Report No. 42160-BI Republic of Burundi Public Expenditure Management and Financial Accountability Review (PEMFAR) Improving Allocative Efficiency and Governance of Public Expenditure and Investing in Public Capital to Accelerate Growth and Reduce Poverty June 2008 Poverty Reduction and Economic Management 3 Africa Region Document of the World Bank, Co-produced with the Government of Burundi Defense o f Democracy) CNI Centre National Znformatique (National Information Technology Center) COMESA Common Market for Eastern and Southern Africa CPI Consumer Price Index CPIP Country Procurement Issue Paper CWIQ Core Welfare Indicators Questionnaire DAC Development Assistance Committee o f the OECD DAF Direction Administrative et Financike (Administrative and Financial Directorate) DBC Direction Gtntrale du Budget et Control (General Directorate o f Budget and Control) DC Direction de Comptabilitd (Directorate o f Accounting) DD Direction de Douanes (Customs Directorate) DDR Demobilization, Disarmament, and Reintegration DFID UK Government Department for International Development DGBCP Direction Gtndrale du Budget et de la Comptabilitd Publique (General Directorate of Budget and Public Accounting) DGHER Direction Gtndrale de I’Hydrauliques et des Energies Rurales (General Directorate o f Rural Water and Energy) DGMP Direction Gtndrale des Marchds Publiques (General Directorate of Public Procurement) DGP Direction Gdndrale de Planz@ation (General Directorate o f Planning) DGR Direction Gdndrale de Recettes (General Directorate o f Public Revenue) DOD Debt Outstanding Disbursed DSA Debt Sustainability Analysis DT Direction de la Trdsorerie (Treasury Directorate) EAC East African Community EFA Education for All EMSP Economic Management Support Project (Projet d ’Appui Ci la Gestion Economique, PAGE) ERSG Economic Reform Support Grant EU European Union F.O.B. Free on board (import price less the costs o f transportation and insurance charges) FAD Fiscal Affairs Department o f the IMF FBu Burundi Franc FNL-PA L I P E H U T U Forces Nationales de Libdration-Partipour la Libdration du Peuple Hutu (National Liberation Forces-Party for the Liberation o f the Hutu People) FNR Fonds National Routier (National Road Fund) FY Fiscal Year GDP Gross Domestic Product GER Gross Enrollment Rate GGTE Groupe de Gestion de la Trdsorerie de I’Etat (State Cash Management Group) GNI Gross National Income GNP Gross National Product GSFM Government Statistical and Financial Manual GWh Giga-Watthour HDI Human Development Index HIPC Heavily Indebted Poor Countries HIV/AIDS Human Immunodeficiency Virus/ Acquired Immunodeficiency Syndrome HRMIS Human Resources Management Information System IDA International Development Association IDF Inspection des Finances (Sectoral inspection body) IFC International Finance Corporation IFMIS Integrated Financial Management Information System IGE Inspection Gtntrale de I ’Etat (State Inspectorate General) IGF Inspection Gtntrale des Finances (Finance Inspectorate General) IMF International Monetary Fund INTOSAI International Organization o f Supreme Audit Institutions I-PRSP Interim Poverty Reduction Strategy Paper ISN Interim Strategy Note ISTEEBU Znstitut de Statistique et des Etudes Economiques du Burundi (Institute o f Statistics and Economic Studies o f Burundi) IT Information Technology JAVC Joint Agreement o f Verification and Control Mechanism JICA Japan International Cooperation Agency LCU Local Currency Unit LIC Low-Income Country MBGOV Ministke de la Bonne Gouvernance, de la Privatisation, de I’lnspection Gtntrale de 1’Etat et de I ’Administration Locale (Ministry o f Good Governance, Privatization and Government Oversight) MDG MillenniumDevelopment Goal MDRI Multilateral Debt Relief Initiative MDRP Multi-Country Demobilization and Reinsertion Program MDTF M u l t i Donor Trust Fund MEFDC Ministry o f Economy, Finance and Development Cooperation MEM Ministry of Water, Energy and Mines MF Minister o f Finance MFP M i n i s t k e de la Fonction Publique (Ministry o f Public Service) MICS Multisectoral Indicators Cluster Survey MINCO Ministry o f Commerce, Industry, and Tourism MINEDUC Ministry o f National Education and Culture MIREG Ministry o f Regional Integration and E A C Affairs MLA Modem Language Association MOT Ministry o f Transport, Post and Telecommunication MPW Ministry o f Public Works and Equipment MSP Ministry o f Public Health MTEF Medium-Term Expenditure Framework NEPAD N e w Partnership for Africa’s Development NGO Non-Governmental Organization NPF National Police Force NPV N e t Present Value O&M Operation and maintenance OCIBU Ofice de Caft du Burundi (Coffee Board o f Burundi) ODA Official Development Assistance OdR Ofice des Routes (Roads Authority ) OECD Organization for Economic Cooperation and Development ONPR Office o f Pensions and Professional Risks (for civil servants) OTBU Ordonnateur Trtsorier du Burundi (Burundi Payment orders Head) PAGE Projet d i i p p u i ci la Gestion Economique (Economic Reform Support Project) PAP Programme d 'Actions Prioritaires (Priority Action Plan) PBCE Plan Budgttaire et Comptable de 1'Etat (Charter o f Accounts and Budget) PEFA Public Expenditure Financial Accountability PEM Public Expenditure Management PEMFAR Public Expenditure Management and Financial Accountability Report PER Public Expenditure Review PETS Public Expenditure Tracking Survey PFM Public Finance Management PHC Primary Health Care PIP Public Investment Program PIU Project Implementation Units PLF Projet de Loi de Finance (Draft Budget Law) PNCE Plan National de Comptabilitt? de 1'Etat (National Plan for Public Accounting) PPBS U S Planning, Programming and Budgeting System PPG Public and Publicly Guaranteed PUP Procurement Reform Action Plan PRGF Poverty Reduction and Growth Facility PRSP Poverty Reduction Strategy Paper PSC Public Service Commission QUIBB Questionnaire des Indicateurs de Base du Bien-&re (Core Welfare Indicators Questionnaire, CWIQ) REGIDESO Rtgie de Production et de Distribution d'Eau et d'Electricitt (State Water and Power Utility) RMSM-X Revised Minimum Standard Model-Extended SCEP Sewice Chargt des En treprises Publiques (Office in Charge o f Public Enterprises) SES Senior Executive Service SIGEFI Systkme de Gestion des Finances Publiques (Integrated Financial Management Information System) SIP Socittt Immobilikre Publique (Public Real Estate Company) SMEs Small and medium enterprises SOEs State-Owned Enterprises SSA Sub-Saharan Africa STA Single Treasury Account STP Secrttariat Technique Permanent (Permanent Technical Secretariat) SWAP Sector-wide Approach TA Technical Assistance TOFE Tableau des Optrations Financikres de 1'Etat (Table o f Government's Financial Operations) TOR Terms o f Reference TSA Treasury Single Account UN United Nations UNDP United Nations Development Program USAID United States Agency for Internationaldevelopment VAT Valued Added Tax VMOP Vice Ministry o f Planning VMPW Vice Ministry o f Public Work WB World Bank WBI World Bank Institute WFP World Food Program WHO World Health Organization ZEP Banque de la Zone d’Echanges Prkfdrentiels (Eastern and Southern African Trade and Development Bank) IDA Vice President: Obiageli Ezekwesili Country Director: John M. McIntire Sector Director: Sudhir Shetty Sector Manager: Yvonne M. Tsikata Jan Walliser Lead Economist: Hinh T. Dinh Task Manager: Jean-Pascal N. Nganou ACKNOWLEDGMENTS This Public Expenditure Management and Financial Accountability Review (PEMFAR) i s a joint World Bank and Government o f Burundi report. The PEMFAR i s a product o f well synchronized teamwork and numerous interactions, and dialogue between Bank staff and many people and institutions in Burundi: the Government o f Burundi, the private sector, Non-Government Organizations (NGOs), academia, and bilateral and multilateral donors. Two key events were organized to launch the PEMFAR. Following the discussion on the PEMFAR concept note in M a y 2007, the team organized a brainstorming session with the main donors involved in Burundi’s development to identify key issues and themes to be developed in the PEMFAR. In June 2007, the PEMFAR team organized jointly with the Government o f Burundi a mission to agree on the scope and methodology o f the PEMFAR as well as to deepen the understanding o f issues to be covered in the report. The outcomes o f these two important events helped to lay the foundations and the strategic directions o f the PEMFAR report. I n September 2008, the report was disseminated to representatives o f relevant line Ministries, donor agencies and other stakeholders and was discussed during the course o f a workshop. The team thanks the participants o f this workshop, whose recommendations were used to further focus and strengthen the PEMFAR report. The team would like to thank H e r Excellency Madame Clotilde Nizigama (Minister o f Economy, Finance, and Development Cooperation, Burundi) and the Government team for the excellent collaboration throughout the preparation o f the PEMFAR report and the valuable advice and inputs provided to complete the report. The report was completed under the overall leadership o f Jan Walliser (Sector Manager, AFTP3). The team received excellent guidance from Hinh Dinh (Lead Economist, AFTP3), who provided invaluable inputs, quality control o f the report, and suggestions throughout the preparation o f the report. Conceptualization o f the work benefited from the guidance o f Yvonne Tsikata (Country Director, LCC3C, and former Sector Manager, AFTP3). The team also received unwaivering support and leadership from John Murray McIntire (Country Director, AFCEl), Kathryn A. Funk (Country Program Coordinator, AFCE1) and Alassane Sow (Country Manager, Burundi). The main author and task team leader o f the report i s Jean-Pascal N. Nganou (Country Economist for Burundi, AFTP3). The PEMFAR core team includes Eric Mabushi (Economist, AFTP3), Carolina Monsalve (Economist, ECSSD), Daniel Albert0 Benitez (Sr. Economist, FEU), Renaud Seligmann (Sr. Financial Management Specialist, AFTFM), Prosper Nindorera (Sr. Procurement Specialist, AFTPC), Maria Eugenia Bonilla- Chacin (Sr. Economist, AFTH3), Harry Garnett (Consultant), Xavier de la Renaudibre (Consultant), MClanie Xuereb (Consultant), and Jean-Pierre L e Bouder (Consultant). Valuable contributions were also provided by Quentin Wodon (Lead Poverty Specialist, AFTPM), Heather Ann Milkiewicz (JPA, AFTP3), Deo-Marcel Niyungeko (Municipal Engineer, AFTU2), Ephraim Kebede (Consultant, AFTP3), and Janine Mans (PA, AFTP3). Additionally, the team appreciates the useful comments o f Michael Stevens (Consultant, QAG) and Lev Freinkman (Lead Economist, AFTP3) on an earlier draft o f the document. The Bank team worked closely with a joint-committee o f various ministries, and the counterpart team included the following members: Donatien Bwabo, Boaz Nimpe, Tharcisse Yamuremye, SClCus Nezenve, Gabin Murekambanze, Claudine Hakizimana, Pasteur Mpawenimana, Ildephonse Bizindavyi, Francois Ngurinzira, Come Nijimbere, Keza Remy, Potame Hakizimana, BCatrice Nanganyinka, Gratien Ninteretse, Pierre Niyubahwe, Sosthene Hicuburundi, LConce Sinzinkayo, Celsius Barahinduka, Louis Nduwayezu, Philibert Nduwayezu, Goreth Ndayininahaza, Gloriose Ndayisenga, Godefroid Nyawakira, AdelaYde Nshimirimana, Georgette Nduwimana, RCvCrien Simbarakiye, SCraphine Rucakumugufi, Daniel Niyonkuru, Lieve Musaniwabo, RCnovat Gatabazi, Aloys Sahiri, Apollinaire Nkeshimana, J.M. Alphonsine Yakire, Graciose Ntibanyunva, Damien Mapfa, Thkrbse Muyuku, Marie Nizigiyimana, Balthazar Batungwanayo, and Pascal Sebushahu. We would also like to thank Sandrine Pierloz from the European Commission for her useful input. The report also benefited immensely from discussions with the peer reviewers Robert R. Taliercio (Senior Country Economist, EASPR), and Abebe Adugna (Senior Economist, FRM). Comments and suggestions from outside the Bank are also gratefully acknowledged, in particular from Paul H. Mathieu, and Joannes Mongardini (former Mission Chiefs, AFR, IMF), as well as Jean-Luc H e l i s (Advisor, Fiscal Affairs Department, IMF). The team i s particularly grateful to Professor Pierre-Richard AgCnor (Hallsworth Professor o f International Macroeconomics and Development Economics, University o f Manchester, and co-Director, Center for Growth and Business Cycle Research, England) for his technical guidance on macroeconomic modeling for Burundi. The team would also like to extend its gratitude to Mariama Daifour Bfi (Program Assistant AFTP3), who processed the document, and Aurore Simbananiye (Program Assistant, AFMBI) for logistics support. Resources received from the Belgian Trust Fund for Poverty Reduction are also greatly appreciated. Further, we acknowledge the financial support from the Japanese PHRD grant for the funding o f a few background studies contributing to the report as well as for dissemination events organized for the validation o f the report in Bujumbura. Republic of Burundi Public Expenditure Management and Financial Accountability Review (PEMFAR) Table o f Contents EXECUTIVE SUMMARY ................................................................................................................................ i ............................................................................................................................................ INTRODUCTION 1 A. Political and Socio-Economic Situation ............................................................................................ 1 B. Recent Reform Achievements ........................................................................................................... 2 C. Rationale for PEMFAR ..................................................................................................................... 4 D . Objectives. Scope and Structure ........................................................................................................ 5 CHAPTER1 . MACROECONOMIC CONTEXT ........................................................................................... 7 A. Structure of Burundi’s Economy and Public Finances ...................................................................... 7 B. Public Finance Structure .................................................................................................................. 10 C. Macroeconomic Outlook and Financing Requirements. 2007-2010 ............................................... 17 D. Solvency and Fiscal Sustainability .................................................................................................. 21 CHAPTER 2 . COMPOSITION OF PUBLIC EXPENDITURE AND KEY SOURCE OF FISCAL PRESSURE ......24 A. Analysis of Public Expenditure by Economic and Functional Classifications ............................... -24 B. Factors Determining the Structure o f Public Expenditure............................................................... 36 C. wage bill - key source of fiscal pressure ......................................................................................... 38 D . Impact o f Public Expenditure on Growth and Poverty - Modeling Results .................................. 43 CHAPTER 3 . FISCAL SPACE ANALYSlS AND SECTORAL EXPENDITURE REVIEW ............ 48 A. Creating Fiscal Space I n Burundi .................................................................................................... 48 B. Education ......................................................................................................................................... 53 C . Health............................................................................................................................................... 62 D. Infrastructure.................................................................................................................................... 69 CHAPTER 4 . REVIEW OF PROGRESS IN PUBLIC EXPENDITURE AND BUDGET REFORMS ..................-74 A. Overview O f Public Expenditure Management Reforms ................................................................ 74 B . Reform O f The Budget Framework ................................................................................................. 78 C. Budget Coverage And Preparation .................................................................................................. -84 D. Budget Execution............................................................................................................................. 90 E. The Interim Integrated Financial Management System (Sigefi) ..................................................... 93 F. Public Accounting ........................................................................................................................... 97 G. Cash Planning And Management ..................................................................................................... 99 H. Audit And Oversight...................................................................................................................... 101 I. Procurement Reforms .................................................................................................................... 106 J. H o w Can The Mtef Be Introduced I nBurundi? ............................................................................ 108 CHAPTER 5 . INCENTIVES AND PROFESSIONALIZATION IN PUBLIC ADMINISTRATION .......... 113 A . Incentives ...................................................................................................................................... 113 B. Professionalization........................................................................................................................ 117 C . Reform Priorities ........................................................................................................................... 120 BIBLIOGRAPHY ................................................................................................................................... 167 List of Tables Table 1.1: Poverty Incidence ....................................................................................................................... 9 Table 1.2: Social Indicators......................................................................................................................... 9 Table 1.3 : Composition o f Government Revenue, 200 1-2006 (in percentage o f GDP) .......................... -10 Table 1.4: Public Expenditure and Net Lending- A Cross Country Comparison (in percentage o f GDP), 2006 estimates .................................................................................................................................... 12 Table 1.5: Composition o f Government Expenditure (in percent o f GDP) ............................................... 12 Table 1.6: Central Government Fiscal Deficit, 2001-2006 ....................................................................... 13 Table 1.7: Main Assumptions for Medium-Term Macroeconomic Projections (in units indicated) ........18 Table 1.8: Balance o f Payments Projections ............................................................................................. 19 Table 1.9: Government Finance Statistics Projections .............................................................................. 20 Table 1-10:National Accounts Projections ................................................................................................. 21 Table 1.11: Public Sector Solvency and Sustainability in Burundi............................................................. 22 Table 2.1 : Economic Distribution o f Public Expenditures, 200 1-2006 .................................................... 25 Table 2.2: Recurrent spending categories in total recurrent, 2001-2006 (in percent) ............................... 25 Table 2.3: Budget Execution Rates by Category o f Domestically Financed Public Expenditures, ..........29 Table 2.4: Functional Distribution o f Public Expenditures, 2001-2006 .................................................... 31 Table 2.5: Comparison o f Spending on Priority Sectors ........................................................................... 33 Table 2.6: Projected Domestic Expenditure by Functional Classification ................................................ 35 Table 2.7: Donor-Financing o f Total Public Expenditure ......................................................................... 37 Table 2.8: Comparative Data on Wage Bill .............................................................................................. 38 Table 2.9: Trends in the Wage Bill ........................................................................................................... 39 Table 2.10: Comparing Staffing and Wage Bill in the Civil Service, 2004 and 2007 (in units indicated) -39 Table 2.1 1: Public Service Staffing and Average Monthly Base Salaries, M a y 2007 (in FBu, unless otherwise indicated) ............................................................................................................................ 40 Table 2.12: Wage Bill and the Budget ........................................................................................................ 42 Table 3.1 : Projections o f Government Financing...................................................................................... 49 Table 3.2: Government Financing, 2001-2006 .......................................................................................... 50 Table 3.3 : Summary o f Baseline Debt ..................................................................................................... -50 Table 3.4: Tax Revenue/GDP Ratios in Selected SSA Countries, 2006 ................................................... 51 Table 3.5: Gross Enrollment Ratios from 2000-01 to 2006-07 ................................................................. 54 Table 3.6: Gross and N e t Enrollment Ratios in Burundi........................................................................... 54 Table 3.7: Public Spending on Education ................................................................................................. 55 Table 3.8: Structure o f Current Expenditure in Education - Functional Classification ............................ 58 - Table 3.9: Structure o f Education Spending Economic Classification ................................................... 59 Table 3.10: Rates o f Execution for Public Expenditures in Education ....................................................... 60 Table 3.1 1: Health Outcome Indicators in Burundi and other African Countries ....................................... 63 Table 3.12: Salaries o f Health Care Personnel in Multiples o f GDP per Capita ......................................... 63 Table 3.13: Public Spending on Health ....................................................................................................... 64 Table 3.14: Public Spending on Roads, Power and Water Supply.............................................................. 69 Table 3.15: Density ofthe Road Network in Burundi and other East Africa Countries ............................. 70 Table 3.16: Donor Contributions to the Road Sector in Burundi and other Low-Income Countries ..........71 Table 3.17: Power and Water Supply in Burundi and Four Other East African Countries ......................... 72 Table 3.18: Donor Contributions to Power, Water Supply and Sanitation (in shares o f GDP) ..................72 Table 4.1: Broad Phases o f the Development o f an MTEF Process....................................................... -111 Table 5.1: Average Salary by Category, M a y 2007 ................................................................................ 114 List of Charts Chart 1.1: Real Output Growth. 200 1-06 (in percent) ................................................................................. 8 Chart 1.2: Burundi GNI Per Capita. 1962-2005 (in current U S dollars. Atlas Method) ............................. 8 Chart 1.3: Real Total Public Spending. 2001-2006 (in FBu billions. unless otherwise indicated) ..........12 Chart 1.4: Balance o f Payments. 2001-2006 ............................................................................................. 14 Chart 1.5: Evolution o f N e t ODA to Burundi Chart 1.6: N e t ODA to Burundi, 1982-2005 .......15 Chart 1.7: External Debt (in percent o f GNI) ............................................................................................ 16 Chart 2.1 : Broad Structure o f Domestically Financed Public Expenditures (average over 2005-2006) ...30 Chart 2.2: A Cross-Country Comparison o f Military Expenditure (in percent o f GDP, 2006) ................32 Chart 2.3: Service Staff by Category. M a y 2007 ...................................................................................... 41 List of Boxes Box 1.1: Tax Regime in Burundi .............................................................................................................. 11 Box 1.2: Debt Sustainability Analysis ...................................................................................................... 23 B o x 2.1: Use o f HIPC Resources .............................................................................................................. 35 Box 2.2: Impact o f Public Expenditures on Poverty ................................................................................. 44 Box 3.1: Key Sector Development, Budget and Financial Management Issues ....................................... 53 Box 3.2: Main Recommendations in the Education Sector ....................................................................... 61 Box 3.3: Key Sector Development, Budget and Financial Management Issues ...................................... -62 B o x 3.4: M a i n Recommandations ............................................................................................................. 68 Box 3.5: Key Sector Development Issues ................................................................................................. 69 Box 3.6: Main Recommendations on the Infrastructure sector ................................................................. 73 Box 4.1: Burundi - Main Texts Organizing the Budget Legal Framework .............................................. 80 Box 4.2: Budget Legal Framework-Main Recommendations ................................................................ 83 B o x 4.3: The Full Poverty Reduction Strategy Paper - PRSP, 2006-09 ................................................... 86 Box 4.4: Recommendations on the Coverage and the Presentation o f the Budget L a w .......................... -86 Box 4.5: Budget Preparation - Main Recommendations .......................................................................... 89 Box 4.6: Government Bank Accounts ....................................................................................................... 91 B o x 4.7: Complex Budget Execution Process........................................................................................... 92 B o x 4.8: Budget Execution - M a i n Recommendations ............................................................................ 93 Box 4.9: Interim Integrated Financial Management System - Main Recommendations .......................... 95 Box 4.8: The Key Principles o f the PBCE ................................................................................................ 97 Box 4.1 1: Public Accounting . M a i n Recommendations .......................................................................... 98 Box 4.12: Cash Planning and Management - Main Recommendations .................................................. 100 Box 4. 13 : Internal Audit - M a i n Recommendations................................................................................ 103 Box 4.14: Main Findings o f the CdC’s report on FY2005 Budget Implementation Report and Loi de dglement .......................................................................................................................................... 105 Box 4.15: External Oversight - Main Recommendations ........................................................................ 105 Box 4.16: Procurement Reforms - M a i n Recommendations ................................................................... 108 Box 5.1 : Grading Structure in the Civil Service ..................................................................................... 119 Box 5.2: Medium-Term Priorities .......................................................................................................... 122 List o f Annexes Annex 1: Proposed Schedule for the Preparation o f the Budget Law .................................................. 124 Annex 2: Cash Management and MTEF Proposals ................................................................................ 125 Annex 3: Wage bill and Salaries ............................................................................................................. 130 Annex 4: Macroeconomic and Government Financial Operation Tables ............................................... 134 Annex 5: Public Expenditure Statistics ................................................................................................... 138 Annex 6: Summary l i s t o f equations for the small model ....................................................................... 148 Annex 7 : Structure o f a Small Model linking Public Expenditure, Growth, and Poverty ...................... 149 Annex 8 : Table 1: Baseline Scenario o f Simulations (2006-20 15) .......................................................... 150 Annex 9: Matrix o f Policy Actions ......................................................................................................... 156 SUMMARY EXECUTIVE 1. This joint World Bank-Government of Burundi report i s the first public expenditure review undertaken in Burundi since 1992, a year before the outbreak of the 1993-2000 civil war. It aims to support the Government’s commitment to transition from a post-conflict economy toward an economic development paradigm, through rationalization o f the public finance management (PFM), and to find fiscal space to finance expenditure programs in key economic and social sectors. The need for P F M reform has become more pertinent as a complement to the Government’s economic and social objectives, which were laid out in the Interim Poverty Reduction Strategy Paper (I-PRSP) in 2003 and the full PRSP in 2006. Those objectives are: (i) peace, security, and good governance; (ii)equitable and sustainable growth to reduce poverty; (iii) development o f human capital (health and education); and (iv) the fight against HIV/AIDS. 2. This report addresses key issues on macroeconomic stability, efficient allocation of public expenditures, public sector staffing, and public finance management. The fiscal sustainability analysis carried out for this report indicates that serious adjustments will be needed to improve Burundi’s fiscal stance; reforms that are complicated by the challenges o f rebuilding the country’s social and physical infrastructure after years o f conflict. While some savings may come from reductions in defense expenditure, this savings will not be sufficient since expenditures for public security (police) are expected to increase. Consequently, external financing (preferably grants and highly concessional loans) will be needed, at least in the short and medium terms, to finance the fiscal deficit. However, given the volatility o f aid to Burundi over recent years and the unpredictable nature o f external capital inflows, the authorities should, in the long term, find ways to rely less on foreign financing, particularly through: (a) improving the tax administration; and (b) strengthening the effective use o f public resources on development goals. Reforming the management o f budgetary resources to improve accountability and transparency i s also crucial to improve fiscal sustainability without deviating from the Government’s overall objective o f implementing i t s national poverty reduction strategy (the PRSP). In order for the reforms to work, capacity development needs to be addressed through training, incentives, and professionalization o f the public administration. 3. The key assumption underlying this report i s that good management of public expenditures has an important role to play in poverty reduction. A simple macroeconomic model developed for this report - simulating the impact o f changes in the level and structure o f public expenditures over 2008-2015 - shows that increasing the share o f capital expenditures, particularly in relation to current expenditures (including military and security spending), would have a strong positive impact on economic growth and poverty reduction. More importantly, reducing the relative size o f wage bill would have an even more significant impact on growth and poverty reduction. In other words, the Government could reduce rates o f poverty faster if i t were able to reduce the size o f the wage bill instead o f increasing i t s share o f GDP - the present trend - or even maintaining i t at i t s current level. 4. The report also examines the measures necessary to establish a sound public expenditure management system aimed at promoting fiscal discipline and the efficient and effective operational performance o f the public sector to ensure appropriate allocation o f resources based on agreed strategic priorities, with particular attention to key economic and social sectors (education, health, and 1 infrastructure). The report also identifies constraints to effective implementation o f PFM reforms and defines the elements o f a short and medium-term action plan to overcome these constraints. The action plan matrix, prepared in close cooperation with relevant Government departments and other stakeholders operating in Burundi, proposes a set o f concrete recommendations for public spending reallocations and priority reform measures. 