliorer notamment : et souhaite rappeler que les • L’exclusion sociale est un phénomène généralisé au Burundi et conclusions tirées de cette étude de nombreuses personnes interrogées ont donné la sensation sont basées sur les résultats des d’être automatiquement exclus de la vie citoyenne différents groupes focus organisés dans 16 communes provenant de • La participation citoyenne est généralement considérée 8 provinces burundaises. Toutes comme un moyen de recueillir les informations venant du les informations contenues dans sommet, des représentants du pouvoir local et non comme un ce chapitre ne relèvent pas du forum de décision et d’échange. point de vue des consultants • La participation peut être considérée comme un moyen internationaux de la Banque d’élargir le contrôle de l’Etat sur les citoyens et le territoire. mondiale mais plutôt du point de • Le Processus d’élaboration des PCDC a encouragé la préparation vue des plus de 300 résidents et à l’identification des priorités devant être transmises au élus locaux des communes ayant sommet mais peu ou aucune information n’est redescendue participés dans cette enquête. vers la base L’équipe souligne aussi que • Le processus d’élaboration des PCDC manque donc de les équipes de consultants qui redevabilité envers la base et ne permet aucune émancipation ont coordonnées cette enquête de la population étaient composées de citoyens Burundais et non des consultants • Avant les effets de la réforme récente du FONIC, les internationaux. représentants de l’Administration locale ont une perception très négative du FONIC, du fait du faible niveau de décision sur l’utilisation des fonds. 130 • La redevabilité sociale au Burundi demeure faible et les citoyens et les communautés ont peu d’occasions de s’engager dans le processus décisionnel et de mettre les autorités locales devant leur redevabilité • L’élaboration du plan communal de développement commu- nautaire, outil de la planification ne peut pas s’accommoder avec le budget communal. C. Contributions des panelistes et des participants Observations/Commentaires Commentaire Réponses par : Observations générales sur le rapport préliminaire L’assistant technique du projet Marc Okry L’assistant du Ministre du déve- « Gutwara Neza » indique que le rapport loppement communal a affirmé que aborde les sujets les plus importants le projet de loi sur le transfert des au Burundi dont la décentralisation, la Gutwara Neza compétences aux communes est en court gestion foncière, la redevabilité sociale et avance et le gouvernement est engagé etc. Les nouvelles modifications de la à l’adopter. loi communale n’ont pas toujours été Cependant M. le député a l’Assemblée adoptées malgré que le gouvernement nationale affirme que le 2 juin ait promis l’adoption de cette loi depuis (aujourd’hui) marque le début de la le début de cette année. Ceci montre un session de Juin à l’Assemblée nationale, manque de volonté politique de la part Honorable pourtant le projet de loi portant sur le du gouvernement face à la question de Aimé transfert des compétences ne figure décentralisation. Sur ce il faut que les Nkurunziza pas sur l’agenda des lois qui seront représentants du gouvernement clarifient débattues et analysées au cours de cette si la décentralisation est une priorité pour session. Par contre l’amendement à la loi le GdB. communale sera traité dans cette session afin que la loi soit en harmonie avec le Jean-Claude nouveau code électoral. Hatungimana Il est alors important que le projet de loi MDC portant sur le transfert des compétences aux communes soit adopté de préférence avant la loi communale afin de l’harmoniser avec cette dernière, ou soit que le projet de loi soit adopté ensemble avec la loi communale présentement au parlement. Par conséquence, le bureau du ministre du développement communal doit accélérer le processus d’approbation du projet de loi au gouvernement. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 131 Résumé executive Le Directeur général de l’Aménagement Ernest L’équipe de la Banque mondiale du Territoire pose la question de savoir Ndikumana/ a expliqué qu’elle ne fait pas une pourquoi à la page 7 du rapport, on fait Directeur recommandation du retour des « Bashin- une référence aux Batwas, ce qui n’a Général de gantahe » mais que l’on ne peut pas aucun lien au contenu du rapport. Il pense l’Aménage- ignorer le rôle important qu’ils jouent cependant que ce paragraphe devrait être ment du sur le terrain en matière de résolution supprimé. Territoire des conflits fonciers malgré qu’ils ne Sur la page 5, il se demande si la Banque soient plus reconnus par la loi. mondiale recommande le retour des « Bashingantahe » en tant qu’institution reconnue en ce qui concerne la résolution des conflits fonciers. Chapitre 1 : Comprendre le contexte de la décentralisation au Burundi Le Directeur général au Ministère du Jean Marie L’équipe de la Banque mondiale trouve Développement Communal a passé en Ntihirageza/ que ses observations et commentaires revue les avancées de la décentralisation MDC sont en accord avec les résultats du tout en signalant qu’il y a des efforts en rapport préliminaire. appui et des renforcements de capacités à fournir aux élus locaux. Il constata aussi quelques autres défis notamment : • Le manque de transferts de certaines compétences aux communes • L’amélioration du cadre institutionnel du processus de décentralisation afin que le gouvernement puisse mieux faire ses transferts de compétences aux communes • La communication a grande échelles du processus de la décentralisation • La coopération décentralisée afin de promouvoir l’intercommunalité entre communes Honorable Aimé Nkurunziza à son Aimé Merci tour a félicité l’équipe projet pour un Nkurunziza rapport de projet bien fait. Il ajoute que le rapport a bien répondu aux demandes du gouvernement. Néanmoins, il souligne quelques observations notamment : • La page 17 du rapport explique le processus de la décentralisation jusqu’en 2010 et aurait pu ajouter une analyse sur l’architecture insti- tutionnelle comme étant une entrée à la décentralisation. Il ajoute aussi que le rapport aurait pu aller de l’avant et soulever d’autres questions sur la compatibilité entre l’accord de paix d’Arusha et la décentralisation. 132 • Une comparaison de la décentralisation du Burundi avec les autres pays notamment le Rwanda n’est pas acceptable car ces deux pays malgré similaires ont de différentes réalités socio-politiques. Il a aussi ajouté que la décentralisation est un choix politique. Chapitre 2: La décentralisation budgétaire- contexte, questions liées aux politiques publiques et possibilités de changement L’Expert en finances publiques du panel Charles La faiblesse de la décentralisation est a fait les constats suivants sur la Nihangaza due notamment au fait que : décentralisation notamment : • Toutes les communes ne maitrisent • Le revenu faible des communes faute pas l’assiette fiscale. de mobilisation financière sauf les • Les exonérations du gouvernement communes qui ont un appui des PTFs qui ne sont pas bien définies. • L’analyse sur la mobilisation des • La question des passations des ressources propres des communes marchés devrait inclure la société civile et le • Le manque de communication budgé- secteur privé. taire • Il signe aussi l’anarchie observée dans • Le manque de contrôle au niveau certaines collines ou certains chefs de comptabilité du fait que le comptable collines jouent le rôle de percepteurs est recruté par l’administrateur. d’impôts • La nécessité d’ouvrir un débat avec le gouvernement sur le transfert des compétences et le renforcement des capacités. Le Représentant de l’OBR a porté une Thacien observation sur l’harmonisation des Nzeyimana collectes de recettes communales et aussi l’élimination de la double imposition au niveau communal afin que l’OBR puisse OBR avoir un partenariat gagnant-gagnant avec les communes. Le Secrétaire exécutif du CERA de la Potame L’équipe de la Banque mondiale a Fonction Publique indique son désaccord expliqué que le problème de regroupage avec les recommandations sur la réduction des communes ne résoudra pas les des communes. Il pense qu’un regroupage Hakizimana faiblesses existantes des communes. Le communal serait plus bénéfique aux problème le plus immédiat a résoudre est communes. CERA/Fonction le transfert des compétences, sans lequel Il ajoute que pourquoi ne pas changer Publique les communes ne pourront pas fournir la procédure actuelle de transferts de des services à leurs résidents. compétences aux communes suivi du Un représentant du gouvernement a financement de FONIC et suggère de indiqué que suite aux nouvelles réformes transférer les compétences aux communes, du FONIC, il y a eu notamment des une fois ces compétences acquises, le nouvelles avancées dont la contribution gouvernement devrait lier les exercices du FONIC a la fourniture des matériaux de performances des communes à leur pour la construction des écoles financement. fondamentales. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 133 La Directrice de la Coopération Suisse Coopération Le GdB est tout à fait d’accord avec a apprécié que cette étude sur la Suisse les observations sur l’absence de décentralisation ait soulevé les questions mécanisme de contrôle au niveau de la gestion et sécurité foncière ainsi que communal. Actuellement, il n’existe des recommandations pertinentes pour le aucun mécanisme de contrôle interne ni futur. externe pour règlementer les finances En outre elle pense que le rapport aurait communales et ceci est l’une des priorités pu aborder les questions de gestions du gouvernement. et contrôle des finances au niveau Un représentant du programme communal. Une recommandation axée “Gutwera Neza” a indiqué que le sur des mécanismes de contrôle aurait été programme avait créé un manuel souhaitable. de procédure pour la comptabilité En ce qui concerne la mobilisation des communale ainsi qu’un projet de loi qui ressources propres des communes elle reste à être discuté et approuvé par le se demande pourquoi l’équipe s’est-elle gouvernement. focalisée sur l’exploitation minière à grande L’équipe de la Banque a expliqué que échelle comme moyen de financement l’exploitation minière a été choisie du fait potentiel des communes et s’il y avait qu’elle représente l’un des principaux d’autres sources de financement. axes prioritaires dans la Stratégie d’aide- Elle a aussi soulevé les questions sur la pays ainsi que la CSLP II. Le choix sur la gestion faible des marchés publics. gestion foncière est basée sur le fait que la reforme foncière n’avance pas. En outre, puisant des expériences d’autres pays, le coût des certificats devient élevé quand il n’est pas un processus d’enregistrement complet donc ce rapport a trouvé qu’il existe une opportunité de combiner la gestion foncière de la décentralisation. Le gouvernement a répondu qu’un nouveau décret vient d’être approuvé pour mieux répondre à la problématique de la gestion des marchés Le représentant de l’ambassade d’Alle- Représentant de Un membre du gouvernement a magne a demandé une clarification du l’ambassade de expliqué que ce terme définit le niveau terme “basket fund” au gouvernement. l’Allemagne de fonds commun du gouvernement destiné à l’investissement communal. Ce fonds sera régi par des critères que le gouvernement devrait tenir en considération avant tout transfert au FONIC. Un représentant du gouvernement a MDC Bien noté. L’équipe de la Banque mondial suggéré que les textes du rapport en modifiera ce paragraphe pour mettre à page 40 mettent en exergue les nouvelles jour les réformes du FONIC. réformes du FONIC et ses nouveaux mécanismes de transferts. 134 Chapitre 3 : Gouvernance foncière décentralisée : défis et opportunités L’expert en politique foncière du panel Damien a noté qu’il y a eu des confusions dans Macumi/ la traduction du rapport de l’anglais au Coordonnateur français notamment la gestion foncière National- Unité qui est un système classique et la gestion de Coordination foncière décentralisée qui sera un nouveau Nationale de système. Il ajoute que le gouvernement la Politique a commencé quelques petites réformes Foncière foncières notamment la lettre de la politique foncière de 2010. Malgré qu’il ne soit pas un décret, il contient la vision du gouvernement sur ce secteur. Cependant il a constaté aussi qu’il manque un cadre institutionnel foncier ainsi qu’un programme national foncier, ce qui relève la question d’appropriation du gouvernement vis-à-vis des questions de la gestion foncière. Le représentant de l’ambassade Représentant de L’expert en Politique Foncière du pan- d’Allemagne a demandé une clarification l’ambassade de el souligne que la CNTB est chargée de sur le rôle de la CNTB par rapport à la l’Allemagne résoudre les conflits liés au retour des gestion foncière au niveau communal. rapatriés tandis que la gestion foncière communale contribue sur le terrain à la reconnaissance et clarté des frontières des terres ; un rôle souvent joué par les “Bashingantahe”. Il a aussi souligné que la CNTB et le service foncier communal travaillent en synergie. Chapitre 4 : Promouvoir la redevabilité sociale par