5. I n moving from a post-conflict to a growth and development scenario in Burundi, policymakers face four key sets of challenges: Improving the allocative efficiency o f public expenditures; Reducing the cost o f public employment; Reforming public expenditure management procedures and practices; and Professionalizing the public service. These challenges are discussed below, along with key recommendations and proposed activities. Improving the allocative efficiency of public expenditures 6. High levels of defense and public security spending are the main obstacle to reallocating public resources toward priority economic and social sectors. Therefore, to the extent possible, and keeping in mind the need for security and stability, i t i s essential to reduce the share o f defense and security expenditures over time in order to create fiscal space for increased spending in the priority economic and social sectors.' 7. A reduction in military and security expenditures will depend on the Government taking several key actions to consolidate peace and political stability in the country. I n the short term, these actions include: (a) ensuring that the agreement signed in September 2006 with the FNL i s fully implemented; (b) continuing with demobilization, disarmament, and reintegration (DDR); and (c) continuing the effort to involve other political parties in the governing process. 8. I n parallel with the reduction in military and security expenditures, the Government and the donor community should increase the share of public expenditures allocated to the education, health, and infrastructuresectors, in line with the PRSP. In education, the most urgent short-term priorities for the sector are to: (a) assess the budgetary implications o f the September 2005 decision abolishing primary school fees to adjust budget appropriations needed to finance costs associated with fiee primary education; (b) increase budget transfers to schools and actual spending on goods and services (notably teaching materials) by ' The share o f defense in total domestic public spending declined from 36 percent, o n average, in 2001-2004, to 24 percent in 2005-2006. A t the same time, however, the share o f public security increased from 3 percent in 2001- 2004 to 12 percent in 2005-2006, mainly due to the creation o f the national police force in 2005, and to the harmonization o f police salaries with the higher pay levels for military personnel. Creating a competent police force i s essential to improve security and promote growth. However, effective implementation o f the demobilization program should eventually result in a more substantial decline in defense and public security spending and make it possible to dramatically increase public spending in other sectors. .. 11 about 20 percent in real terms; (c) accelerate implementation o f the primary school construction program (at least 1200 schoolrooms per year), and better coordinate this program with the allocation o f human resources; and (d) improve the geographic distribution o f teachers and infrastructure in favor o f underprivileged provinces, through assessing the budgetary impact o f providing special incentives to teachers willing to work in distant rural areas. Over the medium to long term, the Government should: (a) strengthen the financial management capacity o f the Ministry o f Education (at the central and provincial levels) and create a directorate responsible for administrative and financial management; (b) pursue ongoing efforts to develop, in cooperation with donors, a comprehensive sector strategy that will serve as a basis for the gradual introduction o f program budgeting and better coordination o f donor programs; (c) put more effort to reduce repetition rates from 30 to 15 percent, to reduce unit costs and improve the efficiency o f the education system; and (d) monitor the creation o f the Common Education Fund and encourage other donors to evaluate the initiative and eventually j o i n the program. 0 In health, a major effort should be made to improve health outcomes, which were severely affected by the political and economic crisis o f the 199Os, and have remained particularly dismal.* The most urgent short-term priorities are to: (a) evaluate the budgetary implications o f the M a y 2006 decision abolishing user fees for certain medical services (children under 5 and women during delivery) in order to increase budget allocations, particularly those to the primary health care (PHC); (b) improve the capacity o f health workers, and encourage the retention o f qualified medical personnel in rural areas; (c) establish a more meaningful and effective staff appraisal system (i.e. a contract based system); (d) reorganize the health sector into health districts; (e) redeploy available human resources to provincial and local health facilities outside o f Bujumbura; and (0 improve the provision o f drugs and vaccines, and other critical goods and services, to district hospitals and PHC services. Over the medium term, the Government should: (a) strengthen the financial management capacity o f the Ministry o f Health (at the central and provincial levels); and (b) intensify ongoing strategic planning activities in collaboration with i t s development partners. The process should be supported by the gradual introduction o f program budgeting (starting in 2009/2010), and by the development o f new donor coordination and financing instruments. 0 In infrastructure, high priority should be given to the comprehensive rehabilitation o f the national road network, and to the overhaul o f the urban power and water systems managed by the state- owned utility, REGIDESO. The creation o f master plans in the infrastructure sector and the use o f public-private partnerships to finance the sector should also be encouraged. - Roads. In the short term, the Government should: (a) allocate adequate resources to road maintenance on the basis o f the plan prepared by the Ofjce des Routes (OdR); (b) assess the condition o f rural roads and how their condition impacts the agricultural sector; and (c) * Actual spending on health stagnated in real terms in 2001-2005, and its share o f domesticallyfinanced expenditures declined from 4.1 to 3.3 percent. In 2006, access to HIPC funds had a positive impact on health budgets, but because o f cumbersome disbursement procedures and low rates o f execution, actual spending increased to only 4.6 percent o f total public expenditures. 1 ... 11 allocate appropriate resources to local governments for the maintenance o f a priority network o f regional and rural roads. Over the medium term, the Government should explore, with other members o f the East African Community, ways to improve traffic on regional and international routes linking Burundi with major seaports in Tanzania and Kenya. - Power und water. I n the short term, the Government should: (a) undertake an assessment o f the condition o f rural power and water facilities, as a first step toward the development o f a sound strategy for the sector; and (b) take measures to improve the financial viability and operational performance o f REGIDESO, through a combination o f tariff adjustments and government subsidies. Over the medium term, the Government should: (a) create enabling environment for the private sector by developing the urban power sector, including through privatization o f the state-owned utility; and (b) explore opportunities for regional cooperation to find alternative sources o f energy (e.g., the Lake Kivu methane project in Rwanda). Reducing the cost ofpublic employment 9. The high wage bill needs to be addressed while improving the allocative efficiency of public resources? To reduce the wage bill, public finance policies must consider: (a) eliminating ghost workers and developing an effective human resources management information system; (b) accelerating the demobilization program, taking into consideration limitations in the absorptive capacity o f the labor markets; and (c) developing appropriate institutional arrangements to control growth in the wage bill. However, efforts to reduce the wage bill will be complicated, in the medium term, by the need to recruit and train large numbers o f teachers, doctors and other health personnel, and by the need to professionalize the civil service (see below), which will involve providing incentives for good performance. Cutting out the non-essential and unproductive workers will allow the Government to save some resources, raise the wages o f key technical personnel and provide incentives for improved performance. Elimination of ghost workers. I n the short term, the Government should conduct a census o f the public service (Le., civil, military and police forces), and establish an effective human resources management information system (HRMIS), which provides for biometric identification o f civil service staff and facilitates the control o f staff needs. A new biometric card (carte sdcurituire) i s expected to be issued in 2008 following the completion o f the census. In addition, technical assistance for the HRMIS i s to be provided under the ongoing Economic Management Support Project (known in French as PAGE) supported by the World Bank. I n the medium to longer term, the Government should include a human resources module in i t s Integrated Financial Management Information System. A major problem for the future is the size o f the wage bill (including the police and the military), which accounts for about 40 percent o f domestically financed expenditures. Until 2004, the Government was able to contain the wage bill under the already high level o f 8 percent o f GDP, but that ratio increased significantly over the past two years, reaching about 10 percent o f GDP in 2006, following a 15 percent increase in civil service and military salaries. iv Accelerated implementation of the DDR program. This should help reduce the size o f the army in the short to medium term, though these reductions may be partly offset by the continued expansion o f the national police force. Over the longer term, as security improves, the total number o f soldiers and police force should slowly but steadily decline. Develotment of appropriate institutional arrangements to control growth of the wage bill. This will depend, in the short term, o n the Ministry o f Economy, Finance and Development Cooperation (MEFDC) and Ministry o f Civil Service (MFP) jointly creating a civil service reform committee to work on staffing and salary issues, with highest priority given to improving the monitoring and management o f the civil service and defining clear staffing targets for line ministries. I n the medium to longer term, in order to maintain the wage bill at i t s present level- 11 percent o f GDP - the Government should: (a) encourage voluntary early retirement o f government officials (plan de reconversion); (b) adopt a replacement policy that aims to replace only two out o f three retiring staff in non-priority sectors; (c) introduce a performance evaluation system o f staff to limit politicization in hiring and to increase the quality o f the public service (see below); and (d) harmonize salaries and benefits o f the public service throughout all ministries. Reforming public expenditure management procedures and practices 10. Improving public finance management procedures and practices i s essential to enhance the performance of the public sector, mobilize additional donor resources, accelerate growth and reduce poverty. Despite recent reforms aimed at modernizing the public expenditure management system and improving budget execution: major weaknesses still affect the quality o f Burundi’s P F M system. The main weaknesses hampering the country’s P F M system include: (a) the legal and regulatory framework has not been updated in more than 40 years; (b) budget preparation practices tend to produce budgets that are not comprehensive or realistic, and do not reflect the country’s strategic priorities; (c) budget execution controls are duplicative and burdensome, which encourages Government officials to use exceptional procedures; (d) the financial management information system i s not fully stable and reliable, which hinders budget execution and monitoring procedures; (e) budget and treasury reports are not based on established accounting procedures; (0 cash management i s fragmented; (g) the internal and external audit systems are underdeveloped; and (h) limited capacity in the country hinders the implementation o f the procurement reform. Moreover, to date the reform process has been entirely donor driven and the implementation o f the reforms i s constrained by the lack o f capacity. The Cellule d’appui responsible for coordinating and monitoring the reform process within MEFDC does not have the means (status, structure, or terms o f reference, TOR) necessary to perform i t s functions effectively. Actions taken in 2007 to improve budget execution include: (a) closing 2006 Government accounts by April 2007; (b) improving accounting systems in customs services; (c) creating a Bureau of Quality Accounting within the Directorate of Accounting in the Ministry of Finance; (d) integrating payroll operations in Government accounts; (e) creating a Bureau of cash management; (f) preparing a cash management plan for the October-December 2007 period; and (g) pursuing ongoing efforts to reduce the number o f Government accounts by consolidating existing accounts o f each public entity and closing accounts opened on behalf of administrative services which do not need to be given administrative and financial autonomy. Additional reform measures are also planned for 2008, including: (h) consolidating current and extraordinarybudget accounts; (i)closing HIPC sub- accounts; and 0’) consolidating them into a single HIPC account, which will then be integrated as a sub-account of the Ordonnateur-Trksorier du Burundi (OTB) at the Central Bank. V 11. The Government should further consider addressing the following vital questions: (a) what will be the role, structure and funding o f the Cellule d’appui; (b) should Burundi move toward adopting an integrated financial management information system (IFMIS) to replace the existing budget execution and monitoring system, SIGEFI (SystBme d’information de Gestion des Finances Publiques), and if so, what prerequisites are needed; and (c) what should be the respective roles o f the Inspection Gknirale de 1’Etat (IGE), and sectoral inspection bodies, including the Inspection des Finances (IDF)? 12. T o stabilize the present PFM system, make it more transparent, and strengthen fiscal discipline, the Government should consider implementing the following measures: 0 Legal and renulatow framework. T o further improve the legal and regulatory framework o f public finance management, the Government should, in the short term, adopt a new accounting regulation adapted to the revised Organic Law o f public finance recently approved by the a u t h o r i t i e ~Technical .~ assistance (TA) for drafting this regulation i s provided by IDA and IMF. Over the medium term, the Government should eliminate both the earmarking o f revenues for specific expenses and the systematic use o f exceptional spending procedures (in conformity with the new Organic Law). 0 Budget presentation and preuaration. T o enable the MEFDC and the line ministries to produce realistic budgets that reflect PRSP objectives and specific sector strategies, the Government should, in the short term: (a) consolidate the ordinary budget, the investment budget, and national road funds; (b) set expenditure ceilings for all ministries within the framework o f a clarified budget production timetable; and (c) identify budget items linked to PRSP priorities. I n the medium term, to gradually introduce performance-based management in conformity with the Organic Law, the Government should progressively: (a) begin the introduction o f a medium-term expenditure framework for broad categories o f expenditure, with TA provided by PAGE; (b) strengthen the capacity o f various actors, including macroeconomic forecasting and budget preparation units, to transition to program budgeting; (c) make budget implementation managers (ordonnateurs) accountable for results; and (d) improve budgetary data collection by regiodprovince. Budget execution. T o reduce the excessive use o f exceptional procedures to bypass burdensome controls, the Government should, in the short term: (a) streamline budget execution procedures through the production o f clear budget execution manuals strictly limiting the use o f exceptional execution procedures, and train relevant staff within MEFDC and the line ministries on new procedures; (b) reintegrate most o f the state’s expenditures into normal budget channels processed and accounted for through a stabilized information system (see below); (c) improve internal controls over payroll by introducing monthly checks o f staff lists as a basis for payroll ’The Government recently prepared and enacted an Organic Law that defines the principles o f public finance management by clarifying the responsibilities o f all actors involved in the PFM, and establishingthe foundations o f sound input-based budget managementpractices. Technical assistance (TA) for drafting the Organic Law was provided by IDA and the IMF. The proposed law recently approved by the Parliament was promulgatedby the President of the Republic. vi data; and (d) introduce more efficient accounting and budgetary procedures for the exhaustive recording, monitoring, and clearing o f arream6 I n the medium term, the Government should: (a) define an action plan to strengthen the absorptive capacities o f ministries in order to improve rates o f budget execution; (b) strengthen the internal audit unit o f the MEFDC, concentrating on i t s efforts in the monitoring o f the budget, reviewing procedures, and improving budget management; (c) establish structures for the monitoring o f budget execution in the sector ministries, and formalize the interaction between MEFDC and the sector ministries on budget execution monitoring; and (d) follow systematically delays in the execution o f expenses at every stage in budget execution procedures. 0 Information svstem. T o stabilize and improve the reliability o f SIGEFI, which currently excludes a number o f important functionalities, including payroll processing and double-entry accounting, MEFDC should address, in the short term, major gaps in application and introduce the functions necessary to generate the required budget execution reports. I n the medium term, the Government will need to decide if it i s necessary to replace SIGEFI with a new system better adapted to the Government’s needs or to adopt the existing integrated information system for financial management. 0 Accounting and financial reuorting. To address discrepancies between revenue and expenditure data produced by various MEFDC departments and the overall budget execution report (Loi de Riglement) submitted to Parliament as well as deficiencies in accounting procedures and internal controls, the Government should, in the short term: (a) produce monthly trial balances and regular reports on revenues and expenditures (including donor-funded projects) to allow for verification and control; (b) establish within MEFDC a unit in charge o f centralizing and analyzing financial and budgetary data on foreign-financed expenditures, which will collaborate with sector ministries, aid coordination sector groups and the national aid coordination agencies; and (c) define project financial reporting requirements in a letter to technical ministries, project coordinators and donors. I n the medium term, the Government should: (a) institute a mandatory reporting system for special funds and special accounts; (b) complete, for each year, the necessary verifications and regularizations on the general balance o f accounts before integrating it into the accounts for the following year; and (c) gradually develop, on the basis o f capacity and need, a unit within line ministries for collecting sectoral foreign-financed expenditures data. Cash manazement. T o improve revenue projections and the ability to develop realistic cash management plans, the Government should, in the short term: (a) progressively reestablish the Treasury Single Account (TSA),’ by completing the ongoing consolidation o f accounts o f different public entities; (b) abolish accounts managed by entities that are not financially This measure includes: (a) defining a monthly plan for the execution o f recurrent discretionary expenses; (b) training managers on their role in public expenditure management and on the procedures included in the new public procurement code; and (c) making commitment (engagement) contingent on availability o f funds, or take all other measures necessary to eliminate arrears. ’The TSA refers to the consolidation o f OTBU (Ordonnateur Tresorier du Burundi) accounts and sub-accountsheld at the central bank and managed by the State Cashier (Cuissier de 1 ‘Etut) within the central bank, by merging ordinary and extraordinary investment budget accounts, eliminating HIPC sub-accounts, and reducing the number o f government accounts. vii autonomous; and (c) produce, on a quarterly basis, a global cash management plan and a sectoral plan for each ministry and disseminate i t publicly to enable each ministry to be aware o f the financial constraints o f the entire Government. I n the medium term, the Government should create a Cash Management Committee under the authority o f the Minister o f Economy, Finance, and Development Cooperation to supervise ongoing reforms and help the country move gradually toward a single treasury account. 0 Internal and external audit svstems. To improve the integrity o f the audit systems, the Government should, in the short term: (a) expand the capacity o f ministerial audit units and the investigative role o f the Inspectorate General o f the State (IGE) for corruption cases; (b) strengthen the Cour des Comptes (CdC) through increased resources and technical assistance; (c) clarify the respective roles o f the IGE and internal audit units in line ministries, and separate the audit and investigation functions; and (d) improve cooperation between the CdC and the IGE. I n the medium term, the Government should: (a) enhance i t s capacity to act upon reports from the CdC and IGE, including sanctioning those involved in misappropriation and responding appropriately to information on poor policy performance; (b) ensure that recruitment, training, quality assurance processes, peer reviews, and certification are in line with international modem practices; and (c) enhance the capacity o f Parliament to review the accounts prepared by the CdC. n that context, the Government would have to define adequate follow-up and monitoring I procedures for both investigative and audit work; and create a technical assistance plan to systematically address the capacity needs and constraints identified for both functions. 0 Procurement reforms. For successful implementation o f procurement reforms, the Government should, in the short term: (a) establish the Procurement Regulatory Agency, the National Directorate o f Public Procurement in MEFDC, and procurement units in the main procuring entities; and (b) assess training schools and institutes that could be involved in the design and delivery o f a comprehensive training program. In the medium term, the Government should: (a) adopt and disseminate a first set o f modem procurement tools (standard bidding documents, guidelines, manuals, procurement filing system, and e-procurement) based on the new Procurement Code; and (b) adopt and launch the implementation o f a procurement capacity building program. Ownership qf PFM reforms. On the basis o f this report and those produced in collaboration with other donors on P F M issues in the past few years, the Government should: (a) develop a realistic timeframe for reforms and capacity building strategy; and (b) prepare the terms o f reference and an operational budget for the Coordination Unit (Cellule d 'appui). Professionalizing the public service 13. Politicizing the civil service weakens the effectiveness of the public administration and i s a barrier to the recruitment of highly qualified personnel. About three-quarters o f staff, including political appointees, generally receive the highest ratings, reflecting a bureaucratic culture that does not penalize poor performance. I n addition, most o f the ministries do not provide their staff with the tools necessary to improve their performance, and the maintenance o f equipment and facilities i s affected by ... VI11 insufficient budget allocations. The lack o f computers, for example, encourages the use o f manual reporting techniques. 14. Inadequate funding of critical current expenditures i s at the heart of the wage bill and incentives issue. The high and rising wage bill i s crowding out investments needed to enable staff to be productive. T o improve the quality o f the civil service and make i t more efficient, the Government should consider the following measures: 0 I n the short term: (a) create staff appraisal and grading systems based on expertise and performance; (b) reduce the number o f positions reserved for political appointees; and (c) increase substantially the budget for equipment, facilities, and in-service training. 0 I n the medium term: (a) upgrade the recruitment commission into a public service commission fully responsible for the professionalization o f the civil service; (b) create within each ministry the position o f Secretary General, occupied by a professional civil servant, who would be the key manager o f the ministry under the control o f the Minister; and (c) restructure ministries and other public institutions, taking into consideration recent changes in the workload and responsibilities o f each department. In other countries, responsibility for leading the restructuring process lies with the Civil Service Ministry or i t s equivalent. Other issues affecting public expenditure management prospects in the medium term 15. Burundi’s medium-term growth prospects depend on: (a) the successful conclusion o f the political transition; (b) a major improvement in security conditions; (c) sustained macroeconomic, public finance, and sector policy reforms; (d) successful integration o f Burundi into East African Community (EAC); and (e) strong technical and financial support from the donor community. Three positive factors - achievement o f the HIPC Decision Point in August 2005; the good revenue performance o f the Government (about 20 percent o f GDP in 2001-2006); and increased donor support (program grants and loans averaging 6 percent o f GDP) - have helped and should continue to help finance the country’s economic recovery. The donor community has recognized that a substantial increase in financial assistance i s critically needed in this post-conflict country, which has made significant progress toward political and economic stability in the face o f major medium and long-term economic reconstruction, rehabilitation, and development challenges. However, a consistently high level o f donor financing, particularly budget support grants, i s not assured in the long run, and the Government must work to further increase tax revenue by reviewing the tax code, broadening the tax base, reducing tax exemptions, and improving tax administration. This effort i s particularly necessary since Burundi’s admission to E A C will require major adjustments in the tax regime. 16. The fiscal space analysis conducted in this report confirms that Burundi does not have enough fiscal space options. In addition to improving the effective use o f public resources (e.g., through declining shares o f police and military spending), as noted in paragraph 2, fiscal space could be increased by: (a) finding more growth opportunities (creating an environment to attract private investment; and rehabilitating the productive sectors); (b) increasing domestic revenues; and (c) avoiding misappropriation o f funds through stronger fiduciary controls. ix 17. The following matrix provides a l i s t o f the most critical recommendations that need to be undertaken by the Government in the next 12 to 18 months. A more detailed policy matrix (including medium and long-term recommendations) i s provided in the annexes. Summary o f Priority Policy Actions (to be implementedin the next 12-18 months) Area Action Timetable Technical Responsibility Assistance PUBLIC EXPENDITURE REVIEW Wage Develop institutional arrangements to control the 2008-2009 National MEFDC, MFP, Management growth o f the wage bill: For instance, setting up a joint Resources Line Ministries team o f M E F D C and MFP to keep the wage bill at sustainable level and address staffing and incentives issues. Eliminate ghost workers and develop an effective 2008-2009 TA MEFDC, MFP, human resources management information system. provided Line Ministries by PAGE Social Sectors Assess the budgetary implications of: (a) the September 2008-2009 TA MEFDC, MFP, 2005 and M a y 2006 decisions abolishing primary required MPH, school fees and user fees for children under 5 and MINEDUC women during delivery, respectively; and (b) the provision o f special incentives to teachers willing to work in distant rural areas to improve the geographic distribution o f teachers and infrastructure in favor o f underprivileged provinces. Infrastructure Allocate adequate resources to road maintenance on the 2008-2009 National MEFDC, MDT, basis o f the plan prepared by the Office des Routes Resources OdR (OdR) . PUBLIC EXPENDITURE MANAGEMENT Budget Improve budget presentation by consolidating the 2008-2009 National DGBCPDBC Preparation ordinary budget, the investment budget, and national Resources road funds. Budget Execution Reduce the excessive use o f exceptional procedures by 2008-2009 TA DGBCPDBC complying with streamlined budget execution resources urocedures as uer the new budget execution manual. needed Introduce efficient accounting and budgetary 2009 TA DGBCPDBCDT procedures for the exhaustive recording, monitoring, resources and clearing o f arrears through: (a) defining a monthly needed plan for the execution o f recurrent discretionary expenses; (b) training managers on their role in public expenditure management and on the procedures included include the new public procurement code; and (c) making commitment (engagement) contingent on availability o f funds. Financial Improve the Computerization o f Financial Management 2008-2009 TA DGBCPDC Management (SIGEFI) through stabilizing SIGEFI and adding resources Information essential budget execution functionalities. needed System X Internal and Improve the effectiveness o f oversight by: (a) 2008-2009 National DGBCPDC External Audit expanding the capacity o f ministerial audit units and the Resources Systems investigative role o f the IGE for corruption cases; and (b) strengthening the Cow des Comptes (CdC) through increased resources and technical assistance. Procurement Establish the Procurement Regulatory Agency, the 2008-2009 National MEFDCDGMP Reforms National Directorate o f Public Procurement in MEFDC, Resources and procurement units in the main procuring entities. xi INTRODUCTION A. POLITICAL AND SOCIO-ECONOMIC SITUATION 1. Burundi i s just emerging from a 13-year civil war. This follows the signing o f a comprehensive peace and reconciliation agreement by 39 politico-ethnic groups in Arusha (Tanzania) in August 2000, which marked a significant political turning point. Following the establishment o f a transitional Government in 2002, a new constitution that provided for appropriate ethnic checks and balances was approved in February 2005. Presidential and parliamentary elections were held in August 2005, and the new Government took office in September 2005. I n September 2006, a cease-fire accord was reached with the last hold-out rebel movement, the FNL-PALIPEHUTU faction. Peace i s s t i l l in the process o f consolidation, as the Government i s in negotiations with FNL-PALPEHUTU. The political situation remains fragile; recently the ruling coalition replaced and imprisoned the controversial chairman o f the main political party (CNDD-FDD). I nNovember 2007, the political deadlock, which had lasted several months, was broken with the nomination o f a more representative Government, and the Parliament resumed i t s legislative role. 2. However, at the beginning o f 2008, a new political impasse emerged and the national assembly did not meet until mid-2008. Earlier in M a y 2008, the rebel movement conducted a two-day armed attack on Bujumbura, which caused civilian casualties. I n June 2008, following pressures from the international community and intervention by regional powers such as South Africa and Tanzania, the FNL leadership returned to Bujumbura, including Agathon Rwasa, the movement’s leader, and they rejoined the Joint Agreement o f Verification and Control Mechanism (JAVC) after the Government reiterated the immunity granted to the FNL, leadership. The President o f the Republic and the FNL leader recently agreed to meet on a bi-weekly basis for peace talks, and to restore public confidence. I n late July 2008, following a contested Constitutional Court decision, 22 parliamentarians who had rebelled against the ruling party were replaced by other members o f the same party in the Parliament, thereby allowing the ruling party to recover the majority in this legislative body. 3. The thirteen years o f recurring conflict has had a devastating effect on Burundi’s economy. With a population o f 7.8 million, Burundi’s GDP per capita f e l l by about 30 percent, from US$152 in 1994 to US$106 in 2005. Among children less than 5 years old, undernourishment reached 45 percent in 2000. According to the United Nations Development Programme’s (UNDP) 2006 Human Development Index, Burundi i s the ninth least-developed country in the world. 4. Recent economic and political developments, however, have created positive prospects for Burundi. The economy has gradually recovered from the shocks experienced during 1993- 2003 (political turmoil, embargos, deteriorating terms o f trade, and severe drought), and annual real GDP growth i s estimated to have risen to 5.1 percent in 2006, from an annual average o f -1.8 percent during the decade o f economic and political turbulence. The authorities successfully concluded a macroeconomic program agreed with the IMF under the Poverty Reduction and Growth Facility (PRGF) arrangement that covered the period January 2004 to January 2008. Burundi has made good initial progress since 2000 in stabilizing the economy, implementing 1 n September 2006, the financial and structural reforms, and beginning to restore social services. I Government approved i t s first full Poverty Reduction Strategy Paper (PRSP), which was presented to the boards o f the International Development Association (IDA) and the International Monetary Fund (IMF) in March 2007. 5. Prospects for growth and poverty reduction appear encouraging, provided that the country i s able to (a) strengthen political stability and consolidate peace; (b) increase public investment; (c) maintain strong external support; (d) make progress on economic governance; and (e) improve social and economic infrastructure. Burundi has now reached a sensitive period. I n the second semester o f 2007, the IMF delayed the conclusion o f i t s sixth review o f the PRGF program because o f a specific governance concern - the illegal and fraudulent budget payment to a private petroleum company (Interpetrol) that occurred in May-June 2007. This followed on the heels o f an illegal sale o f a presidential aircraft (a 1982 Falcon) in 2006. Consequently, the release o f budget support from some multilateral donors was delayed, with the risk o f exacerbating social tensions and jeopardizing the peace and economic dividend o f the past few years, added to the fact that recent political tensions have increased as a result o f the FNL walking away from the peace talks. 6. Attaining the ambitious goals o f the PRSP and making significant progress towards the Millennium Development Goals (MDGs) will require major improvements in mobilizing and managing public resources. Indeed, the objective o f achieving inclusive growth in Burundi will require that the poor reap the full benefits o f higher growth rates. Public expenditure policies have a critical role to play in this regard, by ensuring that the allocative process channels sufficient resources to priority poverty reducing sectors. Within the priority sectors, resources have to be allocated to programs with maximum poverty-reducing and growth impact. Transparency in the expenditure o f public resources and accountability for the uses o f these resources will have to be improved to ensure that programs have their intended poverty-reducing impact. To achieve this, and building o n its recent satisfactory economic management, the Government will need to substantially improve i t s fiscal space, as well as consolidate and expand public expenditure management (PEM) reforms, covering all phases o f public expenditure management from the preparation o f budgets, their approval, execution policies and programs, and accounting for the use o f funds. €3. RECENT REFORM ACHIEVEMENTS 7. Since the peace agreement o f Arusha in 2000, Burundi has undertaken a series o f socio- economic and structural reforms in areas o f public financial management, trade policy, and human development. Public Expenditure Management 8. With the resumption o f financial assistance from donors and the HIPC Initiative, Burundi’s public expenditure management has been under close scrutiny. Given the dire state o f P E M after more than a decade o f civil war, priority has been given to restoring the fiscal base and rehabilitating budget management by fixing the fundamentals at every step o f the budget process. The main weaknesses in the budgeting process include: (a) weak administrative capacity for budget preparation both at the Ministry o f Economy, Finance and Development Cooperation (MEFDC) and sector ministries; (b) fragmentation o f the budget preparation process; (c) lack o f congruence with macroeconomic forecasts and budget constraints; (d) lack o f inclusion o f all 2 sources o f funds and categories o f public expenditure; and (e) a Budget L a w that lacks clarity and i s subject to varying interpretations. 9. As indicated in the Country Financial Assessment Report (2004), significant progress has been made in improving P E M and reducing fiduciary risks. An interim and limited JFMIS (Integrated Financial Management Information System, or SIGEFI after i t s French name), introduced in January 2006 - which can generate standard quarterly budget execution reports and reports on the execution o f pro-poor HIPC expenditure, with a view to strengthening expenditure controls - i s now operational. Also, a new unified functional and economic budget classification system and a double-entry accounting system were adopted and are being implemented. Consequently, based on the new functional classification o f public expenditure, an identification system has been put in place to track pro-poor spending and HIPC expenditures. 10. Nevertheless, many deficiencies remain, including: (a) the lack o f harmonization between the nomenclature o f public accounting system and the budget; (b) the inadequacyhnaccuracy o f records o f the Ministry o f Economy, Finance and Development Cooperation (MEFDC), which relies on the central bank (Banque de la Republique du Burundi) for information o n effective payments; and (c) the weak oversight capacity in government institutions, a problem exacerbated by unclear roles and responsibilities, leading to the possibility o f either duplicated or lacking functions. The creation o f a national Audit Court ( C o w des Comptes) in 2004 was an important step towards the strengthening o f jurisdictional control over public expenditure management. A modem draft Budget Framework L a w (Loi Organique des Finances), recently prepared with IDA and IMF support and adopted by the authorities, will address existing weaknesses and strengthen the management o f public expenditure by clearly defining the core principles for the streamlined and effective management o f public resources. 11. T o address the numerous P E M weaknesses discussed above, in October 2006 the Government reactivated the donor coordination committee on budget support, with responsibility for the entire reform program (including restructuring MEFDC), so that the committee can better coordinate the preparation o f the budget. Finally, under the IMF’s program, the tax administration i s preparing to introduce a value-added tax (VAT), and the Tax Code will be revised in line with Burundi joining the East African Community (EAC), which requires the harmonization o f practices among country members, including the convergence o f fiscal policies. Human Development 12. Education. Improvement in access to and quality o f education i s a priority in the new Government’s national recovery strategy. The Ministry o f National Education and Culture (MINEDUC) has prepared a strategic plan to reach the goal o f Education for All (EFA). In line with EFA’s objective, the Government decided in September 2005 to eliminate school fees. This resulted in a huge increase in first grade enrollment. The primary school enrollment rate has significantly surpassed the pre-crisis, long-term level, with a gross enrollment o f 80 percent in 2003/04 rising to 100 percent in 2005/06. Public spending on education has significantly increased to about 6 percent o f GDP in 2006 from 4.6 percent in 2005, mostly explained by the recruitment o f 6,500 additional teachers and the construction o f new classrooms (capital expenditures were financed with HIPC resources). These reforms, however, still leave Burundi far from reaching the MDG for education, which i s to achieve universal primary education by 2015. 3 13. Health. Through a presidential decree, all fees for health care services for children under five and women during deliveries were eliminated. However, due to delays in the release o f HIPC funds meant to finance these services, this policy could not be implemented effectively, as planned, since the huge rise in the demand for health services overwhelmed the capacity o f the Government to provide those services. Also, health facilities which provided the service were not reimbursed o n time. In 2006, the Government has continued i t s efforts t o operationalize i t s National Health Development Plan and to coordinate development partners’ support in health. The Government has taken important steps, including the creation o f a new Department o f Resources, to remedy the various weaknesses in public expenditure management that affect health service delivery. C. RATIONALE FOR PEMFAR 14. This report integrates the analysis o f a Public Expenditure Review (PER) and a progress assessment o n existing PEM reforms, referred to here as Public Expenditure Management and Financial Accountability Review (PEMFAR). These reforms resulted from the adoption o f recommendations from several donor reports, including IDA (Country Financial Accountability Assessment (CFAA) 2004, Country Procurement Issue Paper (CPIP) 2004, and Economic Reform Support Grant (ERSG) 2006); IMF’s Fiscal Affairs Department; USAID; and others. The motivation for carrying out such an integrated analysis i s three-fold: First, the analysis o f fiscal discipline and inter and intra-sectoral allocation o f expenditure (PER exercise) depends o n a complementary assessment o f the legal framework, procedures, and accountabilities in the existing PEM system (CFAA exercise). Public procurement issues and reforms (CPIP) are best articulated in a broader context o f public finance management reforms (PER and CFAA exercises). In addition, the last PER was carried out six years ago (2001).8 15. Second, the focus o f this P E M F A R o n implementation and capacity building reflects the view that there i s n o need to conduct further studies on P E M systems, given that reports by IDA, IMF, U S A I D and others have already identified what actions the Government needs to undertake to improve public finance management. Moreover, there have been a number o f recent reviews on the progress o f reforms, which have made further recommendations for a ~ t i o n Third, .~ the new political environment and the new Government’s commitment to reforms offer a window o f opportunity to advance public expenditure management and financial accountability reforms in Burundi. The current context also favors donors’ commitment to provide more resources to finance Burundi’s development needs. T o support Burundi’s development efforts, it will be important that increased donor funding be provided through the budget. However, attracting and sustaining higher aid flows will require that minimum fiduciary standards are met. In that connection, the proposed PEMFAR will provide input to the preparation o f donor assistance strategies, including IDA’SCountry Assistance Strategy (CAS) (August 2009), as well as to the preparation o f an anti-corruption and governance strategy t o be derived from the anticorruption This was not a core World Bank PER, but a technical assistance (TA) PER that led IDA and the Government to the design o f the Economic Management Support Project (PAGE, after its French name), which was approved by the Bank’s Board in January 2004. Moreover, in August 2006, an adjustment operation, ERSG, was approved by the Bank’s Board with the aim of supporting various reforms, including PEM reforms. T h e IMF’s Fiscal Affairs Department (FAD) report from December 2006 indicates that some progress has been made, especially in the execution of the budget, but also identifies the problems that remain to be addressed. 4 and governance survey recently conducted by the W o r l d Bank Institute (WI) at the Government’s request. 16. Finally, the full PRSP adopted in 2006 stresses the importance o f addressing weaknesses in the PEM systems and capacity. A P E M F A R i s required to address the many PFM institutional and capacity issues that impede the effective allocation o f resources and execution o f the budget in line with the PRSP, as well as the factors that prevent accounting for the use o f funds as approved by Parliament. The PEMFAR could also lay the groundwork for conducting PERs in key sectors o n an annual basis. The results o f these reviews could be presented in the annual progress reports o f the PRSP. This would enable a close monitoring o f the allocation and execution o f budgetary resources in PRSP priority sectors, and an assessment o f their impact on growth and poverty. D. OBJECTIVES, SCOPE AND STRUCTURE 17. The PEMFAR assesses past budget outcomes, existing systems and capacities for public expenditure reform, in order to learn lessons from previous reform efforts. The findings would be translated into a proposed action plan for reform o f public expenditure and financial management to be considered by the Government. More specifically, the principal objective o f the PEMFAR i s to identify the key actions that will have to be taken to put in place a PEM system to facilitate the preparation o f a budget in line with the Government’s policy priorities, in particular in terms o f stimulating growth and reducing poverty; efficiently and effectively executing the budget; and fully accounting for the uses o f funds against the uses approved by the National Assembly. The plan prioritizes and sequences the proposed measures, and coordinates the actions and interventions o f the various actors. The PEMFAR will build upon the many studies o f public expenditure allocations and outcomes, and o f public expenditure management, that have been carried out in recent years; as well as the plans to improve public expenditure management already agreed to by the Government. The review will focus o n the problems that have been encountered in implementing the agreed-upon reforms, including a lack o f professional capacity in the public service, ineffective administrative processes, and weak performance incentives. 18. The PEMFAR format was adopted because it allows for combining the assessment o f budget results through a Public Expenditure Review (PER) with an assessment o f public expenditure management and financial accountability. Each chapter o f this report focuses on a key question: 0 Chapter 1 (Macroeconomic Context): Are Burundi’s fiscal policies sustainable in the context o f the current and projected macroeconomic framework? 0 Chapter 2 (Public Expenditure Review): Are allocations o f public expenditures and their execution in line with the Government’s overall priorities? Chapter 3 (Sectoral Public Expenditure Review): A r e intra-sectoral allocations o f public expenditures in education, health, and infrastructure consistent with the Government’s sectoral objectives? 0 Chapter 4 (Public Expenditure Management Reform): Do public expenditure/financial management systems ensure effective use o f resources and fully account for the uses o f resources? 5 0 Chapter 5 (Salaries and Professionalization): What salary levels and structures create the incentives needed to achieve public expenditure objectives, and at the same time result in a wage bill that does not crowd out the expenditure needed to make staff productive in meeting service delivery objectives? 19. The PEMFAR thus provides a diagnosis o f the inter- and intra-sectoral allocation o f expenditures for key sectors (education, health, and infrastructure), and also helps to operationalize the recommendations for improving P E M made in previous reports, by identifying the cross-cutting constraints that are impeding the implementation o f reforms. 20. Based on the findings o f i t s analytical work and the diagnosis o f constraints to existing P E M reforms, the PEMFAR identifies priorities for a short-, medium-, and long-term action plan to overcome these constraints. This action plan was prepared in collaboration with the Government o f Burundi, with the active participation o f donors, and consists o f a concrete set o f recommendations for priority reform measures and possible reallocations. It will be a key input to the development o f the strategy for public finance management (PFM) reforms currently under preparation by the Government o f Burundi. 6 CHAPTER 1. MACROECONOMIC CONTEXT 1.1 T h i s chapter analyzes the structure o f Burundi’s economy, describes recent developments, evaluates economic prospects, and estimates financing requirements. 1.2 Three factors have shaped Burundi’s recent economic performance: (a) the effects o f the 1993-2000 civil war, which led to a sharp decline in economic activity, the destruction o f capital stock, and a plunge in donor assistance; (b) the vulnerability o f agriculture, which accounts for 45 percent o f GDP, to unfavorable weather conditions; and (c) sharply fluctuating coffee production, which accounts for 68 percent o f exports. 1.3 Burundi has repeatedly been beset by episodes of political violence since its independence in 1962. The latest conflict erupted in 1993, following the assassination o f the democratically elected president Ndadaye, which led to large-scale communal violence and civil war. Following the signature o f the h s h a Peace Accord in 2000, Burundi has been in a period o f transition, reform, and rehabilitation. Increased donor confidence has led to a large rise in external assistance, which has helped finance an increase in capital expenditure and the demobilization, reinsertion, and reintegration (DDR) o f ex-combatants. Since the election o f a democratic Government in 2005, Burundi’s f i r s t full poverty reduction strategy - which aims to strengthen political stability, consolidate peace, and reduce poverty through accelerated, sustainable, and equitable growth - has been finalized, and in November 2006 the country was admitted to the East African Community (EAC). 1.4 There has been a noticeable improvement in Burundi’s economic performance in 2001-06 relative to the previous decade, although real GDP growth, averaging 2.7 percent, remains low, reflecting cyclical fluctuations o f the agriculture sector. There has been a shift in the composition o f public expenditure in the last six years, with an increasing share o f resources allocated to education, health, and infrastructure, as well as a doubling o f capital expenditure. Nevertheless, Burundi faces a number o f significant challenges, including the need for a more diversified agricultural and economic base and higher agricultural productivity, while investing in human resources and physical infrastructure to increase economic growth and reduce poverty. A. Structure of Burundi’s Economy and Public Finances Economic Structure and Growth 1.5 With an estimated per capita income in 2006 of US$109, Burundi i s one of the poorest countries in sub-Saharan Africa. I t i s a predominantly rural and agrarian economy, with the agricultural sector accounting for nearly half o f economic output. Commercial agricultural production, primarily o f coffee, contributes less than 5 percent o f GDP but generates more than 50 percent o f export earnings. Agricultural production i s volatile and highly dependent on favorable climatic conditions and, like other sectors, has been adversely affected by the effects o f prolonged civil strife. Burundi’s small industrial sector consists mainly o f construction, agricultural processing, brewing, and energy. Services comprise 35 percent o f GDP (2006) and 7 are largely concentrated in the growing public sector, which accounts for 64 percent o f the services sector. 1*6 Growth performance has Chart 1.1: Real Output Growth, 2001-06 (in percent) been volatile, as changes in 1 Burundi’s annual growth rates are largely driven by growth in agricultural output (Chart 1.1). Agricultural performance has been adversely affected by structural constraints, including highly 2006 fragmented land ownership, V declining soil productivity, and a lack o f inputs. Coffee production has been extremely volatile, with ’ --RedGDPgmrr(h -AghuIturdoulpul gmkth Source: Central bank. arabica production falling by 84 percent in 2005 and rising by 386 percent in 2006.’0 Such large fluctuations in coffee production have impacted export growth, although rising coffee prices in the last six years have helped to dampen this negative effect. Given the weight o f the primary sector, raising overall agricultural productivity and diversifying Burundi’s export base - by moving up the value chain and identifying new agrarian and agro-industry export niches - will be essential if the country i s to significantly raise economic growth and reduce its volatility, while sustaining poverty reduction. level. Burundi’s annual 8oo. .; population growth, at an estimated 3.6 percent in 2005, i s among the highest in sub- 4w- Saharan Africa. T h i s suggests 3w- that i t will be difficult to 2w- 1w - increase living standards in , i , , , , , , , , i , i i , , , , , l I , , , , I # / I / I / ## , # I ?I I I I , I f coming years without a slow- ~ t ~ m o ~ ~ w m o ~ t w m o ~ t w m o ~ t down in population growth or a ------ w m w w w ~ ~ ~ h h m m m m m m m m m m o o o m m m m m m m m m m m m m m m m ~ m o o o -a-Burundi --SubSahaanAfha - - N N N 1.8 Inflation, as measured by the consumer price index, has averaged 7.2 percent in 2001-2006 and 8.4 percent in 2007, from a low o f 2.7 percent in 2006. This average masks loT h i s i s partly explained by the old age o f the coffee trees, which cannot carry two consecutive years o f high production. 8 quite large fluctuations, with inflation peaking at 13.6 percent in 2004, before falling to a l o w o f 2.7 percent in 2006. Inflation in the medium term i s expected to remain in the single digits, as bank financing o f the budget i s limited. The consumer price index (CPI) i s based on a basket o f goods which has remained unchanged since 1991, and includes the price o f goods sold in the central market o f Bujumbura. A new household consumption survey i s planned and i s likely to lead to a revision in the basket o f goods included in the calculation o f the index. Poverty Profile and Human Development Indicators Table 1.1: Poverty Incidence Table 1.2: Social Indicators (in percent o f population) (in percent, unless otherwise indicated) 2006 Indicator Burundi SSA National 66.8 Population growth 3.6 2.5 Rural 68.8 Population density (persons sq km) 318.3 33.1 Urban 34.0 Rural population (% of total) 90.2 64.6 Province GNI per capita (US$) 100.0 858.3 Bubanza 57.0 Life expectancy at birth (years) 49 50.5 Bujumbura Rural 64.3 Infant mortality rate (per thousand) 109 93.8 Bururi 56.7 HIVIAIDS prevalence rate 3.3 5.8 Cankuzo 67.7 Literacy rate (percent of population) 59.3 61.1 Cibitoke 59.5 Gross primary enrollment rate 103.2 93.8 Gitega 68.2 Male 108.1 99.3 Karuzi 69.0 Female 98.4 88.2 Kayanza 75.5 Net primary enrollment rate 74.6 69.5 Kirundo 82.3 Male 75.9 72.3 Makamba 57.1 Female 73.3 66.6 Muramvya 70.0 Source: World Bank, World Development Indicators, 2006 Muyinga 70.5 Mwaro 61.5 Ngozi 75.4 Rutana 72.9 Ruyigi 75.9 Bujumbura Mane 28.7 Source: World Bank 1.9 For many years, conflict displaced a large portion o f the rural population and destroyed important economic and social infrastructure. As a result, poverty i s particularly widespread in rural areas, which account for more than 90 percent o f the population. Surveys estimate that the poverty incidence stood at 66.8 percent in 2006, and has probably not changed significantly since 1998 (Table 1.1). Although a household survey was conducted in 1998, this survey was less exhaustive than the 2006 survey, and poverty figures cannot be compared, as the earlier survey i s likely to have overestimated poverty." However, these national figures mask " The 1998 survey only included food consumption for purchased products and self-consumption, but excluded food grants from government and non-governmental organizations, such as the World Food Program (WFP), as well as private gifts. The 1998 survey also excluded some categories o f food, and these factors combined are likely to have resulted in an overestimationof poverty rates. The PRSP used the poverty data from the 1998 survey, which estimated 8 1.1 percent at a national level. Both the 1998 and 2006 surveys used cost of basic needs for estimating poverty rates. 9 important differences between rural and urban poverty rates, with rural poverty estimated at twice the rate o f urban areas. Regional differences are particularly acute, with poverty rates fluctuating from a low o f 28.7 percent in Bujumbura to a high o f 82.3 percent in the province o f Kirundo.” Available social indicators (Table 1.2) show some improvement, with life expectancy rising from 42.8 years in 2000 to 44.6 years in 2006, while the net primary enrollment rate rose from 43 percent to 74.6 percent over the same period. With the introduction o f free primary education in September 2005, this figure has risen considerably in 2006.’3 Despite recent progress, Burundi was ranked 16gthout o f 177 countries in the UNDP’s Human Development Index in 2006. B. Public Finance Structure Government Revenue 1.10 M o r e than 90 Table 1.3: Composition of Government Revenue, 2001-2006 Dercent of Burundi’s domestic revenue i s (in \ n 2001 - oercentape o f GDP) 2002 - 2003 I 2004 2005 2006 Average Total revenue 23.1 24.9 30.2 31.5 27.5 27.7 27.5 derived from taxes, with Domestic revenue 20.0 20.3 21.1 20.1 20.0 19.0 20.1 -~ an average o f 71 percent Tax revenue 18.8 17.9 18.7 18.3 18.5 17.4 18.3 coming from taxes on Income tax 5.2 5.0 5.0 4.9 4.9 4.9 5.0 Taxes on goods&services 8.9 8.9 9.1 9.2 9.1 8.9 9.0 domestic activity (taxes on Taxes on international trade 3.9 4.0 4.6 4.1 4.5 3.2 4.0 goods and services and Other tax revenue 0.0 0.7 0.0 0.0 0.0 0.4 0.2 income tax) and about 20 Property tax 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Nontaxrevenue 2.4 1.3 2.3 1.8 1.8 1.5 1.6 percent from taxes on Grants 3.0 9.1 4.6 11.4 7.4 7.5 8.7 international trade during Current 3.5 1.9 2.4 6.4 4.0 5.2 4.3 Capital 1.1 5.6 2.2 5.0 3.4 2.3 4.4 the period 2001-2006 Source: IMF, (Table 1.3). Given i t s narrow tax base, domestic revenue as a percentage o f GDP i s high, at 20.1 percent in 2006. Further increases in domestic revenue will depend on improved revenue administration and the elimination o f excessive tax exemptions (Box 1.1). In this regard, the Government has plans to revise the general Tax Code and unify the local and national revenue administrations. As a result o f Burundi’s accession to the East African Community, tax policy will be reviewed to bring it in line with other member countries. The introduction o f a common external tariff will lead to a reduction in international tax r e ~ e n u e . ’Grants ~ represent an increasing source o f revenue in recent years, rising to 45.6 percent o f total revenue, thereby raising government revenue to 27.7 percent o f GDP in 2006. T h i s means that Burundi i s reliant on a category o f revenue that can be subject to significant delays and i s less stable than other sources o f revenue. l2The low level for Bujumbura suggests that the survey results must be treated with some caution. j 3 Following the September 2006 free primary school decision, the primary school enrollment rate has significantly surpassed the pre-crisis level, with a gross enrollment o f 80 percent in 2003/04 rising to 100 percent in 2005/06 and 1 18 percent in 2006/07. l4 World Bank technical assistance i s currently reviewing the impact o f the East African Community membership on government revenue. 10 Box 1.1: Tax Regime in Burundi Burundi’s tax regime i s riddled with large tax exemptions, which make i t overly complex and opaque, leading to important distortions which impact negatively o n producers. Tax exemptions are widely perceived as inefficient, leading to discretionary decisions and delays. The various exemptions, which mainly concern imports, constrain revenue collection and make tax administration more complex. Most tax exemptions concern import operations. According to a recent study, 60 percent o f imports entered Burundi in 2006 with partial or total exemptions, costing FBu 103.7 billion. This i s equivalent to 10.7 percent o f GDP and 65.6 percent o f tax receipts. The largest customs tax exemption item i s tax-free imports for diplomatic missions, which account for 71.4 percent o f total import exemptions. The Government’s decision to introduce a broad VAT, together with adherence to the policies o f the East African Community (EAC), w i l l lead to broad changes in the tax regime. Under the assumption o f a broad- based VAT introduced at 17 percent and without Burundi E A C adherence, it has been estimated that this w i l l lead to a reduction in tax revenue o f FBu 1.6 billion, equivalent to 0.16 percent o f GDP. However, the combined effect o f the introduction o f VAT and joining E A C i s expected to have a large effect o n overall receipts, which are expected to decline by FBu 15 billion, equivalent to 1.53 percent o f GDP. This highlights the importance o f introducing a broad-based VAT regime, reducing import exemptions, and improving tax administration. ’.Gerard Chambas and Bertrand Laporte, Introduction de la TVA au Burundi: Etude prdalable de 1’impact budgdtaire, dconomique et social (draft), Centre d’Etudes et de Recherches sur l e DCveloppement International (CERDI), April 2007. Government Expenditure 1.11 Government expenditure has been rising rapidly in recent years, from 27.2 percent of GDP in 2001 to 38.4 percent of GDP in 2006 (Chart 1.3).15 This i s a very high share o f public expenditure, and considerably higher than the figure for a number o f other sub-Saharan African economies (Table 1.4). Recurrent expenditure accounts for 60 percent o f total expenditure and net lending, although its share has risen only slightly over the last six years. The wage bill has been rising over time, representing nearly 10 percent o f GDP in 2006. The rise since 2005 reflects the hiring o f new teachers (primary school enrollment has doubled since primary school fees were abolished in September 2005) and a 15 percent salary increase in 2006, which was the first increase since 2001. Capital expenditure has doubled from 6.4 percent o f GDP in 2001 to 12.1 percent in 2006, as a result o f a steep rise in foreign grants. 1.12 Government expenditure has been rising much more slowly, if military expenditure, exceptional expenditures, and externally financed capital expenditure are excluded. Excluding these categories, public expenditure has risen f r o m 16.2 percent o f GDP in 2001 to 22 percent in 2006 (Table 1S ) . In large measure, the significant rise in public expenditure reflects a rise in foreign-financed capital expenditure, which has been rising from a l o w base o f 3 percent o f GDP since the start o f the decade. ’’These figures are based on IMF data, which include domestically and externally financed expenditures. 11 I Chart 1.3: Real Total Public Spending, 2001-2006 I Table 1.4: Public Expenditure and Net (in FBu billions, unless otherwise indicated) Lending- A Cross Country Comparison (in percentage of GDP), 2006 estimates 4w Percentage 350 Burundi 38.4 300 Liberia 17.1 250 Rwanda 27.8 2w Sierra Leone 21.4 150 100 Tanzania 28.3 50 0 Uganda 21.5 2w2 2W3 ZOW 2005 2ooB I' I Source: IMF. I Source: IMF. Table 1.5: Composition of Government Expenditure (in percent of GDP) 2001 2002 2003 2004 2005 2006 Expenditure and net lending 27.2 25.9 34.9 38.9 36.8 38.4 Current expenditure 21.6 20.4 22.0 21.8 23.3 23.0 Salaries 7.3 7.9 8.3 7.8 8.4 9.8 Civilian 3.9 4.0 4.8 4.6 4.9 5.9 Military 3.4 3.8 3.6 3.2 2.8 2.3 National Police 0.8 1.5 Goods and services 8.0 6.6 7.4 7.2 7.6 6.3 Civilian 3.3 3.3 3.6 3.8 3.1 2.7 Military 4.7 3.3 3.7 3.4 3.4 2.3 National Police 1.1 1.2 Transfers and subsidies 2.9 2.8 2.5 3.5 3.5 4.1 Interest payments 3.4 3.2 3.8 3.3 3.7 2.8 Domestic 1.7 1.3 2.1 1.9 2.3 1.6 Foreign 1.6 1.9 1.8 1.4 1.5 1.2 Exceptional expenditure 0.0 0.0 0.0 1.4 3.9 3.5 DDR 0.0 0.0 0.0 1.o 1.o 3.5 Elections 0.0 0.0 0.0 0.4 2.8 0.0 Capital expenditure 6.4 5.7 13.2 16.0 9.8 12.1 Domestic resources 3.4 1.1 4.1 4.8 2.3 3.8 External resources 3.0 4.6 9.1 11.2 7.5 8.3 Net lending -0.7 -0.2 -0.3 -0.3 -0.2 -0.2 Memorandum item Total expenditure and net lending Excluding military expenditure 19.2 18.8 27.6 32.3 30.5 33.8 Excluding military expenditure, exceptional expenditure and externally financed capital Expenditure 16.2 14.1 18.5 19.7 19.2 22.0 Sources: MEFDC, IMF. 12 1.13 Rising government expenditure in 2006 also reflects, in part, increased pro-poor social expenditure, which i s being financed by the enhanced HIPC initiative. Improved payroll management and continued demobilization o f ex-combatants will help to contain expenditure on salaries in the medium term, but controlling salary rises will also be crucial. The share o f expenditure allocated to social spending has been increasing over time, reflecting the Government’s commitment to increase pro-poor spending. Fiscal Deficit 1.14 Burundi’s fiscal deficit on a commitment basis averaged 4.3 percent of GDP in 2001- 06 (Table 1.6). Excluding grants, the fiscal deficit averaged 13.8 percent over the same period, and has been rising in recent years, reaching 20.4 percent in 2006. This reflects an expansionary fiscal policy, which has been financed by increased external resources. Over 2001-06, the share o f the fiscal deficit covered by external financing has been rising, from a l o w o f 27 percent in 2001 to a high o f 93 percent in 2005. Shortfalls in external budget support in 2006, reflecting in part donors’ governance concerns, were offset by spending cuts in non-wage spending and increased domestic borrowing. Securing sufficient and timely budget support i s critical to ensure financing o f expenditures and to prevent reductions in planned discretionary expenditures. 