la décentralisation : la voix des citoyens Le représentant de la société civile du Alexis L’équipe de la Banque mondiale panel a trouvé ce rapport préliminaire Nzohabonimana explique qu’il est important de rappeler très enrichissant et la définition de la à l’audience la méthodologie suivie dans redevabilité sociale compréhensive. Par ce rapport issu d’une étude et donc sa contre il trouve le rapport un peu trop nature académique et théorique. académique et théorique. Association L’équipe a ajouté que le chapitre sur L’étude sur les 2 communes (Rutegama Burundaise des la redevabilité sociale est basé sur et Makebuko), 2 communes de la même Elus Locaux les résultats des groupes focus de 16 province partageant des traits similaires (ABELO) communes provenant de 8 provinces. ont bien éclairé la situation actuelle au niveau communal. Cependant, il aurait été souhaitable que le rapport puisse examiner en profondeur la différence entre la performance de ces 2 communes. Republic of Burundi Fiscal Decentralization and Local Governance: Managing Trade-Offs to Promote Sustainable Reforms 135 Il souligne alors que l’administrateur Toutes les informations contenues dans communal de Rutegama est un ingénieur ce chapitre ne relèvent pas du point de agronome alors que l’administrateur vue des consultants internationaux de communal de Makebuko n’a pas de la Banque mondiale mais plutôt du point diplôme universitaire d’où le grand de vue des plus de 300 résidents et élus écart dans leurs performances en ce qui locaux des communes ayant participés concerne la collecte des recettes ainsi que dans cette enquête. Elle souligne aussi l’application de certains mécanismes de que les équipes qui ont coordonnées redevabilité sociale. cette enquête étaient composées de citoyens Burundais et non des consultants internationaux. La recommandation du rapport devrait Le directeur du département des mettre l’accent sur la nécessité des finances communales du ministère autorités gouvernementales d’éviter le de l’Intérieur est d’accord sur militantisme en ce qui concerne le choix la problématique soulevée par le des candidats sur la liste bloquée de représentant de la société civile et candidature au niveau communal. Plutôt, il souligne la nécessité du gouvernement faudra sensibiliser les autorités centrales d’organiser un atelier de sensibiliation et locales afin de mettre les candidats sur l’importance d’ériger des candidats les plus qualifiés, compétents et mieux compétents pour la gouvernance de la éduqués sur les listes électorales. commune. 136 The Government’s decentralization program is an integral part of the Burundi Strategic Framework for Poverty Reduction II. The present study, carried out by the World Bank in collaboration with the Government of Burundi and other development partners, has the objective of providing rigorous analysis to allow the formulation of evidence-based policies and decision-making in this policy area. The key messages from this study are as follows: First, it is imperative that the general responsibilities devolved to communes be clarified. Second, given the limited fiscal space and fragility of the Burundi economy, the central government has little room to improve the budgetary means of Communes. For this reason, efforts are needed to strengthen both the local tax administration system and the link between improved tax collection and better services to citizens. Related to this is the need to improve social accountability and to promote the creation of more transparent and systematic mechanisms for tax collection and for the preparation or implementation of communal development plans. Finally, improving land management requires the establishment of sustainable systems for issuing titles throughout the country and calls for substantial financial support from the central government, coupled with the creation of a central and integrated coordination mechanism. While commending the efforts of the national authorities, we hope that this report will provide a useful contribution to help identify priority reforms needed to advance the decentralization and local governance agenda in Burundi. Rachidi B. Radji, Country Manager For more information, please visit: www.worldbank.org/en/country/burundi Also join the discussion on: http://www.facebook.com/worldbankafrica http://www.twitter.com/worldbankafrica http://www.youtube.com/worldbank