1.15 Government-led improvements in donor coordination and aid management may help, but the Government also needs to monitor progress on key structural reform criteria and other conditionalities to ensure the timely release o f budget support. This can help reduce the need to turn to domestic bank financing, which in the past has been a source o f inflationary pressures. At the same time, there i s a need for donors to disburse aid on a more predictable basis. 1.16 Over 2004-06,50 percent of external finance received by Burundi was in the form of grants. The grant element has been rising in recent years, reaching 63 percent in 2006, with debt relief providing 19 percent, and the remaining 17 percent o f external finance coming from loans. The increasing share o f grants in external finance i s in line with the debt-sustainability analysis o f the HIPC Decision Point, which stresses that the provision o f external financial assistance needs to be heavily weighted towards grants, given the country’s vulnerability to external shocks and volatile export performance. Table 1.6: Central Government Fiscal Deficit, 2001-2006 (in FBu billions, unless otherwise stated) 2001 2002 2003 2004 2005 2006 Overall balance (commitment basis) Including grants -28.4 -7.8 -40.1 -39 -51.6 -20.3 % of GDP -5.2 -1.3 -6.2 -5.2 -6.0 -2.1 Excluding grants -39.6 -33 -88.6 -144.1 -144.5 -200.9 % o f GDP -7.2 -5.6 -13.7 -19.2 -16.8 -20.4 Change in arrears (reduction -) 21.3 9.4 -2.2 -58.5 -10.8 -21.6 External (current interest) 18.5 6.6 4.2 -49 -10.7 -0.2 Domestic 2.8 2.8 -6.4 -9.5 -0. I -21.4 Overall balance (cash basis) Including grants -7.1 1.6 -42.3 -97.5 -62.4 -41.9 Excluding grants -18.3 -23.6 -90.8 -202.6 -155.3 -224.5 Financing 18.3 23.6 90.9 210.3 165.8 224.5 External 5.0 41.3 71.8 143.7 154.9 184.1 Privatization proceeds 0.0 0.0 0.0 0.0 0.4 4.7 Domestic 13.3 -17.7 19.1 66.7 10.4 35.7 Banking sector 23.4 -13.0 4.9 60.5 16.4 33.7 Nonbank sector -10.1 -4.7 14.2 6.2 -6.0 2.0 Unidentified financing 0.0 0.0 0.0 -7.6 -10.4 0.0 Source: Central bank, IMF. 13 Quasi-Fiscal Deficit 1.17 Looking at the fiscal deficit to assess fiscal sustainability i s an imperfect exercise, as it ignores the impact o f non-financial public enterprises, public sector banks, the central bank, and other obligations that are outside the budget. A more appropriate concept i s the quasi-fiscal deficit, which refers t o obligations o f the public sector arising from operations that have only indirect effects o n the budget. These activities could have significant macroeconomic implications, as they distort the Government’s real fiscal position and may generate substantial contingent liabilities. In the case o f Burundi, there i s limited data o n the transfers from the Treasury to state-owned enterprises (SOEs), and there i s a clear need for these to be made explicit in the budget. The high level o f SOE indebtedness represents a potential risk to the Government in case o f bankruptcy. At end-2005, these debts totaled FBu145 billion (US$138 million). T h i s highlights the fact that the quasi-fiscal deficit i s likely to be significantly higher than the fiscal deficit. The External Sector 1.18 Burundi has a three-year transitional period, from 2007 to 2009, to implement the East African Community (EAC) trade area common external tariff. Membership i s expected to boost trade in the medium term, while stimulating competition in key areas. Burundi has revised its tariff rates to align i t to the Common Market for Eastern and Southern Africa (COMESA) external tariff. As a result, there has been a significant reduction in top import tariff rates, from 100 percent in 2003 to 30 percent in 2005. However, tariff rates for equipment are still relatively high (10 percent), which i s a concern for investment and private sector development. Efforts to identify the revenue impact o f joining E A C are currently underway, with W o r l d Bank technical assistance. 1.19 Burundi’s trade and Chart 1.4: Balance o f Payments, 2001-2006 current account deficits are (in percent o f GDP) structural in nature (Chart 1.4). 2001 2002 2003 2004 2005 2006 The trade deficit reflects the 0 country’s narrow range o f exports, -5 which s t i l l consist almost entirely -10 o f tea and coffee, and its -15 dependence o n the import o f -20 capital goods and fuel. Export -25 earnings in any given year largely -30 depend o n the size and quality o f -35 the tea and coffee harvests and on -40 international price movements for ITrade b l r c e t Current account o Current acccunt (exclgrants) both products, which suffer from considerable fluctuations. Source: IMF. 1.20 In order to enhance the value-added derived from the coffee sub-sector; Burundi must make important reforms t o develop the capacity to participate in rapidly changing global coffee markets. Significant parts o f coffee production could reach fair trade and other niche markets, which highly value production from fragile countries. However, for this to occur, the 14 competitiveness o f coffee must be radically improved, by promoting best practices to manage costs and define sales strategies. Areas o f improvement include increasing yields to produce larger volumes, reducing costs, and increasing the reliability o f production. 1.21 Burundi’s exports have nearly doubled from US$32.5 million in 2001 to US$60.3 million in 2006, but with coffee still representing 68 percent o f merchandise exports, movements in exports reflect volume and price fluctuations in this sub-sector. Imports have risen more rapidly in the same period, from US$89.9 million to US$256 million, reflecting a strong increase in imports for foreign-financed reconstruction and rehabilitation activities. Likewise, Burundi has consistently run large current account deficits, reflecting the country’s import o f capital goods financed by aid flows. The current account deficit (including grants) rose to 13.8 percent o f GDP in 2006, up from 1.2 percent in 2005, reflecting a sharp rise in the services deficit (Chart 1.4). Burundi’s services deficit i s expected to widen in the near term, as increased trade volumes, boosted in part by the country’s accession to the EAC, lead to growing freight costs. External Assistance Chart 1.5: Evolution of Net ODA to Burundi Chart 1.6: Net ODA to Burundi, 1982-2005 (in U S $ million) (in US$ million) 400 400 350 350 300 300 250 250 200 200 150 150 100 100 50 50 0 0 2001 2002 2003 2004 2005 1982 1987 1992 1997 2002 Grants Total -Current prices ~ 2005 constant prices Source: OECD Source: OECD. 1.22 Aid flows, provided by a limited number of donors, have tended to be volatile, and declined sharply in the mid-1990s as a direct consequence o f the political situation. After falling to US$56 million in 1997, net overseas development assistance has been rising in recent years, reaching US$365 million in 2005 (Charts 1.5 and 1.6). In constant prices, net aid flows in 2005 recovered to pre-crisis levels. During 2001-2005, Burundi received an average o f US$253 million o f official development assistance (ODA) per year, surpassing the US$300 million bar in 2004 and 2005. More than 90 percent o f that assistance has been in the form o f grants, and the rest has been in the form o f loans on concessional terms. Bilateral and multilateral aid flows were about the same magnitude over 2001-05. Multilateral assistance was largely provided by the EU, which i s the largest donor in the country, accounting for 50 percent o f multilateral ODA, followed by IDA, which accounted for 21 percent. The largest bilateral donor and the second 15 largest donor i s the United States, which accounted for 14 percent o f net ODA. Other important bilateral donors are France, the Netherlands, and Belgium. 1.23 Burundi suffers from a heavy Chart 1.7: External Debt external debt burden, with debt (in percent of GNI) totaling US$1.56 billion in 2006 (Chart 250 I I 1.7). However, the country was granted access to debt relief under the enhanced 200 . H I P C Initiative in August 2005, This will reduce debt by more than 90 percent in 150. net present value terms, and scheduled debt service by some US$30-40 million 100 - per year for the next 30 years. However, 50 . Burundi remains at high risk o f debt distress, even after the full delivery o f 07 ~ 7 debt relief, which i s expected to occur in 2008. The results o f the baseline scenario o f the debt-sustainability analysis Source: IMF. indicate that Burundi’s NPV o f debt-to-export ratio remains above the indicative threshold o f 100 percent until 2013, before falling. The ratio o f debt service t o exports in 2005 was 47.1 percent. Burundi will be eligible for further debt relief under the Multilateral Debt Relief Initiative (MDRI) after reaching HIPC Completion Point. However, the additional debt relief under MDRI i s expected to be limited. The baseline scenario assumes GDP growth o f 6.5 percent over the medium term, which i s high relative to past performance, suggesting that external debt could remain a significant burden in the absence o f a sustained rise in economic performance. Debt sustainability indicators could deteriorate if real GDP and export growth are lower. Foreign Reserves and the Exchange Rate 1.24 Burundi’s foreign exchange position has fluctuated widely in the last decade, reaching a l o w o f US$18 m i l l i o n in 2001, with the fluctuations reflecting export performance and volumes o f aid flows. Increased macroeconomic stability and increased donor funding have resulted in a build-up o f foreign reserves, which at the end o f 2006 stood at US$13 1 million (excluding gold), equivalent to about 3.2 months’ o f import cover. Since 2000, the exchange rate has been set by the market in public auctions rather than by the central bank. The difference between the official and parallel rates, equal to 47 percent when the auctions began, has fallen significantly. Increased aid flows from donors have contributed to a further narrowing o f the gap, as has the legalization o f private foreign exchange bureaus. 1.25 The Burundi franc has enjoyed an unusual level of stability over the last three years, owing to the weakness in the U S dollar, donor inflows, and a modest improvement in export performance. There was a nominal appreciation o f 1.7 percent in 2005 and 4.9 percent in 2006 vis-&vis the U S dollar, whereas historically the Burundi franc has tended t o depreciate. The central bank’s interventions in the foreign exchange market are for the purpose o f achieving i t s international reserves objectives and avoiding excessive exchange rate volatility. 16 C. Macroeconomic Outlook and Financing Requirements, 2007-2010 Economic Performance for 2007 1.26 The Government’s program for 2007 was to continue macroeconomic stabilization, overcome delays in implementing some structural reforms, and continue strengthening public finance management. According to preliminary estimates, real economic growth declined to 3.6 percent from 5.1 percent in 2006, despite increased donor-financed capital expenditure, reflecting primarily a poor agricultural harvest. Inflation (measured by the average change in the consumer price index, CPI) i s estimated to rise to 8.3 percent (above expectations o f 6.8 percent), after a very l o w inflation rate o f 2.8 percent in 2006 due to good agricultural performance. Several factors can explain the rise in inflation, among others, the price increase decision o f end- 2007 by the authorities on certain products (e.g., sugar), the rise in international food prices, as well as the implications on Burundi’s imports o f the recent Kenya’s post-election violence. The current account deficit i s estimated to increase from 14.5 percent in 2006 to 16.0 percent in 2007, as a result o f a decrease in exports coupled with an increase in imports. 1.27 The 2007 Budget Law continued with reorienting expenditure toward poverty reduction while containing the budget deficit (excluding grants). The budget deficit (excluding grants) i s estimated to rise to 20.1 percent o f GDP in 2007, from 19.3 percent o f GDP in 2006. It i s also estimated that total revenues will decline from 18.9 percent o f GDP in 2006 to 18.7 percent o f GDP in 2007. Furthermore, public spending i s expected to increase from 38.2 percent in 2006 to 38.7 percent in 2007. T h i s increase i s mainly the result o f an increase in recurrent expenditure. One main component o f the increase i s the wage bill. Wages and salaries increased from FBu 93.9 billion (9.9 percent o f GDP) to FBu 114.0 billion (10.8 percent o f GDP) in 2007. Control o f the wage bill must become an important Government priority. In addition to wage restraint, this will depend on measures aimed at strengthening management o f the wage bill and continued demobilization o f ex-combatants. Demobilization and security sector reforms will continue to reorient expenditure away from security and toward social and poverty-reducing sectors. Medium-term Macroeconomic Framework 1.28 The projections for the medium-term development of the economy o f Burundi have been derived by means of a revised minimum standard model-extended (RMSM-X). With a base year o f 2006, the underlying assumptions have been made taking the recent development o f the economy, as well as all other relevant information on i t s further progress, into account. A summary o f the assumptions i s given in Table 1.7. 1.29 The medium-term economic framework assumes an acceleration o f real GDP growth from 3.6 in 2007 to 4.5 percent in 2008, and to 5 percent for the remaining period until 2011 (Table 1.7). This i s the average growth rate, but yearly growth i s likely to exhibit fluctuations around this trend given the economy’s susceptibility to weather conditions. The main driving force behind the real GDP growth rate remains the primary sector, which contributed, on average, 49 percent to GDP between 2000 and 2006; although a slow shift to the secondary and tertiary sectors can be observed. After a disappointing year with no growth in 2007, the primary sector i s expected to recover and grow modestly in 2008 (below i t s 2006 level),partly reflecting 17 the continued rise in the fuel price which affects transportation costs and the price o f fertilizers, and to further increase in 2009-201 1. The main contribution t o the primary sector comes from food crops (on average, 80 percent between 2000 and 2006), while the share o f export crops i s s t i l l small. Food crops are expected to recover from 2008 onwards. A strong reform process in the agricultural sector will be necessary to achieve this improved performance. Table 1.7: Main Assumptions for Medium-Term Macroeconomic Projections (in units indicated) 2000-06 2007 2008 2009 2010 2011 average Prel. Est. Projections Real GDP growth rate (%) 2.2 3.6 4.5 5.0 5.0 5.0 Inflation (GDP-deflator) (%) 8.7 8.2 18.5 10.0 8.7 6.6 Exports (volume) Coffee -0.1 -6.9 50.0 15.0 15.0 10.0 Tea 0.4 8.9 10.0 10.0 10.0 10.0 Hides and skins 36.2 129.9 5.0 6.0 6.0 6.0 Other exports 91.3 30.0 20.0 15.0 15.0 10.0 Prices Food 4.8 25.6 38.0 -5.5 -5.0 -2.4 POL and other energy 19.6 10.8 52.4 -2.3 -7.1 -5.0 Coffee 7.1 -18.7 5.0 0.5 0.5 0.5 Tea -1.o -15.7 1.o 1.o 1.5 1.5 Manufacturing- unit value index 1.4 3.9 6.8 0.7 1.4 0.4 Exchange rate 9.5 5.6 5.9 -3.0 2.5 4.6 Source: World Bank, IMF, own calculations. 1.30 It i s assumed that the secondary sector will grow only moderately in 2008 (below i t s 2007 performance) t o reflect the increase in the cost o f production resulting from the rise in the price o f food price and fuel. A higher increase in the secondary sector from 2009 onwards i s due to higher value-added from ago-processing. It can be expected that the coffee washing stations will become more efficient, leading to higher output and quality o f coffee. Improvements in the mining sector as well as reforms in the petroleum sector will also lead to a higher growth rate o f the secondary sector. Moreover, i t i s expected that future growth will revolve around small and medium enterprises (SMEs), which appear to be the most vibrant agents o f pro-poor growth. 1.3 1 The expected increase in the tertiary sector will become more apparent, as it i s projected t o grow at a higher rate than in past years. In effect, the service sector will accompany growth in the primary and secondary sectors (trade, transport, and telecom). Additional stimulus i s expected fkom strong public expenditure in the priority sectors (with released HIPC expenditure), renewed investment in the telecom sector following the expected privatization o f ONATEL, and increased dynamism o f the financial sector with the recent reopening o f the headquarter o f the Eastern and Southern African Trade and Development Bank (Banque de la Zone d 'Echanges Prdfdrentiels, ZEP) in Bujumbura. 1.32 Annual average inflation (measured by the GDP deflator) i s projected to subsequently decline from the high level o f 18.5 percent in 2008 to 6.6 percent in 2011 (Table 1.7). According to the World Bank Prospects Group, food and energy prices will eventually decline although not to their initial level. Improved liquidity management by the central bank and limited financing o f government activities by the central bank could also explain the projected decline in inflation. 18 1.33 The current account deficit, excluding official transfers as percentage o f GDP, narrows from 2008 onwards, but remains substantial (Table 1.8). Without official transfers, the current account deficit, given as a percentage o f GDP, i s projected to decrease over time from 36.3 percent in 2006 to 30.3 percent in 201 1. The improvement stems from an increase in exports due to the success o f the reform process, especially in the coffee sector, as well as developments in the mining sector. Combined with a slower increase in imports as the main reconstruction and rehabilitation efforts have taken place, a lower deficit can be recorded. Taking official transfers into account, the current account deficit increased from 14.5 percent o f GDP in 2006 to 16.0 percent o f GDP in 2007 and i s projected to fluctuate around an average o f 18.7 percent in the following years. Table 1.8: Balance of Payments Projections (in units indicated) 2005 2006 2000-06 2007 2008 2009 2010 2011 Actual average Prel. Projections lmports f.0.b. annual percentage change 27.6 28.7 14.6 20.6 26.8 8.4 4.0 3.5 percent o f GDP 23.7 26.6 21.0 30.3 32.8 29.9 27.9 27.0 Exports f.0.b. annual percentage change 19.5 2.6 2.7 -14.3 40.6 14.2 14.9 11.8 percent o f GDP 7.1 6.4 6.6 5.2 6.2 5.9 6.1 6.4 Current account balance incl. official transfers (percent of GDP) -1.2 -14.5 -6.4 -16.0 -19.1 -20.1 -17.8 -17.9 excl. official transfers (percent o f GDP) -29.1 -36.3 -22.5 -37.7 -36.3 -34.2 -31.9 -30.3 Terms o f trade (annual percentage change) 10.9 -4.9 -1.4 -16.0 -11.4 5.3 6.4 3.7 Source: IMF, World Bank, own calculations. Government Finance Outlook 1.34 The outlook for the Government finance statistics, given in Table 1.9, shows a continued fiscal deficit over 2008 to 2011 if grants are included, except for 2009 because o f large payments due to reaching the HIPC completion point. Recurrent expenditure increases in 2007, mainly due to a rise in wages and salaries from 9.9 percent o f GDP in 2006 to 10.8 percent o f GDP in 2007. Mostly reflecting the delays in the implementation o f demobilization program and the need to hire more teachers, wages and salaries will only slightly decrease at 10.7 percent o f GDP in 2008, whereafter they are expected to progressively decline to 9.6 percent o f GDP by 201 1, driven by demobilization and wage restraint. I t i s important to note, however, that the post-conflict nature o f the country requires a higher level of police and army presence. Total revenue as a percentage o f GDP i s projected to remain fairly constant, averaging 19.1 percent over 2008 to 201 1. The fiscal deficit, excluding grants, i s projected to decline from 20.1 percent o f GDP in 2007 to 17.5 percent o f GDP in 201 1, while the deficit including grants i s expected to decrease slightly between 2008 and 2011 (if 2009 i s not considered) as well. The debt stock i s expected to fall substantially after 2009 due to Burundi reaching the HIPC Completion Point in the first quarter o f 2009. 1.35 Aid pledges for 2007 are expected to meet projected requirements and consist largely of external grants. External grants for 2007 are estimated at FBu 221.3 billion, o f which 45 percent i s budget support (including special programs for demobilization and reintegration o f 19 ex-combatants), 37 percent project grants and 18 percent HIPC relief. The financing gap in the following years i s likely to be covered by additional bilateral and multilateral assistance, provided Burundi stays o n track with its reform agenda. Table 1.9: Government Finance Statistics Projections (percent o f GDP, unless otherwise indicated) 2005 2006 2007 2008 2009 2010 2011 Actual Prel. Projections Total revenue (excl. grants) 20.0 18.9 18.7 19.1 19.1 19.1 19.2 Tax revenue 18.5 17.3 17.2 17.7 17.7 17.8 17.9 Non-tax revenue 1.5 1.6 1.4 1.4 1.4 1.4 1.4 Expenditure and net lending 36.8 38.2 38.7 43.8 40.3 39.0 36.7 Recurrent expenditure 23.3 23.4 24.9 26.5 24.3 23.2 22.6 Wages and salaries 8.4 9.9 10.8 10.7 10.4 9.9 9.6 Transfers and subsidies 3.5 4.2 4.4 4.7 4.6 4.5 4.5 Interest payments 3.7 2.6 3.0 3.9 2.1 1.7 1.3 Capital expenditure 13.6 14.8 13.9 17.3 16.1 15.8 14.1 Primary fiscal deficit" -1.7 -7.5 -5.6 -8.3 -8.5 -6.3 -7.3 Overall fiscal deficit (excl. grants) -16.8 -19.3 -20.1 -24.8 -21.3 -19.9 -17.5 Overall fiscal deficit (incl. grants) -6.5 -1.4 0.8 -2.2 7.0 -1.1 0.1 Financing (in FBu billions) Primary fiscal deficit* -14.6 -71.1 -59.6 -108.2 -128.3 -108.7 -142.3 Overall fiscal deficit (excl. grants) -144.3 -182.2 -212.8 -324.8 -322.2 -343.9 -338.3 External grants 100.8 168.9 221.3 295.5 1457.0 325.5 341.1 External borrowing 22.1 7 .O 30.3 22.8 18.8 22.5 Adjustment to scheduled interest payments 4.8 18.8 4.6 -1149.0 0.0 0.0 Domestic borrowing -13.5 -34.3 -5.6 -8.6 -0.4 -25.2 Debt (in FBu billions) Total debt stock 1431.3 1544.6 1680.9 1665.0 1686.7 686.5 720.4 Domestic debt 193.1 243.3 209.1 203.5 194.9 194.4 169.2 Foreign debt 1238.2 1301.3 1471.9 1461.5 1491.8 492.1 551.2 Source: IMF, World Bank, own calculations; (*) excluding interest payments and externally financed projects. 1.36 The reliance of the country on external assistance i s reflected b y the high inflow of foreign savings. These results have to be seen in the post-conflict context o f the country, and similar results have been observed for countries in comparable situations. The need to reconstruct and rehabilitate the economy leads to levels o f spending in the short and medium term that surpass national savings. T h i s can also be seen in the projected high investment rate over the coming years (Table 1.10). The investment rate increases substantially fkom 16.3 percent o f GDP in 2006 to an estimated 19.3 percent o f GDP in 2007 and i s projected to rise further to 27.5 percent o f GDP in 2010, due to increased reconstruction and rehabilitation efforts (including the school construction program). 20 Table 1.10: National Accounts Projections (percent o f GDP, constant prices) 2005 2006 2000-06 2007 2008 2009 2010 2011 Actual Averwe Prel. Proiections Consumption 108.4 110.5 104.5 116.8 109.1 111.1 111.6 111.3 Private 82.4 86.4 81.3 99.8 92.1 93.9 94.6 94.3 Public 26.0 24.0 23.2 17.0 17.0 17.3 17.1 17.1 Investment 10.5 16.3 9.5 19.3 25.0 26.5 27.5 27.5 Private 4.0 8.0 3.1 7.1 9.8 10.5 11.5 11.5 Public 6.5 8.3 6.4 12.2 15.2 16.0 16.0 16.0 Source: IMF, World Bank, own calculations Risks to the Macroeconomic Framework 1.37 The baseline scenario presented above i s subject to a number o f risks. First, Burundi’s economic growth i s subject to variable growth performance o f the agricultural sector. At times, annual growth has been significantly below the projections o f the medium-term economic framework. Second, Burundi i s highly dependent o n aid inflows, and any difficulty in maintaining donor support or delays in mobilizing such support can derail fiscal performance and reduce overall economic growth. Third, delays in the implementation o f structural reforms, particularly in the coffee sector, as well as reforms o f the business environment, would adversely affect private sector investment. Fourth, a mishandled integration into the E A C could also jeopardize the macroeconomic framework. Fifth, Burundi i s a post-conflict country with a long history o f ethnic conflict that has had deleterious effects o n economic growth and poverty reduction. Negotiations with the last rebel group, FNL-PALIPEHUTU, are ongoing; until a negotiated agreement i s reached, this will represent an important security risk. D. Solvency and Fiscal Sustainability’6 1.38 I s Burundi’s current fiscal stance sustainable? If not, how much additional fiscal reduction i s needed? H o w will these fiscal adjustments be affected by external shocks? T h i s section addresses these questions by summarizing the analytical results o f solvency and sustainability analysis conducted on Burundi’s fiscal deficit. 1.39 The conventional practice o f scrutinizing only the budget deficit to assess fiscal efforts and the sustainability o f a fiscal stance i s often inadequate, because fiscal policy i s an integral part o f the overall policy framework and cannot be assessed independently. Solvency and sustainability analysis addresses these shortcomings o f conventional analysis by incorporating the debt dynamics and other macroeconomic targets o f growth, inflation, and foreign and domestic debt. According to this analysis, the public sector i s solvent when i t s discounted revenues are sufficient to cover the public sector debt (including external and domestic debt); and the debt i s sustainable if it can be serviced without incurring new debt. In order to evaluate Burundi’s fiscal stance, the level o f the primary deficit that would make the public sector solvent was first l6T h i s section follows Dinh (1999). 21 determined, given the stock o f public debt at the end o f the previous period, the real interest rate, and the real revenue growth rate.” 1.40 With a level of primary budget deficit (excluding grants for budgetary assistance, foreign-financed capital expenditure, and interest payments) averaging 10 percent of GDP in 2001-2006, Burundi’s fiscal stance appears unsustainable, since this ratio i s higher than the difference between the targeted real interest rate (10 percent) and the average real GDP growth rate (3 percent). For fiscal sustainability, the primary deficit needed to achieve debt sustainability (that is, to maintain the ratio o f debt to GDP at the same level) i s first calculated. In a second step, the required adjustment i s calculated by comparing this sustainable primary deficit with the actual (or expected) primary deficit. All variables and resulting measures (in percent o f GDP) are presented in Table 1.11 for 2001 and 2006. An average for the period 2001-2006 i s also provided. Table 1.11: Public Sector Solvency and Sustainability in Burundi Average 2007 2008 2009 2010 2001-06 Public Sector Solvency Primary Balance (LCU billions) -81.6 -180.5 -272.4 -300.1 -306.6 Primary Balance/GDP -10.4% -17.0% -20.8% -19.8% -17.7% Real GDP Growth 3 .O% 3.6% 4.5% 5.0% 5.0% Difference between 10% and Real GDP Growth 7.0% 6.4% 5.5% 5.0% 5.0% Public Sector Solvency Index 0.4% 15.6% -13.7% -9.4% -7.5% Public Sector Solvency Adjustment (% o f GDP) 11% 32.6% 7.1% 10.4% 10.2% Public Sector Sustainability Exports/GDP 9% 8.9% 9.9% 9.4% 9.9% DODIGDP 188.8% 155.6% 136.5% 128.6% 57.9% Devaluation Rate 6.7% 1.5% -0.4% 2.1% 2.0% Capital Grants/GDP 5.0% 11.5% 12.4% 87.9% 9.4% Sustainable Primary Balance (- =deficit) 3.0% 15.9% -15.5% -0.1% -6.4% Actual Primary Balance (- =deficit) -10.4% -17.0% -20.8% -19.8% -17.7% Public Sector Sustainability Adjustment (YOof GDP) 13.0% 32.9% 5.3% 19.7% 11.3% Memorandum Items: Target InternationalInterest Rate 10.0% 10.0% 10.0% 10.0% 10.0% Target Domestic Interest Rate 10.0% 10.0% 10.0% 10.0% 10.0% Domestic Debt/GDP 20.8% 35.1% 23.0% 19.7% 16.6% Domestic Revenue/GDP 20.1% 18.7% 19.0% 19.1% 19.1% Source: World Bank staff calculations. 1.41 The results o f the analysis reveal that, as o f the end o f 2007, Burundi’s fiscal balance remained far from the level required for solvency and sustainability. If grants were excluded and Burundi was to borrow at non-concessional rates, i t would have needed to adjust i t s deficit by 32.6 percent o f GDP.’* With respect to fiscal sustainability, a similar adjustment o f 34 percent i s ” Both the real interest rate and the real revenue growth rate were assumed to be constant throughout the period o f the analysis. The fiscal sustainability analysis was also performed under alternative hypotheses: (a) reduction o f the targeted real domestic interest rate from 10 percent to 5 percent; (b) increase o f the targeted real domestic interest rate from 10 percent to 15 percent; and (c) reduction o f real export growth projections (2008-10) from 21-24 percent to 10.8 percent, based on an historical average. When the above scenarios were considered in the analysis, our earlier findings did not significantly change. However, the public sector solvency results seem to be modestly affected by the real export growth scenario. 22 required for 2007. Past international experience indicates that an annual fiscal adjustment o f more than 3 percent would be difficult for any country to achieve, so it would take about 11 years for Burundi to become fiscally sustainable and solvent. I t i s neither feasible nor desirable for Burundi to bear the burden o f an adjustment o f such magnitude by expenditure cuts andor revenue- enhancing measures. The country will depend o n concessional financing in the f o r m o f g r a n d g a n d o r highly concessionary external borrowing for the foreseeable future. 1.42 The analysis also reveals that despite the complete debt relief that Burundi i s expected to be entitled to in 2008 in the context o f the HIPC Initiative, the country will remain fiscally unsustainable, with an adjustment o f 24 percent required (as o f the end o f 2008). Meanwhile, the results suggest that the country’s public sector solvency will improve with a fiscal adjustment o f only 7 percent o f GDP, which will require creating fiscal space. The key findings o f the fiscal sustainability analysis presented in this section are consistent with those o f the latest World Bank- IMF debt sustainability analysis (see Box. 1.2 below). The next chapter explores potential sources o f fiscal space. Box 1.2: Debt Sustainability Analysis The latest debt sustainability analysis (DSA), dated December 2007, assesses Burundi’s external public debt dynamics using the forward-looking debt sustainability framework for low-income countries @IC). At end-2006, Burundi’s stock o f public and publicly guaranteed (PPG) external debt was US$1,464 million (US$1,5 17 .million including arrears), US$5 1 million higher than at end-2005. The baseline scenario assumes continued interim HIPC debt relief through the medium term. An alternative scenario models the HIPC Completion Point and the delivery o f additional MDRI debt relief in September 2008. Under the baseline scenario, Burundi would be in debt distress, as the present value (NPV) o f PPG external debt at end-2006 was 88 percent o f GDP, 841 percent o f exports, and 462 percent o f government revenues. As a result, most debt burden indicators substantially exceed their policy-based thresholds under the baseline scenario. Under the alternative scenario, the risk o f debt distress i s significantly reduced but remains elevated. Burundi i s currently in debt distress. Even after reaching the HIPC Completion Point and benefiting from debt relief under the enhanced HIPC and MDRI initiatives, Burundi remains highly vulnerable to shocks that could lead to lower exports and economic growth. For Burundi to lower i t s high debt stock ratios with respect to exports, and to maintain the other debt indicators below their respective thresholds, particular emphasis needs to be placed on five aspects o f future economic performance. First, the authorities need to implement a strong and sustained reform effort, especially to develop a diversified export base, in order to support robust growth in exports and GDP. Second, the authorities need to implement prudent fiscal and monetary policies to ensure fiscal discipline and macroeconomic stability. Third, together with prudent debt management, future financing needs to be in the form o f grants and highly concessional loans in the short and medium term. Fourth, attention should be given to strengthening debt management policies and institutions. Fifth, accelerated efforts are needed to meet the HIPC Completion Point triggers as soon as possible. Capital grants accounted for an average o f 4.5 percent o f GDP over the 2001-2006 period. 23 CHAPTER 2. COMPOSITION OF PUBLIC EXPENDITURE AND KEY SOURCE OF FISCAL PRESSURE 2.1 The objective o f this chapter i s two-fold: first, i t provides a discussion on the economic and functional composition o f public expenditure in Burundi. Second, it takes a closer look at the wage bill - the cost o f public employment - as a key driver o f fiscal pressure in Burundi. 2.2 The lack o f reliable data remains a major obstacle to a comprehensive analysis o f economic and public finance developments. In effect, (a) no national accounts have been produced since 1998; (b) the coverage o f budget execution reports (Reddition des Comptes) i s s t i l l incomplete, as they provide information on the execution o f public expenditure according to the economic classification and not the functional composition; (c) data on the wage bill do not include all the public personnel expenditures; and finally, (d) Burundi’s budget monitoring and execution system does not provide data on donor-financed expenditures, which constitute 43 percent o f total public expenditures. 2.3 The chapter i s organized as follows. Section A reviews the economic and functional composition o f public expenditure across key spending categories in 2001-2006. The most important aspect o f that analysis i s to compare the structure o f government expenditures and donor programs with the broad policy objectives o f the recently approved PRSP. Section B discusses two factors that influence the structure o f public expenditure. Section C provides an in- depth discussion o f the wage bill as a key source o f fiscal pressure in Burundi. The last section, Section D, models the impact o f alternative public spending scenarios on economic growth and poverty. A. Analysis o f Public Expenditure by Economic and Functional Classifications Economic Classification 2.4 The composition of government expenditure i s dominated by a large share of recurrent expenditures (wages and salaries, other goods and services, interest payments, and transfers and subsidies). They accounted for more than 66 percent o f total expenditure over 2001-2006, leaving about 30 percent for capital expenditure and less than 4 percent for exceptional expenditures (DDR and elections).*’ However, in recent years, mainly due to an increase in foreign-financed capital expenditure, there has been a moderate and gradual shift toward more capital expenditure in Burundi’s budget. The share o f capital expenditure in total expenditure increased slightly over 2001-2006, from 22.9 percent in 2001 to 32.2 percent in 2006. In the meantime, the share o f recurrent spending contracted, from 77.1 percent o f total expenditure in 2001 to 61.3 percent in 2006. Yet, accounting for about 23.5 percent o f GDP, on average, over 2005-06, recurrent expenditure s t i l l dominates the structure o f expenditure, while 2oT h e economic composition o f public spending indicates that, on average over 2001-06, while 37 percent of recurrent expenditures were absorbed by wages and salaries, another 33 percent were consumed by other goods and services. Meanwhile, transfers and subsidies, and interest payments claim the remainder (about 30 percent). 24 capital expenditure (excluding foreign-financed outlays) has been relatively modest, at about 3 percent o f GDP over the same period (see Table 2.1 below). Table 2.1: Economic Distribution o f Public Expenditures, 2001-2006 In percent of total expenditure (net lending) In percent o f GDP Average Average 2001 2002 2003 2004 2005 2006 2001-04 2005-06 Current expenditure 77.1 78.2 62.5 55.6 63.0 61.3 21.5 23.5 Wages & Salaries 26.1 30.3 23.6 19.9 22.7 26.0 7.8 9.2 Goods and services 28.6 25.3 21.0 18.4 20.5 17.7 7.3 7.2 Transfers and subsidies 10.4 10.7 7.1 8.9 9.5 10.9 2.9 3.9 Interestpayments 12.1 12.3 10.8 8.4 10.0 6.8 3.4 3.2 Domestic 6.3 5.0 5.8 4.8 6.1 3.4 1.a 1.8 Foreign 5.9 7.3 5.0 3.6 3.9 3.4 1.7 1.4 Exceptionalexpenditure 0.0 0.0 0.0 3.6 10.5 6.5 0.4 3.7 DDR 0.0 0.0 0.0 2.6 2.8 6.5 0.3 2.3 Elections 0.0 0.0 0.0 1.o 7.8 0.0 0.1 1.4 Capital expenditure 22.9 21.8 37.5 40.8 26.5 32.2 10.4 11.2 Domestic resources 12.1 4.2 11.6 12.2 6.2 9.8 3.4 3.1 External resources 10.7 17.6 25.9 28.6 20.3 22.4 7.0 8.1 Source: IMF. Wages and Salaries 2.5 The past six years were characterized by continued increase in the share o f government wage bill (Table 2.2). The share o f wages and salaries in total current expenditure increased from an average o f 36.5 percent in 2001-04 to an estimated 39.2 percent in 2005-06. This i s the result o f very l o w levels o f current revenue and expenditure. In 2006, total spending o n wages and salaries was 53 Table 2.2: Recurrent spending categories in total recurrent, 2001- percent of central 2006 (in percent) government current revenues Average Average and 42.4 percent of 2001 2002 2003 2004 2005 2006 2001-04 2005-06 Wages& 33.8 38.7 37.7 35.8 36.1 42.4 36.5 39.2 government current Salaries expenditures.2' At about 10 Goods & 37.0 32.4 33.6 33.0 32.6 28.8 34.0 30.7 percent Of GDP in 2o06y the government wage bill i s Ez%s& Subsidies 13.4 13.7 11.4 16.1 15.0 17.8 13.6 16.4 Interest 15.7 15.7 17.3 15.1 15.9 11.0 16.0 13.4 significantly above the levels payments o f 6-7 percent of GDP found Domestic 8.1 6.4 9.3 8.7 9.6 5.5 8.1 7.6 Forei 7.6 9.3 8.0 6.4 6.3 5.5 7.8 5.9 in Other SourcesZEFDC, SIGEFI. countries.22 The increase in *' These ratios are higher than the average for IMF's Poverty Reduction Growth Facility (PRGF)-eligible countries, estimated at about 28 percent and 26 percent, respectively (see Haiti PEMFAR, 2008). However, for many sub- Saharan African countries, wages and salaries typically range between 35 to 60 percent of recurrent expenditures. Based on these figures, Burundi i s less of an outlier in the region. ** I t was estimated that the average central government wage bill represents about 6-7 percent of GDP for PRGF eligible countries (see Haiti PEMFAR, 2008). 25 the wage bill has been particularly pronounced since 2005, reflecting the hiring o f new teachers (primary school enrollment has doubled since primary school fees were abolished in September 2005), and a 15 percent salary increase across the board in 2006. However, this figure substantially underestimates the true level o f the wage bill, since contract workers in health, education, and a few other ministries are paid from resources generated by the sector, and their salaries are not included in the official payroll. Moreover, it does not include other wage bills concealed under transfers to semi-autonomous institutions, such as the University o f Burundi. 2.6 The ability of the Government to attract and retain quality personnel i s hampered by the low level of the public sector salary. As a result, there will be an increasing need to pay higher salaries to key technical and professional staff. A comparison o f public sector salaries to those paid outside the public sector (e.g., NGOs and donors) reveals that the former are significantly lower than the latter. For instance, some donors pay doctors up to ten times as much as the US$60 monthly average for doctors paid by the Government (see Chapter 5 for details). This raises the critical issue o f the need to generate more domestic revenue and/or analyze the options to reduce the size o f the civil service in order to eventually finance a more efficient and better-paid civil service. Nevertheless, a major challenge for the future remains the control o f the growth o f the wage bill, while supporting the expansion o f education and health services, which depends on the hiring o f large numbers o f teachers, doctors, and other qualified health personnel. A closer examination o f the wage bill i s provided in Section C. Interest Payments 2.7 Despite its relative decline in recent years, interest payments on the debt account for 11-17 percent o f total recurrent spending over the period 2001-06 (see Table 2.1). The share o f interest payments (on both domestic debt and foreign debt) in the total recurrent expenditure declined from about 16 percent in 2001 to 11 percent in 2006. This i s partly explained by the World Bank’s mobilization in 2003 o f bilateral donors (Belgium, France, Italy, Norway, and United Kingdom), in the context o f a Multi Donor Trust Fund (MDTF).23 The fact that Burundi was granted interim r e l i e f under the enhanced HIPC Initiative could be advanced as another explanation for the decline in interest payments. The implementation o f the Government’s strategy o f domestic debt arrears in 2006 explains the reduction in the domestic debt’s interest payments. Together with wages and salaries, interest payments made up about 53 percent o f total government recurrent expenditure, on average, over 2001-06. On the one hand, these “fixed and quasi-fixed expenditures” (about FBu 118.3 billion or US$115 million in 2006) imply that the 23 The MDTF, a grant which closed on December 3 1,2007, included two components: an in-servicetraining component and a debt component. Arrears clearance and debt servicing were supported by the latter component. The clearance o f debt service arrears to African Development Bank (AfDB), InternationalFund for Agricultural Development (IFAD), and European Investment Bank (EIB), and the payment o f debt service to the International Development Association (IDA), AfDB, IFAD, and EIB in a timely manner, was expected to require a contribution o f US$68 million from bilateral donors to meet these obligations. But until the closure o f the Trust Fund, total contributions amounted to US$20.9 million, o f which about US$14.7 million and US$5.8 million were used to service Burundi’s debt to IDA and AfDB, respectively. The average monthly debt service paid through the MDTF stood at roughly US$l 00,000 (compared to US$2 million prior to the HIPC Decision Point). In short, the main objective o f the MDTF was to support Burundi in meeting its obligations to its multilateral creditors, including AfDB, IFAD, EIB, and IDA, during the interim period, leading to access to the enhanced HIPC Decision Point, which Burundi finally reached in August 2005; in order to increase inflows o f donors assistance to the country. 26 Government has little control to shift expenditure from one category to another. On the other hand, capital expenditure i s mainly foreign financed (see below). Consequently, unless foreign assistance i s directed toward achieving i t s own public investment program (PIP), the Government has little room to maneuver in using the budget as a policy tool t o influence economic and social outcomes. Put differently, the current structure o f Burundi’s budget provides little fiscal space. The issue o f fiscal space i s discussed in more detail in the next chapter. However, Section D o f this chapter investigates the poverty and growth impact o f some o f the basic fiscal space issues (e.g., reallocation o f budget resources, allocation o f HIPC and MDRI debt relief). Other Goods and Services 2.8 The share o f expenditures on other goods and services (including critical inputs for the operational efficiency of key ministries) in total current expenditures declined from an average of 34.0 percent in 2001-04 to 30.7 percent in 2005-06 (see Table 2.2). This reflects the constraints faced by the Government originating from the relative weight o f the fixed or quasi- fixed expenditures discussed above. But most o f this decline i s explained by the reduction in military expenditures (from 22 percent o f recurrent expenditures in 2001 to 10 percent in 2006). This decline i s less pronounced (22 percent to 16 percent) when security (military and police) expenditures o n other goods and services are considered due to the creation o f the National Police in 2005. Despite these efforts by the Government to move away from a post-conflict context, and from military expenditures t o a development economy, the relative decline o f the share o f civilians’ other goods and services i s worrisome, given the importance o f this expenditure category for growth and development purposes (e.g., for maintenance o f infrastructure, education materials and health supplies). The relative peace stability in Burundi suggests that less priority be given to military spending and more to civilian expenditures, which are believed t o be growth enhancing and poverty reducing (e.g., other goods and services o n operations and maintenance, teaching materials, et^.).'^ Improving the allocative efficiency o f budgetary resources, for instance by reducing the size o f the wage bill or further reducing military expenditures on other goods and services, will help achieve such an objective. Transfers and Subsidies 2.9 Transfers and subsidies have been increasing since 2002, reflecting subsidy allocations to state-owned enterprises and increasing the welfare obligations o f the Government. Their share in total current expenditures increased from an average o f 13.6 percent in the period 2001-04 to 16.4 percent in 2005-06 (see Table 2.2). Due to the unavailability o f reliable data, we are unable to separate the analysis o f transfers o n the one hand and subsidies on the other hand to determine the relative weight o f each o f these expenditure sub-categories. However, w e can argue that the relative importance o f the share o f transfers and subsidies over the last seven years i s partly due to the recent creation o f many semi-autonomous public agencies (e.g., Sewice ChargB des Entreprises Publiques, SCEP; Inspection GBnBrale d ’Etat, IGE), and the return to normalcy o f the payment o f pensions after many years o f conflict. Finally, i t should 24 Capital expenditures, which are generally believed by economists to have a growth-enhancing impact, are related to recurrent expenditure through the other goods and services (operations and maintenance) spending category. 27 also be noted that in Burundi, wages and salaries o f staff in universities, including the University Hospital o f Kamenge (which trains nurses and other health personnel), are classified in the transfers and subsidies expenditure category. A natural recommendation i s therefore to improve the public expenditure classification by specifying the budget allocation to semi-autonomous institutions and others according to the economic classification o f public expenditure. Capital Expenditure 2.10 Capital expenditure has increased significantly, but remains dominated by foreign- financed expenditures (Table 2.1). Moreover, unlike countries such as Rwanda, Burundi has not benefited from a post-conflict momentum. From a very l o w level in 2001 (6 percent o f GDP), capital expenditure increased to 16 percent o f GDP in 2004 due to increased international assistance following the signing, at the end o f 2003, o f a global peace agreement which includes factions that were not part o f the 2000 Arusha peace agreement. On average, capital expenditure increased from 10.4 percent o f GDP in 2001-04 to 11.2 percent in 2005-06. Meanwhile, about 72 percent o f capital expenditure was financed by foreign assistance in 2005-06, compared to 67 percent in 2001-04, but this assistance remains very volatile. Despite the availability o f HIPC resources (which allocate about 73 percent to capital expenditure), the Government’s capital spending remains quite modest (at 3.1 percent o f GDP in 2005-06, down from only an average o f 3.4 percent in 2001-04). T h i s appears worrisome, since it reflects the impression that Burundi’s growth strategy i s greatly influenced by donors.25 But given i t s limited fiscal space, Burundi has to rely on foreign assistance to support i t s growth and development strategy, at least in the short run (see Chapter 3, Section D on Fiscal Space, for details). 2.1 1 The importance of investment for economic growth and poverty reduction, and the complementarity between public and private investment, means that the Government needs to make efforts to increase budgetary allocation toward domestically financed capital spending. This will only be possible through a more effective allocation o f public resources, away from non-productive expenditures such as those related to security, and more toward spending on other goods and services associated with capital outlays and social sectors. Section D, Impact o f Public Expenditure on Growth and Poverty, which provides, through a small macroeconomic model, a quantitative evaluation on growth and poverty o f scenarios for the allocation o f public expenditure, underscores the importance o f capital spending. Budget L a w o f 2007 by Economic Classification 2.12 The increasing trend of public expenditure has continued even in the 2007 Budget Law. Public spending i s expected to increase from 38.4 percent o f GDP in 2006 to 40.3 percent o f GDP in 2007. This increase o f 1.9 percentage points results from an increase in both recurrent and capital expenditure. One main component o f the increase i s the wage bill. Wages and salaries are expected to increase from FBu 93.9 billion (10.0 percent o f GDP) to FBu 113.5 billion (10.7 percent o f GDP) in 2007, and even further in 2008. Control o f the wage bill must become an ’’The increase o f domestically financed capital expenditure i s mainly explained by the impact o f HIPC resources, which account for more than 50 percent o f domestically financed capital expenditures. Moreover, for a number o f sectors, H I P C expenditures account for the bulk o f domestically financed capital expenditure. 28 important Government priority. In addition to wage restraint, this will depend o n measures aimed at strengthening management o f the wage bill and continued demobilization o f ex-combatants. Demobilization and security sector reforms will continue to reorient expenditure away f i o m security and toward social and poverty-reducing sectors. 2.13 Government expenditure in priority sectors i s discussed in the next sections. The analysis reveals that despite the post-conflict context, the Government’s allocation o f public resources has been broadly in line with PRSP priorities. However, the allocative efficiency will need to be improved in order to create more fiscal space to reorient resources toward growth-enhancing and poverty-reducing sectors. Impact o f Budget Execution on the Economic Structure o f Public Expenditures 2.14 Before 2006, rates of budget Table 2.3: Budget Execution Rates by Category of execution did not have a major Domestically Financed Public Expenditures, negative impact on the structure o f 2002-2006 - Ratio of Commitment to Initial Budget public expenditures. Table 2.3 presents Allocation (percentages) rates o f execution for each category o f 2002 2003 2004 2005 2006 domestically financed expenditures. The Wages and Salaries 98 97 101 104 119 Goods and Services 90 94 92 106 96 most striking changes between 2002 and Interest Payments 127 97 105 85 87 2005 are: (a) the decline in rates o f Transfers and Subsidies 98 99 102 101 128 execution for interest payments; and (b) Capital Expenditures - 96 - 94 - 102 - 103 - 52 a major increase in rates o f execution for Total Domestically- 100 96 100 100 100 wages and salaries (+6 percentage Financed Expenditures points), goods and services (+16 percentage points), transfers and Note: Excludes Special Funds, net lending, contingency line and foreign- subsidies (+3 percentage points) and financed expenditure Sources: MEFDC, SIGEFI. capital expenditures (+7 percentage points). In other words, budget execution procedures and practices protected not only wages and salaries, but also goods and services and capital expenditures. Several factors could explain the declining execution rate for interest payments; e.g., interest payments might have not been a priority for the transition Government, which was ending in 2005. Also, the lack o f capacity for debt management forecasts could lead t o calculation errors in the original budget figures o f interest payments. 2.15 The Government’s budget execution performance changed in the wrong direction between 2005 and 2006. Rates o f execution increased dramatically for wages and salaries (+15 percentage points), but declined sharply for goods and services (-10 percentage points) and catastrophically for capital expenditures (-5 1 percentage points). Across-the-board salary increases, and the recruitment o f new teachers in response to major improvements in primary school enrollments, explain high rates o f execution for wages and salaries. Complicated disbursement procedures for HIPC-financed expenditures are the main factors behind the decline in rates o f execution for capital expenditures and goods and services. Available data indicate that rates o f execution for HIPC-financed expenditures improved significantly in 2007. Again, controlling the growth o f the wage bill i s a major issue both at the budget planning stage and during budget execution. 29 Functional Classification 2.16 The Government’s public expenditure allocation to priority sectors has significantly improved over the past six years, but more needs to be done (Table 2.4). A look at the composition o f government expenditure by functions shows that non-priority economic and social sectors (including general services and security sector) averaged, in recent years (2005-06), 64 icture o f Domestically Financed Public ‘--wage over 2005-2006) percent for the priority sectors. In 200 1, prionty expenditures Other expenditures, General Services, accounted for 32 percent o f 2% exsludmg Seeunty, Pnonty Econonuc 26% total domestically financed and Socal Sectors expenditures. The evolving 36% structure o f government budgets and actual (domestically Secunty (Defense, financed) public expenditures lntenor and Public secunty), 36% reflect, in part, the PRSP priorities outlined in the Interim PRSP o f November 2003 and 2.17 The slow reduction in the share o f security expenditure in 2001-2005, and the availability o f HIPC funds in 2006, helped increase the share of economic and social priority sectors from 28.6 percent o f total public expenditures in 2001 to 34.1 percent in 2005 and 38.2 percent in 2006. From the beginning o f the 2000s, the Government accorded a very high priority t o the education sector, whose share o f total domestically financed expenditures increased from 23 percent in 2001 to 25.5 percent in 2004, 27.2 percent in 2005 and 29 percent in 2006. In contrast with the strong support given to education, the share o f health, agriculture, and infrastructure began to decline from 8.6 percent in 2001 to 6.9 percent in 2005, before increasing to 9.1 percent in 2006. The availability o f HIPC funds was a major factor in that increase. General Public Services 2.1 8 Expenditure on general public services (excluding security) increased from 4.6 percent in 2001-04 to 5.6 percent o f GDP in 2005-06, reflecting the need to establish national institutions following the 2005 elections. These services absorbed, o n average, 26 percent o f domestically financed public expenditure in 2005-06 (Chart 2.1). Despite the relatively high spending, the general view remains that the delivery o f services in Burundi i s poor, as standards within the public administration are not optimal and staff are often poorly trained and equipped for the functions they are required to perform. I t i s therefore essential to improve standards and accountability in the delivery o f public services. Chapter 5 discusses, among other issues, the incentives and professionalization o f the civil service that can help achieve that objective. 30 Table 2.4: Functional Distribution of Public Expenditures, 2001-2006 Percentage o f total Percentage of GDP expenditure Average Average Average Average 2001-04 2005-06 2001-04 2005-06 Domestically-Financed Expenditure General Public Services (excl. Security Sector and debt 26.9 26.0 4.6 5.6 servicing and amortization) Security Sector 38.7 35.7 6.6 7.6 Defense 36.2 24.0 6.2 5.1 Interior and Public Security 2.5 11.7 0.4 2.5 Priority sectors 31.9 36.1 5.5 7.8 Education 23.9 28.1 4.1 6.1 Health 3.8 4.0 0.7 0.9 Agriculture 1.9 1.4 0.3 0.3 Infrastructure26 2.3 2.7 0.4 0.6 Other expenditure 2.4 2.2 0.4 0.5 Domestically- and Donor-Financed Expenditure Priority Economic and Social Sectors 53.3 58.4 17.3 22.0 Exceptional expenditures 0.9 10.1 0.4 3.8 Elections 0.3 3.9 0.1 1.4 DDR 0.7 6.2 0.3 2.4 Other expenditure 39.6 38.2 12.3 14.4 Total donor-financedpublic expenditures (excluding 22.0 29.5 7.4 11.1 budget support) Total donor-financed public expenditures (including 38.6 49.8 12.8 18.7 budget support) Total domestically-financed public expenditures 17.2 21.5 Source: MEFDC. Defense and Public Security 2.19 Security spending (defense/military, interior, and public security including police) account for a large though slowly declining share of total public expenditures, a s expected in a country emerging from a long series of civil conflicts. Assuming that peace prevails, military spending should continue to come down: at about 5 percent o f GDP in 2005-06, it remains twice above the average (2.6 percent) for the East African Community (EAC). The signing o f the Arusha Peace Accord in 2000 and various agreements with rebel groups led to significant improvements in the security situation. Consequently, the share o f defense spending significantly declined from an average o f 36.2 percent o f total public expenditures in 2001-04 to 24.0 percent in 2005-06. However, as i s typical in a post-conflict setting, the budget o f the Ministry o f Interior and Public Security increased rapidly, and i t s share o f total public expenditures went up from 2.5 percent in 2001-04 to 11.7 percent in 2005-06, following: (a) the creation o f the National Police Force (NF’F) in 2005; (b) harmonization o f the pay scale o f police officers with that o f civil servants in 2006; and (c) the assignment to the NPF o f staff from the Ministry o f Defense. Together, the share o f security expenditures (defense and interior and public security) went down from 40.3 percent in 2001 to 32.9 percent in 2006. 26 Estimates o f public spending in infrastructure displayedin Table 3.14 differ from those in Tables 2.4-2.6 because the former includes an estimation o f off-budget spending (REGIDESO and FNR) while the latter i s simply based on budget laws. 31 2.20 A comparison of Chart 2.2: A Cross-Country Comparison of Military Burundi's military expenditure Expenditure (in percent o f GDP, 2006) with that for other EAC countries and Sierra Leone reinforces the argument that Burundi has to further reduce its defense 5.0 spending to increase its 4.0 expenditure toward priority 0 economic and social sectors if it 3.0 wants to compete with these 2.0 countries (Chart 2.2). Military 1 .o expenditures in Burundi accounted for 5.8 percent o f GDP in 2006,27 0.0 compared to only 1.3-2.8 percent I I Burundi Rwdnda Uganda Kenya SierraLeone Tanzania for i t s selected comparators. One Source: SIPRl Military Expenditures Database; IMF; Staff calculations. may argue that this i s due to the fact that the war in Burundi ended only recently. However, we investigated these expenditures for Rwanda just 6 years after the official end o f i t s war, and a similar conclusion, that Burundi's military expenditures need to continue to go down, was drawn even though these expenditures for Rwanda represented 3.3 percent o f GDP in 2001. Moreover, even a country such as Sierra Leone, which emerged from conflict not long ago, reported very l o w defense expenditure to GDP ratios compared to Burundi. 2.2 1 Finally, demobilization, although financed through donor resources, will play a major role in reducing military expenditure. The impact o f the demobilization, disarmament, and reintegration (DDR) program undertaken by the Government, with the help o f the donor community, should not be underestimated. The immediate impact o f the DDR on domestically financed expenditures i s minimal, as most o f the program i s financed by the donor community. However, the DDR program will reduce military and police staff and should facilitate the decline o f total security expenditures to more sustainable levels. P r i o r i t y Economic and Social Sectors 2.22 The slow reduction in the share of defense and public security expenditure in 2001- 2005, and the availability of HIPC funds in 2006, helped increase the share of economic and social priority sectors (education, health, agriculture, and infrastructure) from an average of 31.9 percent of total domestically financed public expenditures in 2001-04 to 36.1 percent in 2005-06. With respect to GDP, the share o f priority expenditure also significantly increased, from 5.5 percent in 2001-04 to 7.8 percent (about an increase o f 2 percentage points) in more recent years (Table 2.4). T h i s increase i s mainly explained by the priority recently given by the Government to the education sector, whose spending share with respect to GDP increased by two percentage points over the two periods. In contrast with the strong support given to education, the 27 That is, six years afier'the signing o f the Arusha agreement, that was considered to be the turning point o f the political situation in Burundi, after the longest o f its civil conflicts (1993-2003) since its independence in 1962. 32 share o f health, agriculture, and infrastructure taken together began to decline from 8.6 percent in 2001 to 6.9 percent in 2005, before increasing to 9.1 percent in 2006. The availability o f HIPC funds was a major factor in that increase. 2.23 The Burundi distribution of public expenditures among priority economic and social sectors subcategories i s greatly biased toward education, a s 78 percent of total priority sectors expenditures go to education only. In terms o f resources allocated to the education, health, agriculture, Table 2.5: Comparison of Spending on Priority Sectors and infrastructure sectors (Percent of GDP) relative to GDP, with the Burundi Kenya Rwanda Tanzania Uganda exception o f Rwanda, Burundi Average Average Average Average Average 2005-06 2004-05 2005-06 2003-04 2003-04 fares below other E A C Priority 7.8 9.3 7.7 12.0 8.8 countries (see Table 2.5), Sectors which reinforces the Education 6.1 6.6 4.2 5.1 4.5 importance o f creating more Health 0.9 2.0 1.4 2.7 2.2 fiscal space to find priority Agriculture 0.3 0.7 0.6 3.2 0.4 economic and social sector Infrastructure2’ 0.6 ... 1.6 1.o 1.7 (e*g*,through reducing security Source: MEFDC; IMF; Bank o f Tanzania. expenditures). Allocations for education have been relatively high over the past 2-3 years. However, given that expenditure categories might be defined differently in the countries selected, these comparisons should be treated with caution. Education 2.24 Burundi’s spending on education, at about 6.1 percent of GDP in 2005-06, i s significantly above that for other EAC countries except Kenya (e.g., Rwanda, 4.2 percent) (Table 2.5). However, education outcomes are yet to match those for post-conflict counties such as Sierra Leone and Rwanda (a detailed discussion i s provided in Chapter 3, Fiscal Space Analysis and Economic and Social Sector Expenditure Review). From the beginning o f the 2000s, the Government accorded a very high priority to the education sector, whose share o f total domestically financed expenditures increased from an average o f 23.9 percent in 2001-04 to 28.1 percent in 2005-06. Driven by high wages, it i s now the second largest category o f spending, behind security expenditures. Health Care 2.25 Although health outcomes are higher than those of some post-conflict countries, they remain below the sub-Saharan average. This i s partly explained by the lower allocation of public expenditure for health. With a share o f Government’s spending on health averaging 4 percent (Le,, 0.9 percent o f GDP) in 2005-06, Burundi fares far below all i t s EAC comparators in the allocation o f budgetary resources to health care. In contrast to education public spending, health expenditure increased only marginally (0.7 percent to 0.9 percent o f GDP, 28 Estimates o f public spending in infrastructuredisplayed in Table 3.14 differ from those in Tables 2.4-2.6 because the former includes an estimation o f off-budget spending (REGIDESO and FNR) while the latter i s simply based on budget laws. 33 on average, from 2001-04 to 2005-06). Its allocation share in domestically financed expenditures also increased, o n average, by only 0.1 percentage point. As will be discussed in the health sector section o f Chapter 3, the lack o f qualified personnel and the concentration o f available personnel in Bujumbura are cited as two o f main factors behind the poor performance o f the health sector. This report recommends increasing budgetary allocation to health to address the most urgent priorities o f the sector, including expanding and improving training facilities and providing incentives for health personnel working outside Bujumbura. Note that, owing to HIPC resources, nominal health allocations for 2007 are 62.7 percent higher than the executed health budget in 2006, although 23 percent lower than the original budget allocation in 2006. Economic Sectors 2.26 Burundi’s spending on economic sectors (including agriculture and infrastructure) remains extremely low. At less than one percent of GDP (4.1 percent of total expenditure), public spending in economic sector ranks below EAC comparators whose share of public expenditure in agriculture and infrastructure represents between 2.1 and 4.2 percent (Table 2.5). Despite the availability o f HIPC resources in the last two years, the allocation o f public spending to economic sectors remains weak: the GDP share o f public expenditure on infrastructure only increased by 0.2 percentage points from 2001-04 to 2005-06. Similarly, the share o f these expenditures in the total domestically financed expenditures also increased only marginally, from 2.3 percent to 2.7 percent. 2.27 Moreover, the lack of priority given to agriculture, a s evidenced by the low allocation of public resources to the rural sector, i s a cause for concern in an agrarian economy. Since agriculture i s the predominant sector o f Burundi’s economy, with more than 800,000 households living on coffee production proceeds, the authorities should allocate more resources to the development o f the rural sector. A recent World Bank study on the sources o f rural growth (World Bank, 2007) calls for enhancing productivity in agriculture. The study emphasizes the need to improve public expenditure management systems, especially in the coffee sector, which i s plagued by poor governance and corruption, due largely to years o f c i v i l war. This PEMFAR argues that actions are needed to both: (a) increase allocation o f public resources to the sector, to finance specific activities pertaining to improving productivity; and (b) strengthen the capacity to design and control public expenditures. The national budgetary resources allocated to the agricultural sector should be increased progressively in conformity with the declaration made at the N E P A D summit in Maputo, which set targets o f 10 percent o f the total budget and an improvement in budget execution rates in the agricultural sector. The development o f rural infrastructure i s also crucial. A more detailed analysis o f the structure and trends in infrastructure (but not in agriculture) i s provided in Chapter 3, the sector expenditure review. Budget Law o f 2007 by Functional Classification 2.28 As with the 2005 and 2006 budgets, the 2007 Budget L a w allocates an increasing share o f resources to health, education, and infrastructure (Table 2.6). In certain key sectors, large rises in projected expenditures are being driven by HIPC resources. For a number o f sectors, HIPC expenditures account for the bulk o f domestically financed capital expenditure, and thus explain the projected large increases in overall capital expenditure. I f actual budget execution i s in line with the original Budget Law, then i t can be argued that Burundi i s increasingly allocating more resources to key priority sectors. The following section discusses the factors determining the structure o f public spending. 34 Table 2.6: Projected Domestic Expenditure by Functional Classification (in current FBu billions, unless otherwise indicated) 2006 2007 YOChange Actuals Budget Law General public services (excl defense) 70.0 87.5 25.0 Interior and Public Security 27.5 34.9 27.1 Defense 47.9 53.6 11.8 Education 66.4 87.5 31.8 Fight against AIDS 0.8 2.5 225.1 Health 9.9 16.0 62.2 Agriculture 3.2 6.4 98.7 ~nfrastructure~~ 7.2 14.7 104.7 Other expenditure 88.3 79.9 -9.5 Total 293.6 348.0 18.5 Note: The data for 2006 i s actual expenditure on a commitment basis as per SIGEFI, while 2007 figures are based on the 2007 budget law. HIPC resources are included for both 2006 and 2007. Other expenditure includes other ministries, as well as debt service and the contingency line. Source: MEFDC, SIGEFI and 2007 Budget Law. Box 2.1: Use of HIPC Resources In August 2005, Burundi reached i t s decision point under the enhanced HIPC Initiative and qualified for interim debt relief. According to the Budget Law passed by Parliament, HIPC relief appropriations totaled FBu 8 billion in 2005, rising to FBu 39.5 billion in 2006. Although the Government projected that expenditure from HIPC funds would reach FBu 39.5 billion in 2006, delays caused by procurement and capacity issues meant that expenditures were only incurred from July. As a result, the volume o f actual HIPC expenditures in 2006 was only FBu 15 billion, representing an execution rate o f 38 percent. Nearly half o f HIPC funds were allocated to the education sector, with health receiving a quarter o f the resources, and infrastructure 11 percent o f funds. HIPC funds were used predominantly to finance capital expenditures (73 percent), although for some sectors, such as health, 24 percent o f total HIPC funds were used to finance recurrent expenditure-such as programs that promote the health o f mothers and infants, subsidies for caesarians, and community-based family planning activities-while only 0.2 percent financed health infrastructure. Education received nearly half o f total HIPC funds, and these were used exclusively for infrastructure investments. Over 98 percent o f these funds financed the construction o f primary schools, in line with the large expansion o f primary education, following the Government's decision in September 2005 to make primary school education free. HIPC resources help explain in part the rising share o f total expenditure directed at health (including HIV/AIDS), education, and infrastructure. In 2006, actual HIPC expenditures were equivalent to 4.3 percent o f total expenditure, including donor funded spending. However, these resources w i l l rise considerably when Burundi attains the HIPC Completion Point, which i s expected to occur in the first half o f 2008. I t i s desirable to continue using HIPC funds for both capital and recurrent expenditures, in order to ensure adequate funding o f priority recurrent expenditures. With considerable delay, in June 2007 a commission was set up in the Ministry o f Economy, Finance, and Development Cooperation (Commission de suivi de depenses PPTE) with the aim o f monitoring HIPC expenditures. This w i l l make i t easier to monitor HIPC expenditures and to analyze their actual impact o n key indicators. The PPTE will be responsible for monitoring HIPC-financed activities and evaluating their impact o n key economic and social indicators. 29 Estimates o f public spending in infrastructuredisplayed in Table 3.14 differ from those in Tables 2.4-2.6 because the former includes an estimation o f off-budget spending (REGIDESO and FNR) while the latter i s simply based on budget laws. 35 B. Factors Determining the Structure o f Public Expenditure 2.29 T w o factors stand out as major determinants o f the structure o f public expenditure in Burundi: (a) the level o f domestic resources; and (b) the influence o f external financing. Level o f Domestic Resources 2.30 Because of Burundi’s current low domestic resources, the structure of public expenditure presents a double characteristic. The limited resources o f the Government are primarily allocated to recurrent expenditures, while capital expenditures are mainly financed by external sources. Fixed and quasi-fixed expenditures (wages and salaries, interest payments, and transfers and subsidies) account for a large part o f government resources, leaving little room for other goods and services - the other category o f recurrent expenditures. Expenditures on other goods and services appear as an item that can be easily compressed in the recurrent budget. In other words, it i s a “residual” item operating like an adjustment factor. 2.3 1 National resources to finance the investment budget are limited and their allocation i s preset. Despite recent efforts by the authorities to develop a framework for the size and the structure o f the ongoing and future public investment program (PIP), investment expenditures are s t i l l not identified and planned in a coherent framework. The Government began to prepare PIPs in 1988 to improve investment planning and mobilize donor financing. A few PIPs were also prepared during the crisis o f the 1990s, but proved to be virtually useless. Improved relations with the international community in the 2000s induced the Government to resume preparation o f comprehensive three-year PIPSas a donor coordination and resource mobilization mechanism, the quality o f which needs to be improved. This situation has led to a separation between public sector capital spending on the one hand, and socioeconomic development needs on the other hand. Consequently, public sector investment has been driven by donor-funded and NGO projects, further exacerbating the separation between current expenditure and investment budgets. The influence o f the Government o n the investment budget i s limited. Impact of Donor Financing 2.32 External financing has an increasingly strong influence on the overall pattern o f public expenditure, as external project finance funded more than 70 percent of capital expenditure in 2005-06 and program and project donor assistance financed about 50 percent of Burundi’s total public expenditure (Table 2.7);’ Program loans and grants, equal to 18 percent o f total expenditure (including debt service) in 2006, i s another mechanism by which international assistance has shaped public expenditure decisions. As a share o f GDP, budget support (program grants and loans) increased Erom 2 percent in 2001 to a peak o f 8 percent in 2004 and 2005, but slightly declined to 7 percent in 2006. As a whole, externally financed 30 Donor contributions increased from 18 percent of total public expenditures in 2001 to 53 percent in 2005, and then declined to 50 percent in 2006. On average, about 41 percent of these contributions were in the form of budget support, which helped fund domestically financed expenditures. Externally financed project spending, which accounted for 1 1 percent o f total public expenditures in 2001, reached a peak of 29 percent in 2004 but declined to 21-23 percent in 2005-06. External funding i s even higher once exceptional financing for elections and demobilization i s added to program and project loans and grants (Annex 3). 36 expenditure, which includes budget support, exceptional expenditure, and externally financed project expenditure, accounted for 5 percent o f GDP in 2001 and 20 percent in 2006. The strong preference o f donors for financing projects and programs in priority sectors has led to an increased allocation o f funds for education, health care, and infrastructure in recent years. In addition to influencing how public funds are allocated, external financing for fiscal reforms has led to improvements in public finance management, which have helped to improve budget processes. 2‘33 Burundi’s On Table 2.7: Donor-Financingof Total Public Expenditure external budget support (including debt service) reinforces the influence o f Averaee Averaee donor contributions on the 2001-04 200506 External project finance (% capital expenditure) 66.7 74.5 structure Of public spending. External project finance (%total expenditure) 20.9 21.6 Although government data do not External project finance (% GDP) 7.0 8.3 the distribution by sector - Budget support loans and grants (%total expenditure) 16.4 19.9 o f donor funds, it can be safely Budget support - loans and grants (% GDP) 5.4 7.6 assumed that at least 95 percentOf DDR and elections financing (% total expenditure) 0.9 9.9 donor contributions have been DDR and elections financing (% GDP) 0.4 3.8 to typesOf projects: Externally financed expenditure (% total expenditure) 38.3 51.3 (a) priority economic and social Externally financed expenditure (% GDP) 12.8 19.8 Source: MEFDC. sectors; and (b) exceptional expenditures (i.e,, the electoral process and DDR activities). Based on this assumption, the bottom panel o f Table 2.4 presents the estimated share o f priority sectors (excluding exceptional expenditures) in total domestically and donor-financed public expenditures. I t shows that the share o f priority economic and social sectors in total public expenditures increased, on average, from 53.3 percent in 2001-04 to 58.4 percent in 2005-06.31 Thus, the share o f priority expenditures in total i s significantly higher in this case than when only domestically financed expenditures (excluding debt service) were considered (see top panel o f Table 2.4). 2.34 The sluggish increase in the share of economic and social sectors in donor funds and in total public spending i s due to the increase o f exceptional expenditures (DDR and electoral process), which accounted for about one fourth o f donor-financed project funds in 2005 and 2006. Since these exceptional expenditures also are o f high priority to achieve the objectives o f the PRSP, i t can be argued that the share o f priority sectors in total public expenditures rose from an average o f 54.3 percent in 2001-04 to 68.5 percent in 2005-06. Clearly, the success o f the DDR program i s essential to improve the overall structure o f public spending, as i t will help reduce the share o f defense in domestically financed expenditures and will enable donors to allocate more funds to economic and social sectors. Section C below discusses in detail wage bill as a key source o f fiscal pressures. 3’ The share o f priority economic and social sectors increased from 41 percent in 2001 to about 57 percent in 2003 and 2004, but declined to 50-5 1 percent in 2005 and 2006. 37 c. WAGE BILL - KEY SOURCE OF FISCAL PRESSURE 2.35 This section will examine the level o f staffing costs, usually referred to as the wage bill. I t i s important to analyze issues related to the level o f staffing costs as part o f a public expenditure management study because salaries, staff allowances, and pension payments are always the largest element o f recurrent public expenditure, typically ranging f i o m 35 to 60 percent o f total recurrent spending. Further, they are a component o f public expenditure under the control o f Government, in contrast to capital expenditures, which are largely financed by donors in low-income countries, especially post-conflict ones. In addition, personnel expenses are usually the first call o n any cash available to the Government. All o f these factors are true in the case o f Burundi as was discussed earlier in this chapter. The issues discussed in this section can be summarized as follows: 0 The wage bill i s very high and i s squeezing out other expenses needed to deliver basic services; 0 There i s some scope to reduce staffing and the wage bill by eliminating ghost workers; 0 Substantial restructuring and reengineering will be required if the wage bill i s to be reduced and at the same time the delivery o f services improved; and Improvements will also have to be made to the institutional arrangements for managing salaries and staffing in order to bring down the wage bill to a sustainable level and at the same time improve incentives. 2.36 The wage bill i s Table 2.8: Comparative Data on Wage Bill - high relative to GDP when compared to other Country fin units indicated) Real Wage Bill Wage Bill Wage Bill GDP (%GDP) (%Domestic (%Recurrent Sub-Saharan African Der Revenue) exoenditure) countries. The high civil capita (2000 service wage bill (masse US$, salariale) has become a 2006) Ghana 316 5.3 27 30 serious issue over the past Kenya 441 9 42 38 few years because it i s Malawi 157 6.3 35 34 squeezing out other Mozambique 319 6.7 53 44 Rwanda 268 5.2 42 33 recurrent expenses needed 503 5.4 29 44 to make civil Servants Tanzania 350 4.3 35 31 55 The Gambia FY O6'O7 343 4.5 29 27 policy and service delivery Burundi (2007 budget) 110 11 58 45 The wage Sources: World Bank files (Data for 2003 except where stated); IMF World Economic (Salaries plus allOWanCeS) Outlook October 2007; UN Statistics Division. was estimated by the IMF *Datafor2005. to be 8 percent o f GDP and 20 percent o f total expenditure in 2004 and i s rising: the wage bill was 7 percent o f GDP in 2002 and i s estimated to have risen to as much as 10.6 percent in 2007, making it one o f the highest in Sub-Saharan Africa. Data from some other l o w and middle- income countries are shown in Table 2.8; only Zanzibar's wage bill as a percentage o f GDP i s close t o Burundi's. However, in a number o f countries as well as Burundi, the wage bill as a proportion o f recurrent expenditure i s around 45 percent, suggesting that Burundi i s less o f an 38 outlier than if one restricted one’s attention to the wage bill as a fraction o f GDP. The high wage bill relative to GDP reflects in part the fact that Burundi i s emerging from a protracted period o f l o w economic growth. 2.37 The high wage bill situation i s not expected to improve significantly in the medium term, as shown in Table 2.9, which i s based on World Bank projections. As these projections are based on a high economic growth forecast for 2008-2010, the wage bill expressed as a percentage o f GDP could climb rather than fall if growth were to be below the forecast average o f 5.9 per annum for 2008-2010. Therefore, a major challenge for the future will be to control the growth o f the wage bill, while supporting the expansion o f education and health services, which depends on the hiring o f large numbers o f teachers, doctors, and other qualified education and health personnel. 2.38 Most of the growth Table 2.9: Trends in the Wage Bill in the wage bill will have (gross salaries, including allowances) taken place between 2005 (FBu billions, unless otherwise indicated) and 2007. The wage bill for 2005 2006 2007 2008 2009 2010 the civil service will have Total 72.6 93.9 113.1 131 145.5 157.4 risen considerably over that 85.9 94.5 Civil 41.9 55.3 66.4 78.2 period, while the wage bill Military 24 22.9 28.1 28.4 29.8 31.3 for the police increased Police 6.7 14.7 18.7 19.7 20.1 20.5 threefold (Table 2.9). ‘ 7 0 GDP 8.4 10 10.6 10.6 10.1 9.7 Increases in the numbers o f Note: IMF data for 2005 and 2006, and World Bank projectionsfor 2007-2010; teachers accounted for most Sources: I M F and World Bank. o f the r i s e in the civil service wage bill, as shown in Table Table 2.10: Comparing Staffing and Wage Bill in the Civil 2.10. The teachers’ wage Service, 2004 and 2007 (in units indicated) bill almost doubled, while 2004 2007 the number o f non-teachers Staff (excluding military and Teachers 26,579 36,266 Non-teachers 9,799 5,276 uolice) actuallv f e l l over the Contract 7,312 7765 same period. Total 43,690 49,307 2.39 Military wages are WageBill a major but falling Teachers 18,159 31,884 component of the wage bill. Non-teachers 5,430 2,028 The- high level of Contract 2,494 2,940 expenditure on military Total 26,083 36,852 wages, and defense generally, has been a Note: Salaries include allowances and are in FBu millions. Source: MEFDC. particularly severe problem, although the situation i s improving. Military expenditures (salaries, goods and services, and total expenditures) decreased sharply after the peace agreement signed in 2000, a trend that accelerated after 2004. Military salaries, which were 46.1 percent o f total salaries in 2001, fell to 24.4 percent in 2006, and expenditures on good and services also decreased, from 58.6 percent to 32.8 percent during the same period. Military salaries as a share o f total spending decreased significantly, from 12.3 percent in 2001 to 8.1 in 2006. 39 2.40 The wage bill of the police rose by nearly 200 percent in 2005-2007. This reflects the harmonization o f the salary structure o f the police with that o f the military, after the creation o f the Police National du Burundi in 2005. In addition, there was a large increase in the number o f overall staff, due to the transfer o f gendarmes from the Ministry o f Defense to the Ministry o f the Interior, the transfer o f certain categories o f the military and o f certain police from the Ministry o f Justice (Oficiers de la Police Judiciaire) to the Ministry o f the Interior. Prior to the creation o f the new national police force in 2005, the police were paid by the civil service. 2.41 Teachers, Table 2.11: Public Service Staffing and Average Monthly Base soldiers, and police Salaries, M a y 2007 (in FBu, unless otherwise indicated) dominate the numbers in Numbers Average Monthly base monthly salary the public service. The base salary expenditure composition o f the public (FBu) (FBu millions) service reflects i t s post- Police* 18,626 40,684 758 Militaire 28,108 36,023 1,013 conflict status. The Non-teachers 2 1,079 41,758 553 distribution o f staffing by Teachers 37,750 52,656 1,838 main category i s shown in TOTAL STAFF 105,563 39,427 4162 Table 2.1 1; more than a Note: The data for the police are for September 2007. Military and police staff numbers exclude civilians. Average salaries exclude staff that are 'sous-contrat' (contract workers). third are t ~ ~ ~ e and r s Sources: NP, MINDEF, MFP, CNI. almost a third are soldiers. Teacher salaries represent about 41 percent o f the total wage bill. Since health workers are included in the first figure (there are about 2,800 members o f the statutory staff and about 2,100 sous contrat), fewer than five percent are in what i s often referred to as the core civil service, the non-frontline service providers. Staff members under contract are almost all in very low-paid jobs such as cleaners, messengers, and custodians, although this category also includes a small number o f highly qualified and high-salary staff. 2.42 Official figures underestimate the true size of the public service. Some staff are not accounted for in Table 2.1 1. For example, schools can hire temporary teachers (vacataires) who do not appear on any official l i s t o f personnel or payroll. The review o f public expenditure in education o f this PEMFAR reports that there were about 49,000 teachers in 2005, well above the figure in Table 2.1 1. Some staff members do not appear on the payroll managed by the firm C N I for the MEFDC and the Ministry o f Public Service (MFP). These include the staff members o f the Presidency, the Vice Presidency and the Secretary General, who are not under the MFP. However, the numbers o f staff not specifically accounted for in Table 2.1 1 are thought not to be high. According to the 2007 budget, the cost o f these additional staff i s about FBulOl million a month, or FBu 1 billion a year. 2.43 There are inconsistencies in the staffing numbers. The numbers collected from the National Police, Ministry o f Defense, and CNI, which prepares the Fonction Publique payroll, vary somewhat from the figures supplied by the Ministry o f Economy, Finance, and Development Cooperation (MEFDC). MEFDC reports about 1,000 more teachers, slightly fewer contract workers and police and slightly more military. I n the case o f the Ministry o f Education, these differences are likely to be due to the vacataires, whose pay i s not managed by MFP, but between the Ministry o f Education and MEFDC. The military and policy numbers were collected directly from those two agencies. At the very least, this indicates a need to improve record-keeping. Ideally, working with the ministries, MFP should have up-to-date numbers for those officially 40 employed in the public service, as well as their personnel files (except for military and police). That information should then be supplied to the MEFDC, which would prepare the payroll (or contract with an entity such as CNI to do so), and payments would then be made to staff by MEFDC (all staff have to have bank accounts). With the development o f a payroll module in the Integrated Financial Management Information System, SIGEFI, scheduled for 2008, the management o f the wage bill will revert to the MEFDC, and payroll expenditure calculated by MFP should be incorporated into the expenditure chain, as i s already the case for the police and military payroll. 2.44 The civil service wage Chart 2.3: Service Staff by Category, M a y 2007 bill bulges in the middle of the I grading structure. The civil 80 service i s not shaped like a 70 pyramid. Most staff are in the middle “collaboration” 60 Category (i.e., the category below management), as shown in Chart 2.3, which shows the -5 50 40 a, breakdown o f staff in the civil n 30 service by “category” (category i s the first level in the grading 20 structure; each category i s 10 divided into four grades). This reflects the educational 0 Direction Collaboration Ewecution Sous contrat attainment o f civil service staff, as access to different categories Source: MEFDC. i s determined by educational qualifications. There i s also a high number o f staff sous contrat, equivalent to 20 percent o f total numbers. Opportunities for Cost Savings in Public Employment 2.45 The wage bill could be reduced if ghost workers were eliminated. There i s anecdotal evidence that there are some ghost workers in the c i v i l service. Administrative systems are so weak that some staff are not taken o f f the payroll when they retire, die, or leave for another job. M a n y countries have successfully eliminated ghosts by carrying out a civil service census. However, the key to eliminating ghosts i s to have in place a good human resources management information system (HRMIS). The best o f these systems include biometric identification o f staff. A n e w biometric card (carte securitaire) i s scheduled to be unrolled in 2008, following the completion o f the public servants census. Most public administration reform programs include designing and implementing HRMISs. Some have favored including a human resources module in their Integrated Financial Management Information System (IFMIS), as i s planned in Burundi. This integrated human resources system, scheduled for implementation in 2008, will be connected to the MEFDC’s SIGEFI and i s therefore expected to greatly improve the management o f payroll. 41 2.46 I n the long term, the number of staff and the wage bill have been, and may continue to be, on a rising trend. Civil service employees have increased over time, even during the post- conflict period: there were about 23,000 staff in the civil service in 1983, compared with more than 50,000 in 2007. The current policy i s to only replace staff who leave, except for new recruitments in education, health, and the judiciary. The number o f soldiers i s being reduced: the target for the end o f 2007 was 27,000 demobilized, although this depended upon the remaining rebels being demobilized. By September 2006, about 21,000 soldiers had been demobilized. The medium-term target i s to reduce, in 2008, the military from about 27,500 soldiers today to 25,000 and the National Police Force from about 18,000 today to 15,000. Despite pressure from the donor community to implement demobilization rapidly, ultimately its speed will depend on the security situation and on the conclusion o f successful negotiations with the one rebel movement which has yet s i g n a peace accord. So, although the numbers are expected to decline for the military and police3* and hiring will be restrained in the core civil service, hiring more teachers and health workers to meet high-priority poverty reduction needs and more magistrates to meet the needs o f an improved judiciary i s likely to put upward pressure on the size o f the public service in the future. In addition, the policy o f reintegrating all exiled public servants means that in the future, these could be a further source for an overall rise in staff numbers. 2.47 The current Table 2.12: Wage Bill and the Budget proposal to increase (in percent, unless otherwise indicated) allowances will raise the 2005 2006 2007 2008 2009 2010 wage bill substantially. Wage bill/GDP 8.4 10 11 11.4 11.5 11.3 Wage bill/current Proposals have been expenditure 36.2 42.4 45.1 48.3 46 45.3 prepared to increase Goods and 7.6 6.8 6.9 6.7 7.3 7.5 allowances by 34 percent. services/GDP Goods and Although this was services/current 32.8 28.8 26.7 21.8 29.3 30.1 initially planned for the expenditure revised 2007 budget, the Memorandum items (in FBu billions) Current expenditure 201 22 1 261 279 316 341 Government decided to Wages and 13 94 114 135 145 157 postpone the increase, Salaries and it has been included Civilians 42 56 13 89 98 109 Military 24 23 29 28 29 30 in the 2008 budget. Police 7 15 19 18 18 19 Salaries have been raised Goods and 66 64 71 78 93 105 across the board twice in services Nominal GDP 861 938 1029 1150 1264 1395 this decade: by 10 percent Source: Burundi authorities; and IMF staff estimates and projections. in 200 1 and by 15 percent in 2006, although every year there are performance-based salary increases, which can be a maximum o f 7 percent. The Government decided to raise allowances in part because there had been basic salaries in recent years, while allowances had been raised i n f k e q ~ e n t l y The . ~ ~ World Bank estimates that because o f these rises, the wage bill will increase to 11-11.4 percent o f GDP 32Once the rebels sign the peace accord they will be integratedinto the army and police, which will lead to a rise in overall numbers, although demobilization in the medium term would cause overall staffing levels to fall. 33 The family allowance i s the same as it was during colonial times. 42 in 2007-2008, and then fall marginally t o 11.3 percent in 2010, still a high level. The high wage bill keeps expenditure on goods and services at less than percent o f total current expenditure. I t i s projected that the wage bill will increase to 48.3 percent o f current expenditures in 2008 (Table 2.12). 2.48 I t may be possible to reduce the wage bill by restructuring the public administration. Almost every public administration reform program in Africa includes rightsizing and restructuring, the objectives being to reduce the wage bill (or at least prevent it rising), create fiscal space to increase salaries, and make the administration more effective. Kenya and Zambia, for example, have reduced the sizes o f their c i v i l services quite substantially. Although the number o f staff in the public service in Burundi may not be particularly large by low-income country standards, there i s much anecdotal evidence that some staff, particularly in non-frontline service provision jobs, are surplus to requirements. In some cases, this i s because they are not equipped to do their jobs, but in others cases it i s clear that they either do not have a real j o b to do or are not motivated to do their job. The scope for reducing staffing i s probably limited in Burundi, since all but a few thousand are service providers. Planned reductions in the number o f police and military have already been mentioned. W e will return to the subject o f restructuring in Chapter 5; in this case the objective may be to increase efficiency rather than to reduce the level o f the wage bill, although the Government should ensure that the wage bill as a proportion o f GDP falls, or at least does not increase, by 2010. A study to investigate pay reform options i s recommended in the medium term. 2.49 I t i s difficult to avoid the conclusion that several steps will have to be taken to reduce the wage bill over time. These include: 0 Conduct a census o f military, police, and civil service; 0 Freeze recruitment o f all staff, except for medical personnel, teachers, and magistrates; 0 Establish a coherent policy o n salary increases; 0 Revise the policy o f replacing all retiring stafc and 0 Enhance the powers o f the Recruitment Commission. 2.50 These measures, together with robust economic growth, should lead to a reduction in the wage bill-to-GDP ratio to less than 10 percent by 2010. The next section deals with the growth and poverty effects o f public expenditure reallocations, including the reduction o f wage bill based o n a small macroeconomic model. D. IMPACT OF PUBLIC EXPENDITURE ON GROWTH AND POVERTY - MODELINGRESULTS 2.51 To assess quantitatively the impact that changes in fiscal consolidation and the composition o f public expenditure would have on growth and poverty in Burundi over the medium term, a simple macroeconomic modelj4 was developed. The model simulates the impact o f external debt relief and changes to the level and composition o f public expenditure on 34A more detailed technical presentation o f the model, as well as a description o f the calibration procedure, i s available in J-P. Nganou and E. Pinto Moreira, “Public Expenditure, Growth and Poverty: Exploiting Cross-CountryRegressions for the Burundi PER Report.” Unpublished,World Bank (July 2007). 43 real per capita economic growth and poverty over 2008-2015. Due to the lack o f reliable national accounts data in Burundi,35 the model uses parameter estimates derived from three recent cross- country studies - a growth equation, a public investment equation, a private investment equation for low-income countries, together with some basic budget id en ti tie^.^^ T o link the model with poverty, two partial elasticity formulae are used: (a) a formula estimated by Ravallion (2004) which relates the rate o f change in poverty headcount index to the rate o f growth o f real income per capita, adjusted for a measure o f inequality; and (b) alternative values (-0.5, -1, and -1.5) o f growth-poverty elasticity. A key feature o f the model i s that public investment and private investment are treated endogenously. T h i s captures the complementarity effects between the two types o f investment. The model i s summarized in a schematic diagram in Annex 7. Box 2.2: Impact of Public Expenditures on Poverty The analysis o f public expenditure allocation and budget execution provides an important indication o f how the Government i s increasing resources to priority sectors, which should translate into improved social indicators and lower poverty rates. Ultimately, however, what i s also required i s to assess the extent to which budget funds are being targeted towards provinces with high levels o f poverty. Currently, no systematic assessment o f the link between public spending and poverty i s carried out in Burundi. Regional poverty rates are estimated to range from a l o w o f 28.7 percent in Bujumbura to 82 percent in the province o f Kirundo in 2006, but there are some issues regarding the accuracy o f such poverty numbers. The main limitation to conducting such an assessment i s the lack o f data o n the geographical destination o f budgetary funds. Burundi’s highly centralized budget management system and lack o f geographical budget data make it virtually impossible to identify the regional distribution o f public expenditures. The knowledge gap on the poverty impact o f public spending can be partly addressed through analytical work aimed at informing policymakers about the impact o f public expenditures in Burundi. More importantly, there i s an urgent need for l i n e ministries to systematically include geographical information for all items in their budget proposals and for a monitoring system to be put in place to track the amount o f public funds being directed to different provinces. Currently, monitoring o f HIPC fund usage through SIGEFI i s o n a national program-by-program basis for both recurrent and capital expenditures. As a first step, it would be important to create a mechanism for tracking the geographical distribution o f HIPC resources, which are meant to be pro-poor budget outlays. 35 Several years o f civil conflict affected Burundi’s statistical infrastructure. 36 T h e growth equation was derived by Gupta, Clements, Baldacci and Mulas-Granados (2005) for a panel of 39 low- income countries over 1990-2000. Results from the baseline regressions find evidence that fiscal consolidation achieved through curtailing current expenditures, are, in general, more conducive to growth. More specifically, a percentage point of GDP increase in spending on wages and salaries reduces real per capital growth, while a one percentage point increase in the ratio of capital outlays to GDP increases growth by more than half a percentage point. An implication o f these results i s that protecting capital expenditures during periods o f fiscal adjustment leads to higher growth, while reducing the wage bill does not have an adverse impact on growth. For the public investment-to-GDP ratio equation, the model uses the regression results estimated by Clemens, Bhattacharya, and Nguyen (2003) for a group of 55 low-income countries for the period 1970-1999. This regression makes public investment a s a percentage of GDP a finction of foreign aid in percent of gross national income; the urbanization ratio; total debt service as a percent of GDP; real per capita income; an openness indicator and on the external debt stock. The private investment- to-GDP ratio parameters have been taken from Mlambo and Oshikoya (2001), who estimated investment equations for a group o f African countries for the period 1970-1976. Their results show that an increase in a country’s growth rate, real GDP per capita, or public capital expenditure stimulate private investment, whereas an increase in foreign debt or a deterioration in the fiscal balance tends to reduce this. 44 2.52 A baseline scenario was calculated in order to assess the impact of changes in public expenditure policy. I n this scenario, the budget deficit as a share o f GDP i s taken as fixed to make aid endogen~us.~’ Table 8.1 in Annex 8 presents the baseline scenario, which assumes real GDP per capita growth o f 1.3 percent in 2007, rising to an average 1.6 percent for 2008-2015. The projected growth rates lead to a significant drop in the poverty rate. Using Ravallion’s adjusted elasticity, poverty incidence declines from 66.9 percent in 2006 to 44.8 percent in 2015, declining by significantly less using alternative growth-poverty elasticities. These simulation results suggest that current rates o f external assistance would not be enough to allow Burundi to attain the Millennium Development Goal (MDG) for poverty o f reducing poverty by h a l f by 2015. 2.53 Using baseline fiscal and macroeconomic data for 2006, mainly from the International Monetary Fund (IMF) and the World Bank, the model takes the parameters derived from three regressions together with budget identities, to simulate what impact changes in the economic composition of public expenditure and reductions in external debt would have on economic growth and poverty (Annex 6). Five different policy alternatives are considered: (a) an increase in public investment financed by aid; (b) a budget-neutral reallocation o f spending f r o m other recurrent expenditures to capital expenditures; (c) a reduction o f debt; (d) total forgiveness o f foreign debt; and (e) a reduction in wages and salaries. I n line with the favorable international environment that Burundi currently faces, the simulations are conducted under the assumption that the overall budget deficit i s constant and aid i s the balancing item in the Government budget. 2.54 The main findings from the simulations are: 0 The higher capital expenditure scenario assumes that capital expenditure as a percentage o f GDP rises by 4 percentage points3* above the baseline scenario in 2008 and peaks at 4 percentage points in 2008-2012 before falling to 1 percentage point above the baseline scenario in 2015. This scenario assumes that other components o f public expenditure remain unchanged and that this surge in public investment i s entirely financed by foreign grants. In other words, there i s a fiscal stimulus because o f increased public expenditure and at the same time an increased share o f public funds going to public investment. Using the Ravallion-adjusted formula, the simulation results reveal that by 2015, the poverty rate would fall t o 18.2 percent in 2015, as real GDP per capita rises.39 A second scenario assumes that capital expenditure as a percentage o f GDP rises by 4 percentage points in 2008-2012 in relation to the baseline and slows thereafter, while other current expenditure rises to offset the fiscal stimulus. N o t surprisingly, the impact on poverty i s lower than in the previous scenario, with poverty rates falling by 15.8 37In other words, the baseline scenario assumes that for a given change in debt, any shortfall in resources i s filled by foreign aid. I t should be noted that this closure rule can be changed. The shares of all other components of spending remain constant at base period values. 38The budget of 2008 shows an increase o f capital expenditure to GDP by about 3.4 percentage points relative to the 2006 figure (Le. from 12.45 percent to 15.8 percent). ’’ Based on an assumed growth-povertyelasticity of -0.5, poverty will fall by 5.2 percent. 45 percentage points in relation to baseline rates, as real GDP per capita rises, but not by as much as before. A third scenario assumes a reallocation o f savings to public investment as a result o f partial debt relief (i.e,, foreign debt stock i s reduced by half). Capital expenditure rises by 0.6 to 0.8 percentage points above the baseline scenario, while interest payments decline by 0.6 percentage points throughout 2009-2015. As a result, the poverty rate by 2015 i s 5.4 percentage points below the baseline rate o f 44.8 percent. A fourth scenario assumes a reallocation o f savings to public investment as a result o f total debt relief. Capital expenditure rises by 1.2 to 1.5 percentage points above the baseline scenario, while interest payments decline by 1.1 percentage points throughout 2009-2015. As a result, the poverty rate by 2015 i s 10.2 percentage points below the baseline rate o f 44.8 percent. A fifth scenario assumes a reduction in the wage bill by 1.5 percentage points o f GDP in 2009 in relation to the baseline scenario. This has a powerful impact o n poverty rates, reducing poverty by 17.5 percentage points below the baseline rate, in line with the parameters o f the growth equation, which finds that real per capita income rises when the wage bill declines. 2.55 A policy implication of the simulation results i s that the composition of public expenditure has a large impact on economic growth and poverty. Increasing the share o f capital expenditure has a dynamic effect o n economic growth and has a large impact o n both growth and poverty rates. Likewise, reducing the wage bill has an even larger impact than an increase in capital expenditure. Burundi’s wage bill, expressed as a percentage o f GDP, has been rising f r o m 7.3 percent in 2001 to 10 percent in 2006, and the baseline model assumes that it will reach 11.8 percent in 2007. The simulations suggest that poverty could be reduced more quickly i f Burundi tried to reduce the size o f the wage bill instead o f keeping it constant over the same period. The model also clearly indicates that a continued rise in the wage bill, everything else being constant, will have a dampening effect on both growth and poverty. Alternatively, Burundi could step up i t s public investment program in order to increase capital investment as a share o f GDP, which i s estimated at 12 percent o f GDP in 2006. Conclusions and Recommendations 2.56 Since 2001, Burundi has made considerable efforts to increase the shares of health, education, and infrastructure spending in domestically financed public expenditure. The share o f domestically financed resources allocated to the ministries o f health and education increased from an average o f 27 percent in 2001-04 to 32.0 percent in 2005-06, while the share allocated to infrastructure rose from 2.3 percent to 2.7 percent. Increases in public expenditure for priority sectors accelerated in 2006, as a result o f HIPC resources. HIPC fund expenditures are expected to rise in the medium term and will help finance increased social spending and key infrastructure. Nevertheless, domestically financed health and infrastructure expenditures remain l o w and need to be raised to improve social indicators, reduce poverty, and increase growth. 2.57 Increased social spending has not reflected a significant fall in defense and public security spending. Excluding debt service and amortization, the share o f domestically-financed spending o n defense has fallen, o n average, from 36.2 percent in 2001-04 to 24.0 percent in 2005- 46 06. However, domestic resources going to the Ministry o f the Interior and Public Security have risen from 2.5 percent to 11.7 percent over the same period. Taken together, the share o f security spending in total domestically financed expenditure has not declined significantly, from 3 8.7 percent in 2001-04 to 35.7 percent in 2005-06, despite demobilization. As a share o f GDP, security spending increased on average from 6.6 percent in 2001-04 to 7.6 percent in 2005-06.40 The creation o f the National Police Force explains the large rise in public security spending over 2005-2006. Although the number o f staff in the police i s expected to continue to decline, there i s a need to ensure that the share o f resources allocated to the Ministry o f Interior and Public Security does not continue to rise, crowding out pro-growth and pro-poor expenditures. 2.58 The analysis o f public expenditure by economic category highlights the rapid increase in the wage bill in recent years, which i s expected to have attained 11.O percent o f GDP in 2007. The 15 percent wage increase in 2006 was not included in the original Budget Law, but was incorporated in the revised budget. I t i s important that the Government develops coherent staffing and pay policies and that these policies are discussed and agreed upon during the preparation o f the original budget. The revised budget in 2006 to a large extent accommodated the large wage bill, and this has translated into lower execution rates for other categories o f expenditure. Capital expenditure has been rising rapidly over 2001-2006, but this has been driven to a large extent by an increase in donor funding. The ratio o f recurrent to domestically financed capital expenditure has fallen only modestly, from 6.4 in 2001 to 6 in 2006. The positive changes in the composition o f public expenditure in 2001-2006 reflect in part the allocation o f HIPC resources and increased donor funding. 2.59 As this section has highlighted, there i s a need to improve the comprehensiveness o f the budget and the quality of budget execution data. The Government i s moving in the right direction, and from 2007 SIGEFI has payment order data for the full wage bill. However, differences between budget execution reports and SIGEFI numbers suggest that there i s room for improvement. Likewise, the absence o f a system for tracking donor funding by ministry makes it difficult for the Government to prepare a budget with full knowledge o f overall sector allocations. The lack o f provincial-level public expenditure data means that it i s difficult to target expenditures to provinces which have the highest poverty rates. Although there has been marked progress in budget preparation and budget execution, there remains a continued need for strengthening these processes. 40When taken with respect to total domestically-financedthe spending share on security decreased. However, as a share of GDP, security spending increased. This apparent contradiction i s explained by the fact that total expenditure grew faster than GDP. 47 CHAPTER 3. FISCAL SPACE ANALYSIS AND SECTORAL EXPENDITURE REVIEW 3.1 The analysis o f public expenditure structure and trends in Chapter 2 i s complemented in this chapter by presenting detailed analyses on: (a) creation o f fiscal space; and (b) the budgetary allocation and efficiency aspects within three priority economic and social sectors - education, health, and infrastructure. Following a discussion on the fiscal space analysis, this chapter analyzes the trends in allocations o f public resources to these sectors over the past six years, and the extent to which they are in line with the development priorities o f the Government o f Burundi; and then analyzes the intrasectoral allocation o f public spending, including the relative weight and trends o f recurrent versus investment expenditures. Further, a brief discussion i s provided o n financial management issues in the sectors. The chapter i s organized as follows. Section A presents the fiscal space analysis building o n the findings o f the previous chapter and o f the intrasectoral allocation o f public expenditure; Section B presents a public expenditure review o f the education sector; Section C discusses budget allocation issues relevant to the health sector; and Section D addresses the intrasectoral allocation o f public expenditure in infrastructure. A. CREATING FISCALSPACE IN BURUNDI 3.2 Securing financing for Burundi’s development agenda requires creating/expanding its fiscal space. However, a legitimate question that arises i s what r o o m to maneuver the country has for this purpose? In other terms, h o w much additional fiscal space can be created under plausible scenarios for increased aid, domestic revenues, and economic growth? T h i s section addresses this question by first outlining trends in public expenditure, based o n the discussion o f the previous sections as w e l l as o f the preceding chapter, and foreign aid in Burundi. Burundi’s potential to create further fiscal space i s examined here. 3.3 This section discusses the scope for creating fiscal space for additional priority public expenditure over the next few years from the four components of the fiscal diamond: domestic revenue, grants, sustainable borrowing, and improvements in the efficiency of expenditure. The budgetary resource envelope for primary expenditures amounted to 40.3 percent o f GDP in 2006. This was composed o f domestic revenues o f 19.0 percent o f GDP in 2006, grants o f 18.0 percent o f GDP, and financing o f 3.3 percent o f GDP. The latter consisted o f external borrowing o f 0.7 percent o f GDP, domestic borrowing o f 2.3 percent o f GDP, and privatization proceeds o f 0.3 percent o f GDP. Grants 3.4 I n 2006, foreign grants to the Government budget were an estimated 18.0 percent of GDP. Program support accounts for the largest share o f the external grants, with an estimated 7.2 percent o f GDP in 2006, followed by project grants 4.4 percent o f GDP, HIPC relief 3.9 percent o f GDP, and special programs (DDR) 3.7 percent o f GDP. Future levels o f grant aid to the budget, in the form o f budget support and/or project aid, could be boosted if donors increase their overall aid to Burundi, in line with commitments made at international meetings to expand aid in 48 real t e r n s to the poorest countries, andor if a larger share o f donor aid were to be channeled to the Government budget. The latter will depend on the perceived quality o f public expenditure management as way o f mitigating fiduciary risks in the context o f a post-conflict country, especially with regard to budget support. A more coordinated action o f donors i s also recommended. 3.5 It has been Table 3.1 : Projections of Government Financing4' projected that in the short (percent o f GDP) term, aid inflows in the 2006 2007 2008 2009 2010 2011 form of grants will rise Financing (identified) 18.0 20.9 22.5 96.2 18.8 17.6 considerably, from 20.9 External grants External borrowing 2.4 0.6 0.7 2.3 1.5 1.1 percentOf GDP in 2007 to External borrowing, net 2.9 2.5 3.0 -89.1 1.3 1.5 22.5 percent of GDP in Privatization proceeds 0.3 0.1 0.3 0.3 0.2 0.1 2008 and 96.2 percent o f Domesticborrowing -1.4 -3.3 -0.5 0.7 1.1 0.8 Financinggap/errorsand GDP in 2o09 (due to omissions 0.0 0.0 -0.4 1.1 1.8 7.3 Burundi reaching the Source: World Bank staff estimates. HIPC Completion Point)42 43(Table3.1). The projection also indicates that grants will decline to 18.8 percent o f GDP in 2010 and 17.6 percent o f GDP in 201 1, as i t i s not certain how long the increase in aid volumes will last. Consequently, the Government o f Burundi, at some point in the future, needs to mobilize additional domestic revenues in order to offset a decline in aid flows. As a result, the projected increase in grants makes a significant contribution to fiscal space in the short term, while in the medium and long term; grants will only make a marginal contribution to fiscal space. In general, the volatile nature o f grants, as observed by the projected values, suggests that i t s contribution to fiscal space i s limited and unsustainable. Government Borrowing 3.6 Recent trends in borrowing. N e t borrowing to fund the Government's budget has been relatively modest in the last six years. During 2001-2006, net external financing averaged 4 percent o f GDP, and net domestic borrowing averaged 2.6 percent o f GDP. In addition to limited access, external borrowing was constrained by the need to restore external debt sustainability and to avoid the risk o f falling into the debt trap. Domestic borrowing, on the other hand, was constrained by the Government's intent to reduce the need for borrowing fkom the central bank. 4' I t i s worth noting that it i s difficult to make any feasible quantitative assumption about the volume o f budget grants over the long term, as the range of possible outcomes i s large. 42Note here that the projected figure does not reflect the net resources available for expenditure. Nevertheless, the debt relief (under HIPC) i s an important vehicle of resource transfer as i t may create space for much-neededexpenditure. 49 3.7 The Table 3.2: Government Financing, 2001-2006 little scope for the Net External Financing* 3.3 7.1 3.5 4.0 6.2 0.4 Government to Disbursements 5.0 10.3 7.0 6.4 8.6 4.3 borrow either from Amortization -5.1 -7.3 -6.7 -5.5 -3.9 -4.4 Change in the commercial amortization arrears 3.4 4.1 -1.4 -8.9 -1.7 -0.4 banks or from non- Debt relief 0.0 0.0 4.6 12.0 3.3 0.9 bank financial Net Domestic institutions. borrowing 2.5 -3.0 2.7 8.6 1.2 3.8 Bank sector 4.3 -1.9 2.4 7.8 1.9 3.6 Consequently, the Non-bank sector -1.8 -1.1 0.4 0.8 -0.7 0.2 macroeconomic Privatizationreceipts 0.0 0.0 0.0 0.0 0.0 0.4 imperatives dictate that Source: IMF. Note: * Net external financing includes both external borrowing and grants. domestic financing should be very low. Privatization receipts added 0.4 percent o f GDP to the budget’s financing resources in 2006 (Table 3.2). 3.8 External borrowing. The Government’s net external borrowing will decrease considerably over the short and medium term, as the main intent i s to make most o f external financing in the form o f grants only, which are estimated to make up about 90 percent o f total external financing. Net external borrowing i s estimated to fall from 2.9 percent o f GDP in 2006 t o 2.5 percent in 2007 and will then slightly increase to 3 percent o f GDP in 2008. In 2009, due to a significant decline in the adjustment for scheduled interest payments following the HIPC Completion Point, net external borrowing i s projected to fall to -89 percent o f GDP, before it increases very slightly to 1.4 percent o n average in 2010 and 201 1. 3.9 Despite the enhanced Table 3.3: Summary of Baseline Debt HIPC assistance, the magnitude o f Sustainability Indicators external financing over the period (with continued interim HIPC relief) 2007-20 17 will lift the net present Indicative Average value (NPV) o f Burundi’s public Threshold 2007 2017 2027 2007-27 NPV of debt to GDP (%) 30 85 38 32 45 or publicly guaranteed (PPG) external debt over two o f the five NPVofdebttoexports(%) 100 998 230 158 315 thresholds at which debt i s deemed NPVofdebt torevenue(%) 200 468 192 157 230 sustainable under the IMF/IDA Debt service to export (%) 15 35 15 10 20 L o w Income Country Debt Debt service to revenue (%) 25 17 12 10 15 Sustainability Assessment Source: Joint Bank-Fund Debt Sustainability Analysis, December 2007. Framework (see Table 3.3). The projections for external financing are adopted from the most recent Debt Sustainability Analysis (DSA), which was compiled in December 2007 (IMF and World Bank, 2007). In the short term, all debt sustainability indicators, except for debt-service-to- revenue, will breach the policy-dependent indicative thresholds. The projected net external finance over the short term will lift the NPV o f debt to export above the sustainability threshold, although the indicator will start to decline in 2008, and by 2027 it will fall significantly to 158 percent. The standard stress tests suggest that the evolution o f Burundi’s external debt position i s 50 subject to considerable vulnerabilities, as most debt indicators deteriorate significantly under the bound tests.44 3-10 hmestic borrowing- There are Table 3.4: Tax Revenue/GDP Ratios in Selected three potential sources o f growth in domestic SSA Countries, 2006 revenues as a share o f GDP; first, the income GDP per capita Tax revenue as elasticity o f some taxes, secondly, changes U S dollars percent of GDP Burundi 106.0 18.0 in tax policy (tax rates, thresholds, Kenya 455.8 15.6 exemptions etc), and thirdly, improvements Rwanda 268.0 12.4 in tax administration. Table 3.4 compares Tanzania 335.0 12.9 Uganda 274.0 12.4 the tax revenue collected in Burundi with Sub-Saharan 583.4 21.2 that o f other East African countries. Despite Africa Source: World Bank World Development Indicators Database, itslow income per capita,~ ~is ~ among ~ d i 2006; African Economic Outlook 2005-2006 (OECD) for Kenya’s E A C countries the country with the highest taxfigure. tax revenue t o GDP ratio and approaching the sub-Saharan African (SSA) average o f 21.2 percent4’ As discussed earlier in this chapter, the high tax revenue GDP ratio i s mainly explained by the narrow nature o f the tax base. Consequently, creating fiscal space through increased domestic revenues requires improving tax administration. 3.11 The main challenge to the Government i s the need to maintain macroeconomic stability without crowding out private sector borrowing from the banking system. Government’s domestic debt stock i s estimated at 24.6 percent o f GDP in 2006, and this debt carries a fixed interest rate o f about 5 percent; hence the sustainability o f the domestic debt stock i s also another area o f concern o n domestic borrowing. Besides the existing high level o f domestic debt stock, issuance o f further debt could be constrained by two factors: (a) in the short term, broad money i s expected to grow at 12.4 percent (M2) and 14 percent (M3), which i s significantly above nominal GDP growth (10.1 percent); hence domestic financing o f the budget will be provided principally through the issuance o f new Treasury securities rather than by recourse to central bank advances (IMF, 2007); and (b) the lack o f significant non-bank financial institutions. The poorly developed domestic financial sector also constrains the use o f alternative macroeconomic instruments to create fiscal space. I f the Government were to fund a more expansionary fiscal policy by borrowing from the central bank, the latter would then face the need to sterilize the reserve money thus created, but i t s ability to do so i s constrained by the same characteristics o f the domestic financial system which restrict the scope for Government borrowing from the domestic market. 44The bound tests include (a) real GDP growth at historical average minus one standard deviation in 2008-09; (b) export value growth at historical average minus one standard deviation in 2008-09; (c) US$GDP deflator at historical average minus one standard deviation in 2008-09; (d) net non-debt creating flows at historical average minus one standard deviation in 2008-09; (e) combination o f 1-4 using one-half standard deviation shocks; and (0 one-time 30 percent nominal depreciation relative to the baseline in 2008. 45T h e tax-GDP ratio for sub-Saharan Africa also includes South Africa, which has one o f the highest ratios. When South Africa i s excluded, the tax-GDP ratio for the sub-region i s estimated at 18.8 percent in 2006. 51 PrivatizationReceipts 3,12 The Government budget received modest inflows from privatization receipts, 0.4 percent of GDP in 2006, which i s expected to increase very slightly to 0.5 percent of GDP in 2007 and 2008. This was the result o f a program o f privatization o f mainly small-scale public enterprises. Further privatization o f state assets in the financial and coffee sectors was launched in 2007, and 50 percent o f the receipts above their projected amount will be used to reduce the domestic borrowing requirement. Although there are s t i l l several large-scale enterprises remaining in the public sector portfolio, such as Office National des Telecommunications (ONATEL), Burundi Coffee Company (BCC), S o c i i t i Immobili6re Publique (SIP), it i s not yet possible t o make realistic projections about what proceeds they could generate, as i t i s not yet clear what further plans the Government has for privatizing these public enterprises. The Efficiency o f Public Expenditure 3.13 I n principle, fiscal space can be created by improving the allocation and technical efficiency of Government’s expenditures, thereby allowing more outputs in terms o f public goods and services to be produced for a given amount o f budgetary resources. However, it i s very difficult to quantify in a rigorous manner the magnitude o f expenditure inefficiency and the potential scope for creating fiscal space through efficiency improvements. Reducing less productive and lower priority spending and increasing expenditure efficiency are the other alternatives for creating fiscal space. 3.14 Allocative efficiency of Government expenditure in Burundi i s likely to be low because budget allocations are still heavily influenced by those which prevailed during the past few years. As discussed in this chapter and in the preceding one, these allocations were heavily biased towards recurrent expenditures, highly centralized service provision, and very rigid expenditure structures. I n 2006, while Government spent around 19 percent o f total expenditure o n defense, i t s spending on priority sectors, excluding education, amounted only 15 percent o f total expenditure. In the education sector, recurrent expenditures account for 75 percent o f the total sectoral expenditure, mostly explained by the weight o f the wage bill, which consumes 67 percent o f the sectoral recurrent expenditure, limiting funds for teaching materials and essential school maintenance. Meanwhile, capital expenditures amount to only 25 percent. The weaknesses in the budget process for allocating expenditures, discussed in Chapter 4, are a constraint to improving the allocative efficiency o f Government expenditures. In addition t o allocative efficiency setback, the overall inefficiency o f education i s another area o f concern. Public expenditure returns on education i s lower than the SSA’s average (see section on Education, below). 3.15 I t i s also likely that technical efficiency i s low in Burundi, given that many aspects of the country’s public expenditure management system are weak. For example, transparency and accountability for spending o f public funds are weak. As pointed out in the next chapter, there are structural and functional weaknesses o n the overall monitoring o f public expenditure in Burundi. T o begin with, only few line ministries have internal audit units; and those that do are heavily inhibited by the lack o f adequate and qualified expertise. 52 3.16 Another potential source of intra-budget fiscal space may be defense spending. Although it i s not the purpose o f this report to assess whether the defense budget corresponds to the perceived threats to national security and the Government’s goals and intentions, international comparison i s relevant. Can Burundi afford t o continue increasing military expenditures, especially now, during a time o f severe fiscal adjustment, or do these expenditures contain a potential source o f fiscal space in the short term? Although Burundi has decreased its armed forces, the number o f police force has increased significantly. This represents a costly payroll problem for a country that needs t o make efforts to streamline i t s wage payroll. 3.17 I n summary, to increase public investment, it will be necessary either to increase the level o f fiscal revenues (which are already a respectable 18-20 percent of GDP) or foreign financing or, ceteris paribus, to constrain domestically financed recurrent expenditures. I n that connection, the government wage bill (civilian and security), which has risen t o about 10 percent o f GDP in 2006 budget, i s on the high side compared to other African countries, with services and transfershubsidies representing another 10 percent or so. Since increasing the tax ratio above i t s existing level i s likely t o be neither easy nor perhaps even desirable o n economic grounds, increasing public investment will continue to be constrained by the availability o f foreign savings unless current expenditures are constrained and their efficiency improved. Thus, public sector structural reform appears likely t o be a subject o f considerable importance in Burundi for the foreseeable future, as the proceeds o f privatization would provide additional fiscal room to the Government. 3.18 Fiscal discipline i s a critical determinant o f macroeconomic stability, mainly because the use o f domestic bank borrowing to finance the deficit i s always inflationary, while excessive external borrowing will create an unsustainable debt burden. I t i s therefore important to restrict aggregate government expenditure even though this will have a crucial bearing o n sectoral spending limits. The Government has t o determine an aggregate spending level that i s in line with maintenance o f a stable macroeconomic situation. Intrasectoral allocation issues o f public expenditure to education, health, and infrastructure are discussed in the subsequent sections. B. EDUCATION Box 3.1: Key Sector Development, Budget and Financial Management Issues Despite a dramatic recovery o f gross primary school enrollment ratios in the 2000s, Burundi’s education indicators remain below the average for Sub-Saharan Africa. Regional disparities are large and high drop- out and repetition rates affect the efficiency o f the system. The basic quality o f Burundi’s education compares favorably with that o f many other low-income African countries. However, delays in the school construction program have led to a major expansion o f the double-shift practice, which may affect quality and learning achievements. Progress achieved i s largely due to a major increase in public spending on education. The use o f HIPC finds, large and growing donor programs, and household contributions are also a factor. The structure o f public spending i s generally adequate (high priority to primary education), but the share o f teaching materials and other goods and services i s dangerously low. Budget planning i s not based o n a well-defined sector strategy, and overly complicated budget execution procedures encourage Government officials and suppliers to use exceptional procedures that bypass existing controls. 53 Sector Performance and Outcomes 3.19 The education sector was severely Table 3.5: Gross Enrollment Ratios from 2000-01 to 2006-07 2000- 2001- 2002- 2003- 2004- 2005- affected by the political 01 02 03 04 OS 06 and economic crisis o f primary 61% 68% 74% 80% 85% 101% the 1990s. Gross primary Lower Secondary 11% 12% 13% 13% 16% 17% Upper secondary 3.2% 3.3% 3.4% 3.6% 4% 4% school enrollment ratios Tertiary (number of 122 139 186 215 258 (GERs) from 73 students per 100,000 percent in 1991 to less inhabitants) Source: Ministry of Education. than 50 percent in 1994- 97. The impact on secondary and tertiary education was less significant. Secondary education GERs continued to progress from 5.5 percent in 1991 to 6.6 percent in 1994-95. 3.20 Improvements in the political and economic environment led to a dramatic recovery in the 2000s (Tables 3.5-3.6). Primary school GERs increased from 61 percent in 2000-01 to 85 percent in 2004-05. Secondary education GERs also went up to 11 percent in 2002-03 and 14 percent in 2004-05. In 2006, the Government abolished school fees for primary education. The measure had a strong and immediate impact on enrollments, and the number o f primary school students jumped from 970,000 in 2003-04 to 1,350,000 in 2005-06, bringing the gross enrollment rates for 2005-06 to 101 percent. 3.21 Although Burundi has made great progress in terms o f primary education, Burundi’s secondary education indicators remain below the average of Sub-Saharan Africa. In 2004-05, primary school GERs were 10 percentage points below the SSA average o f 95 percent. With a 100 percent GER in 2005-06, Burundi may have caught up with most o f the other low-income African countries. However, with a secondary education GER o f 14 percent in 2004- 05, Burundi remains far behind the SSA average o f 32 percent. Table 3.6: Gross and Net Enrollment Ratios in Burundi and in a Sample of Low-Income African Countries, 2006 Total Male Female Total Male Female Gross Primary Enrollment Gross Secondary Enrollment Burundi 103.2 108.1 98.4 Burundi 14.3 16.4 12.1 SSA 93.8 99.3 88.2 SSA 31.9 35.4 28.4 LIC LIC Africa* 97 100.9 92.9 Africa* 24.9 26.7 23.1 Conflict** 140 137.5 142.4 Conflict** Net Primary Enrollment Net Secondary Enrollment Burundi 74.6 75.9 73.3 Burundi SSA 69.5 72.3 66.6 SSA 25.1 27.4 22.9 LIC LIC Africa* 73.3 76.3 70.2 Africa* 19.4 20.5 18.4 Conflict** Conflict** Note. * Benin, Burkina Faso, Chad, Ghana, Kenya, Niger and Uganda, plus Madagascar and Tanzania only for primary education, ** Sierra Leone, Rwanda. Source: Ministry o f Education; World Development Indicators. 54 3.22 Gender disparities are large, but the most critical disparities are regional. I n 2004- 05, the primary school GER was 91.4 percent for boys and 78.4 percent for girls, and the gender disparity indicator was 0.86, close to the SSA average o f 0.89. I n 2003-04, GERs in primary education reached 13 1 percent in Bujumbura but were a l o w 66 percent in the very poor Eastern provinces. Improving the regional distribution o f schools and teachers should be one o f the main objectives o f future education policies. 3.23 The rapid progress o f gross enrollment ratios hides major structural weaknesses. The efficiency o f the primary education system i s reduced by high dropout and extremely high repetition rates. Burundi’s repetition rate (30 percent) i s by far the highest in the selected sample o f comparators. Consequently, retention and completion rates (respectively 37 and 35 percent) are very l o w and the index o f internal e f f i ~ i e n c yo~ the primary education system was only 47 f~ percent in 2003-04. This means that more than half o f the resources invested in the sector are wasted on students who will not complete the cycle. 3.24 The basic quality of Burundi’s education system compares favorably with that o f many other African countries. Standardized Modern Language Association (MLA) tests show that the scores o f grade 5 students far exceed averages in other low-income countries. However, the massive increase in primary school enrollments in 2005-06 was not accompanied by an appropriate effort to provide the necessary classrooms and teaching materials. The double shift practice, which now affects 69 percent o f primary schools, substantially reduces teaching times available to students and i s likely to lower quality and learning achievements. Public Spending on Education - Impact of HIPC Initiative 3.25 Public spending on Table 3.7: Public Spending on Education education increased substantially in 2001-2005. As I n FBu Billions Percent Total Public Percent o f GDP Expenditures shown in Table 3.7, Budget Execution Budget Execution Budget Execution Government budget allocations 1991 18 3.5 increased in nominal terms but 2~01 20.6 20.9 15 15 3.8 3.8 remained stable at around 15-16 2002 22.6 22.3 13 14 3.9 3.8 percent of total public 2003 28.6 27.9 16 16 4.4 4.3 expenditures from 2001 to 2004 34.4 33.8 15 15 4.7 4.6 2005. The increase in actual 2005 42.1 43.5 16 19 4.9 5 spending was more dramatic, 2006 78.5 65.1 22 22 8.4 7 with generally high rates o f 2007 85.5 nla 25 nla 8.3 n/a execution reaching a peak of Source: Ministry o f Education and World Bank calculations. Data on execution are 103 percent in 2005. The share basedoncommitmm. o f the education sector went up from 15-16 percent o f total public spending during the 2001-04 period to 19 percent in 2005. 46 The index o f efficiency compares the number o f years theoretically needed to bring one student through the cycle to the number o f years actually needed to achieve that result, taking into account dropout and repetition rates. 55 3.26 Access to HIPC funds led to a sharp increase in budget allocations, which went up to 22 percent o f total budget allocations in 2006 and to 25 percent in 2007. Despite a slight decline in the rate o f execution, actual spending also reached 22 percent o f total public spending in 2006. The best measure o f the high priority given by the Government to the sector i s the share o f GDP allocated to public spending on education, which went up from 3.8 percent in 2001 to more than 8 percent in 2006 and 2007, much higher than in most low-income African countries. 3.27 High unit costs limit the positive impact o f high public spending. Results obtained are affected by the relative inefficiency o f education expenditures. In 2003-04 for every one percent o f GDP spent on education, Burundi obtained an average o f 1.1 years o f schooling for its population, against an average o f 1.9 years in Sub-Saharan Africa. 3.28 I n 2006, FBu 18.4 billion, Le., 47 percent of HIPC funds available to Burundi, were allocated to education, and about 90 percent o f that allocation was expected to finance the construction o f primary schools. Two factors delayed implementation o f the school construction program: (a) it took time for the Ministry o f Economy, Finance, and Development Cooperation to decide about the distribution o f HIPC funds and to make available to the Ministry o f Education its share o f the resource; and (b) complex commitment and payment procedures also slowed down construction work.47 Delays in implementing school construction exacerbated the school capacity crisis intensified by the dramatic increase in school enrollments in 2006 and 2007. In 2007, only FBu 9.8 billion, i.e., one third o f HIPC funds for Burundi, was allocated to the education sector. External Assistance and Private Contributions 3.29 Donor financing accounts for a large and growing share o f public spending on education. The Government does not have the instruments necessary to monitor effectively donor programs and disbursements, and Table 3.7 does not take into account donor contributions. According to the vice Ministry for Planning (VMOP), donors were expected to contribute FBu 11.3 billions to the sector in 2003, i.e. close to 40 percent o f Government budget allocations, but actual commitments and disbursements did not exceed FBu 5.7 billions, adding only 20 percent to actual Government spending on education. There i s evidence that donor assistance to the education sector increased significantly since 2003. According to unofficial sources donor contributions may now account for close to one third o f total public spending on education. With the help o f i t s main partners, the Government should take appropriate measures to improve the planning and monitoring o f donor programs and better coordinate donor and Government operations in the sector. Three bilateral donors have decided to merge their contributions within a Common Education Fund. This i s a sound initiative that should be tested and then extended if proven effective. 3.30 Contributionsfrom households are substantial. Before the abolition o f school fees in 2006, households contributed FBu 1,500 per primary school student per year. I n 2004 school fees totaled FBu 2.2 billion, including 1 billion for primary and 1.2 billion for secondary education; i.e., the equivalent o f 7 percent o f total Government spending on education (current and 47M o s t o f the school construction contracts were awarded to small enterprises with limited financial capacity, which had to suspend their operations when payments were delayed. 56 investment expenditures), Managed outside the government budget, these funds were primarily used to finance school books and other non-wage current expenditures. Now that school fees for primary education have been abolished, the Ministry o f Education i s expected to transfer to provincial authorities and school facilities frais de compensation equivalent to amounts previously contributed by households. However, since the elimination o f school fees led to a dramatic increase in enrollments, these transfers are insufficient. Consequently, school facilities and local authorities continue to mobilize contributions fi-om households to finance critical current expenditures (see B o x 3.2 which provides a preliminary evaluation o f the impact o f the presidential decision to eliminate school fees in the primary). Box 3.2: Impact of the Provision of Free Primary Education in Burundi The decision to provide free primary education for every child in Burundi's public schools, announced in August 2005 b y the Head o f the State, has led to a significant increase in the enrollment rate. T h i s decision also led to an increase in the resources allocated to education to accommodate the flow o f new enrollments. The 2006 Core Welfare Indicators Questionnaire (CWIQ) survey could be used to assess the impact o f free primary education o n children schooling since the households surveyed were asked about their children schooling before and after the reform. According to the survey analysis, more than 97 percent o f households were aware o f the free primary school decision. If school were not free, 41.4 percent o f the poorest households (top quintile based o n the consumption per equivalent adult household) reported that some o f their children who are in primary school today, would have stayed at home. T h e contrasting figure for the richest household i s only 13.1 percent. Before the school year o f 2006, 32.1 percent o f the children from the poorest quintile did not attend school due to the lack o f money to pay for their fees. Therefore, it i s obvious that the provision o f free primary education has made a major impact on the schooling o f children. The questionnaire also helps to identify precisely who has benefited more from the provision o f free education. For example, 3 1.8 percent o f children, between the age o f 6 and 11, were enrolled in school since the introduction o f free education. Among these newly enrolled children, 28.2 percent are from the poorest quintile, 26 percent from the second quintile, 19.8 percent from the third, 16.6 percent in the fourth, and 9.4 percent are from the more affluent households. The profile o f new school children also shows that the impact o f free education was greater on (a) relatively young children; (b) children who do not work (we have to bear in mind here that the decision to work or not depends partly o n the decision to be i n school or not); (c) children from rural vs. urban areas; (d) children belonging to poor households; (e) children belonging to households with at least one member with a disability; (0 children-headed households had a limited education; and (g) household heads involved in agriculture (as compared to industry and services). By contrast, there i s relatively lit t le difference in the impact, depending o n whether the child has been displaced due to war, according to the gender o f household head or marital status, gender o f child (boy and girl), whether the child i s an orphan or not, and depending on the remoteness o f the child's living area i n relation to different services. Thus, the main result i s not surprising: free schooling has been a pro-poor measure. However, although free primary school has helped to increase the enrollment rate o f children, it does not mean that all children o f primary school age are enrolled. Moreover, the cost o f education continues to be an obstacle for some households, especially after the primary level. Thus, cost i s a key impediment to sending children to school, as indicated by parents for both sexes and for both age groups, especially among the poorest households. For older (high school age) children, cost remains the most fundamental obstacle. I t i s also important to note that the perception o f needlessness o f education i s the main reason for girls aged 12 to 18 years not attending to school, whereas t h i s i s not the case for boys o f the same age. This suggests that there are s t i l l strong cultural barriers to the acquisition o f education that goes beyond primary education, and that the impact i s greater on girls than boys. Source: Estelle Sommeiller and Quentin Wodon (2007): Impact de 1'dducationprimaire gratuite sur la scolarisation au Burundi, World Bank, Mimeo. 57 3.3 1 I n several areas of the country, parents have created associations and management committees whose role i s to provide advice to school managers and local authorities on the use of school fees. The Government should support and help extend these initiatives. User representatives should be systematically involved in the planning, management, and monitoring o f both school fees4' and public funds allocated to their communities. Structure o f Public Spending - Functional and Economic Classification 3.32 The functional distribution of public spending Table 3.8: Structure o f Current Expenditure in Education - Functional Classification on education tends to improve (In FBu Billions Unless otherwise indicated) (Table 3.8). The Government i s 2001 2002 2003 2004 2005 2006 giving a high priority t o primary General 328 550 367 321 540 623 services education, and the recent increase %o f total 1.6% 2.5% 1.3% 1.0% 1.3% 1.1% in the education budget benefited Preschooling 8,105 8,416 11,673 13,871 19,018 27,751 mainly primary education, the and primary %oftotal 40.6% 38,9% 42.4% 43.0% 44.4% 48.9% share o f which rose from 41 Secondary 5,302 5,718 7,021 7,567 10,288 12,904 percent o f total education spending %oftotal 26.5% 26.5% 25.5% 23.4% 24.0% 22.7% in 2001 to 49 percent in 2006. This Technical& 962 1,064 1,116 1,180 1,770 2,018 was achieved by reducing vocational %o f total 4.8% 4.9% 4.1% 3.7% 4.1% 3.6% significantly the share of Tertiary 4,935 5,450 6,916 8.85 10,673 12,719 secondary and tertiary education. %oftotal 24.7% 25.2% 25.1% 27.4% 24.9% 22.4% Despite i t s recent decline, the Kamenge 363 411 462 497 542 775 share o f the tertiary sector remains Hospital (MedicaI high considering the small number school) of students attending the % oftotal 1.8% 1.9% 1.7% 1.5% 1.3% 1.4% university. Total 19,994 21,609 27,554 32,285 42,830 56,791 Note: Data on total current expenditures exclude activities outside the education sector (training, literacy and cultural programs). Except for 2007, figures are budget execution data (on commitments basis). Source: Ministry of Education 3.33 Up to 2005, the quasi totality of the education budget financed current expenditures (Table 3.9). As HIPC funds were allocated to investments, the share o f capital expenditures in budget allocations went up from 1.4 percent in 2005 to 12.8 percent in 2006. However, the rate o f execution o f capital expenditures fell to 43 percent in 2006 f r o m a rate o f 71.9 in 2005 (Table 3.10). Therefore, their share in actual spending increased only to 12.8 percent in 2006 (excluding donor-financed projects). 3.34 Wages dominate. The share o f salaries and wages in budget allocations to current expenditures reached a peak o f 73.9 percent in 2005 but declined to 67.9 percent in 2006. This, however, underestimates the real cost o f personnel expenditures, since transfers and subsidies also include a major salary component. 48 School fees have not been abolished for secondary education. 58 3.35 The share o f teaching materials and other goods and services in budget allocations and actual spending i s very low (3-4 percent o f allocations to current expenditures up to 2005 and 5.9 percent in 2006). Until 2006, school books, teaching materials and other goods and services were partly financed by school fees (outside the budget). Most o f the increase in goods and services in 2006 i s due to a large allocation tofrais de compensation to help schools pay for expenditures previously financed by school fees. Future programs should include a major increase in funds allocated to teaching materials and other goods and services. - Table 3.9: Structure o f Education Spending Economic Classification Current Transfers & Capita' Expenditures Wages Goods & services subsidies Expenditures FBu millions/% FBu millions/% FBu millions/% FBu millions/% FBu millions/% total expenditure current expenditure current expenditure current expenditure total expenditure 200 1 19,995 12,804 610 6,581 863 95.9% 64.0% 3.1% 32.9% 4.1% 2002 21,610 13,602 954 7,054 724 96.8% 62.9% 4.4% 32.6% 3.2% 2003 27,554 18,395 864 8,295 357 98.7% 66.8% 3.1% 30.1% 1.3% 2004 32,282 21,055 912 10,315 550 98.3% 65.2% 2.8% 32.0% 1.7% 2005 42,830 3 1,665 1,416 9,749 627 98.6% 73.9% 3.3% 22.8% 1.4% 2006 56,791 38,074 3,346 15,371 8,317 87.2% 67.0% 5.9% 27.1% 12.8% 2007 74,632 5 1,670 7,258 15,704 10,849 87.3% 69.2% 9.7% 21.0% 12.7% Note: Data on total current expenditures exclude activities outside the education sector (training, literacy and cultural programs) which are included in the estimated cost of specific components (wages, goods and services and transfers and subsidies). Except for 2007, figures are budget execution data (on commitments basis). Source: Ministry of Education. Financial Management Issues 3.36 Improving the performance o f the education sector will require major improvements in budget planning and execution procedures and practices. One o f the main bottlenecks i s the lack o f administrative and financial management capacity within the Ministry o f Education. The creation o f a directorate for administrative and financial management should be envisaged in the very near future. 3.37 Budget planning should be based on a coherent sector strategy aimed at better coordinating Government budgets and donor programs. Budget preparation starts when the Ministry o f Economy, Finance, and Development Cooperation sends to line ministries the Budget Framework Letter (Lettre de cadrage) about the next budget. Central and provincial directorates submit budget proposals, which are reviewed by the Minister's Office and discussed with the Finance Ministry. Budget proposals prepared by the Ministry o f Education focus on staffing needs, are not based on clear strategic objectives, and do not take into account resource constraints, since the Budget Framework Letter does not include indicative budget ceilings for 59 each sector and each ministry. Final budget allocations, including decisions on intra-sectoral priorities, are made by the Ministry of Economy, Finance, and Development Cooperation and usually reproduce previous budget allocations, with adjustments for price increases. The selection and design o f donor-financed investment projects i s done o n an ad hoc basis and does not reflect comprehensive sector strategies. 3.38 The Government i s preparing a 2008-2010 Action Plan, in cooperation with its development partners, and education i s one o f the sectors that plans to gradually introduce program budgeting. Future programs and program budgets will be dominated by five main objectives: (a) universal primary education by 2015; (b) regulate access to post-primary education; (c) improve the quality o f the education system; (d) make it more equitable; and (e) increase its efficiency. A good information system and a school map will be the main instruments o f the policy. Improving the efficiency o f the system will largely rely on reducing the repetition rates. 3.39 High rates o f execution seem to indicate that the overall budget execution performance o f the sector i s generally adequate. While rates o f execution are l o w for capital expenditures - whether financed by the budget or by donors - rates o f execution for current expenditures are generally high, as shown in Table 3.10.49 Table 3.10: Rates of Execution for Public Expenditures in Education (in percentage o f budget allocations) 2001 2002 2003 2004 2005 2006 Current Expenditures 99.5 101.0 97.8 95.4 103.9 96.7 Wages 99.7 103.1 96.8 93.8 105.7 95.9 Goods and Services 91.5 125.4 99.8 88.3 96.5 96.4 Teaching materials 51 99 100 90 IO0 61 Transfers & Subsidies 100.0 94.8 100.0 99.5 99.4 98.7 Capital Expenditures 97.4 58.8 88.1 95.7 71.9 42.8 TOTAL 99.4 98.7 97.7 95.4 103.2 83.2 Source: MEFDC. 3 -40 Nevertheless, the Ministry o f Education argues that budget execution procedures need to be streamlined, a s thirty steps and four controls are necessary to move from the commitment to the payment phase. Since the Ministry o f Education lacks expertise in accounting and financial management, inadequate handling o f any o f these phases i s a cause o f delays and encourages Government officials and suppliers to use exceptional procedures, which bypass most o f the controls. 49An important exception i s teaching materials (within goods and services), whose rates o f execution fell to a low 61 percent in 2006. 60 Box 3.2: Main Recommendations in the Education Sector Short term priorities: Assess the budgetary implications o f the September 2005 decision abolishing primary school fees to adjust budget appropriations needed to finance costs associated with free primary education; Increase by at least 20 percent in real terms budget transfers to schools and actual spending on other goods and services, with special emphasis on teaching materials; 0 Accelerate implementation o f the primary school construction program (at least 1200 schoolrooms per year), and better coordinate this program with the allocation o f human resources,’ Based on an adequate information system, including a good school map, the strategy and the plan should produce a more equitable regional distribution o f infrastructure and teachers. Special attention should be accorded to areas where primary education GERs are below 80 percent. However, assessing the budgetary impact o f providing special incentives to teachers willing to work in distant rural areas is a prerequisite for the success of such a measure; Medium to long term priorities: Strengthen the financial management capacity o f the Ministry of Education (at the central and provincial levels) and create a directorate responsible for administrative and financial management; Improve the eficiency o f the education system by reducing repetition ratesfrom 30 to 15percent along with a reduction in unit costs o f the education system; 0 Pursue ongoing efforts to develop, in cooperation with donors, a comprehensive sector strategy, which w i l l serve as a basis for the gradual introduction o f program budgeting and better coordination o f donor programs; Monitor the creation o f the Common Education Fund. Encourage other donors to evaluate the initiative and eventually j o i n the program. 61 c. HEALTH Box 3.3: Key Sector Development,Budget and Financial Management Issues Despite recent progress, Burundi’s health indicators remain low. The lack o f qualified personnel and its concentration in Bujumbura are the main factors behind the poor performance o f the sector. Domestically financed public expenditures stagnated in share o f GDP and percentage o f total public spending in 2001-05. I n 2006, access to HIPCfunds had a positive impact on health budgets and actual spending, despite low rates o f execution and cumbersome disbursement procedures. The present structure o f current expenditures indicates a bias in favor of hospital care. Donors and NGOs are the main sources offinance for the health sector, but the Government is unable to plan and monitor their programs and their activities. Households also make a substantial contribution. The recent decision to eliminate user fees for specific services was sound but poorly implemented. 0 Despite the creation o f the General Directorate o f Resources (DGR), a new department responsible for budget planning and execution within the Health Minisoy, the lack offinancial management capacity at the central and provincial levels and in primary health care (PHC) facilities is a major problem. 0 The Government and its partners hope that the ongoing strategic planning activities and improved donor coordination w i l l help address the most critical development andfinancial management issues in the sector. Sector Performance and Outcomes 3.41 After fifteen years o f decline, health indicators improved in the 2000s. Ten years o f conflict had a devastating impact on the health o f the population. Infant mortality rates doubled since the late 1980s/early 1990s. Child mortality and the percentage o f stunted children increased by close to 50 percent. The end o f the conflict and measures initiated during the transition improved health outcomes. Infant mortality rates declined from 153/1,000 to 114 in 2005. Child mortality rates also declined from 231/1,000 to 190 and the percentage o f stunted children declined from 57 to 44 percent. 3.42 Despite this progress, health indicators in Burundi remain poor. As shown in Table 3.11, infant and child mortality rates are lower than in four other post-conflict countries, but higher than the average for Sub-Saharan Africa.50 50 Data on maternal mortality rates and the percentage o f stunted children are less relevant, since they date back to 2000 and do not take into account recent improvements in Burundi’s indicators. 62 3.43 One of the main factors behind the poor performance of Table 3.11: Health Outcome Indicators in Burundi and other African Countries Burundi's health sector i s the lack Child Infant Maternal Moderatelsevere of qualified personnel. In 2003, mortality mortality mortality stunting Burundi had only 343 doctors and per 1,000 per 1,000 per in YO of under (in 2005) (in 2005) 100,000 five children (in 2101 nurses (5 doctors and 29 nurses live 2000) *** per 100,000 people), one o f the births lowest ratios in the world. The lack o f (in 2000) *** adequate training capacity i s largely Burundi 190 114 1,000 57 responsible for the shortage o f SSA 169 101 940 38 average physicians. Until recently, the Post conflict 231 142 1,288 38 medical school in Bujumbura countries* produced only 20 doctors per year, Other 145 86 1,128 41 African LIC much less than the number required ** to meet a deficit estimated at 250.51 * Liberia, Sierra Leone, DRC and Rwanda; *, Benin, Burkina Faso, Chad, Another factor i s the decline in the Ghana, Kenya, Madagascar, Niger, Tanzania & Uganda; *** Available data on maternal mortality and percentage of stunted children remuneration o f public health in comparator countries date back to 2000. personnel, which now i s below the Source: UNICEF child info website: www.childinfo.org. average in a sample o f 6-7 other low- income countries (Table 3.12). 3.44 More most Of Table 3.12: Salaries of Health Care Personnel in Multiples the qualified personnel are of GDP per Capita concentrated in Bujumbura. The General Certified practitioners nurses lack o f security led many doctors and Burundi 5.8 - 12.2 3.6 - 5.4 other health staff to leave rural areas Average of 6-7 otherlow-income 8.3 - 15.7 5.2 - 8.5 African countries * during the conflict'As a about Burkina Faso, Cameroon, Chad, Ethiopia, Mauritania, Niger, and DRC (for 80 percent o f the doctors and 50 doctorsonly). Source: Burundi data are based on salaries of health care personnel in May percent Of the n the 2006. Data for other countries are from the World Bank's Human i capital city. I n the very poor Development, Africa Region and Country Status Reports. provinces o f Kirundo and Muyinga, the ratio o f nurses per 100,000 people i s less than 20. 3.45 Expanding and improving training facilities and providing incentives for health personnel working outside Bujumbura should be two main priorities of future health policies. A few measures have already been taken by the Government, and the private sector: (a) the number o f students attending the medical school o f Bujumbura i s increasing; (b) three private universities have begun to train doctors and will produce graduates in 3-4 years; and (c) the Government has also begun to provide primes d 'doignement (distance allowance) for doctors who are willing to work in district hospitals. A private university i s also training doctors, but until recently its diplomas were not certified. 63 Public Spending on Health and HIPC Fund 3.46 The Dolitical crisis of the 1990s led to a drastic decline in public Table 3.13: Public Spending on Health I n 2001 constant In Percent of In Percent of spending on health. Budget allocations FBu (billion) Total Public GDP and actual spending increased modestly Expenditures in nominal terms in 2001-2005 but Budget Actual Budget Actual Budget Actual 1993 nla n/a nla 4.4 nla 1.3 stagnated or declined as a share o f GDP 2ooo nla nla 2.3 nla 0.6 and total public expenditures (Table 2001 3.7 3.7 4.2 4.1 0.7 0.7 n 2002, Burundi was the African 3.13). I 2o02 3.9 3.8 3.9 3.9 0.7 0.7 2003 3.5 3.5 3.6 3.5 0.6 0.6 country with the lowest percentage o f 2004 4,3 3.8 4.2 3.6 0.7 0.6 public expenditures for the health 2005 3.8 3.6 3.6 3.3 0.6 0.6 2006 13.3 6.7 9.3 4.6 2.5 1.1 sector. Source: MEFDC and World Bank estimates. Note: Data on actuals are based on commitments. Total 3.47 Access to the HIPc Initiative expenditureshudget allocation excludes debt service (debt payments and had a strong positive impact on health amortization). sector spending. In 2006, most o f the proceeds o f the HIPC Initiative went to health and education. HIPC-financed budget allocations amounted to FBu 10.2 billion, equivalent to 43 percent o f the health sector budget, which went up from 3.1 percent o f total budget allocations (domestically financed, excluding debt service) in 2005 to 9.3 percent in 2006. The increase in actual spending was less dramatic, as the rate o f disbursement o f HIPC-financed expenditures did not exceed 30 percent and the rate o f budget execution in the health sector - which was high in 2001-05 - fell sharply in 2006 (about 50 percent). This was due to three main factors: (a) it took time for the Ministry o f Economy, Finance, and Development Cooperation to decide about the distribution o f HIPC Funds and to make available to the Ministry o f Health i t s share o f the resource; (b) the adoption o f overly complicated reimbursement procedures for the cost of services that in previous years were financed by user fees;52 and (c) the lack o f financial management and procurement capacity in the Ministry o f Public Health (MSP), which before 2006 was using most o f i t s resources on salaries and other current expenditures. 3.48 The poor disbursement performance of HIPC funds in 2006 led to a reduction in the allocation of HIPC funds for health and an overall reduction of budget allocations to the sector in 2007. Nevertheless, HIPC funds s t i l l account for close to 48 percent o f budget allocations to the sector and total budget allocations remain much higher than in 2005 in nominal terms, as a share o f total public expenditures and as a share o f GDP. The share o f investment in HIPC-financed expenditures went up to 36 percent in 2007. External Assistance and Private Contributions 3.49 Donors and NGOs have become the main source of financing for the health sector. The Government i s unable to effectively monitor donor and NGO contributions to the health sector and data on public health spending in Table 3.13 do not include donor and NGO-financed 52One o f the m a i n objectives o f HIPC financing for health in 2006 was to help finance the reduction o f user fees for specific health services (women in delivery and children under 5), as decided by the Government in 2005. 64 expenditures. However, available estimates seem to indicate that per capita ODA allocations to Burundi are among the highest in Sub-Saharan Africa. In 2006, MSP hoped that planned donor and N G O programs - including emergency assistance - would be 5-6 times higher than Government budget allocations, but actual spending was considerably lower, as rates o f execution o f foreign-financed projects are very poor. Current projections indicate a possible decline in donor assistance, perhaps due to the poor performance o f ongoing projects. Should this happen, the Government would need to take radical measures to increase budget allocations, improve execution, and optimize the structure o f health spending. 3.50 Household contributions also account for a significant percentage of health spending in Burundi. As public spending decreased in the 1990s, households increased their own spending. According to the 2006 Household Survey (CWIQ), their contribution was estimated at US$8.6 per capita and represented about 5.6 percent o f total household consumption. This i s much higher than in many other SSA countries. As cash incomes declined, the increase in out-of-pocket expenditures became a major obstacle to adequate access to health services. The abolition o f user fees for a number o f health services in 2006 improved rates o f utilization and should have significantly reduced household payments. Structure o f Public Spending - Functional and Economic Classification 3.5 1 Three ministries control most of health sector spending in Burundi. A wide variety o f ministries and institutions are involved in the management o f the health sector. However, most o f the public spending i s controlled by three ministries. The Ministry o f Public Health (MSP) i s responsible for 80-87 percent o f total public health expenditures. The second most important institution i s the Presidential Ministry in charge o f the Fight against HIV/AIDS. I t s share increased dramatically in 2006 and 2007, reaching 12-17 percent o f total budget allocations for the health sector. Finally, the Ministry o f Education provides subsidies and equipment to the University Hospital o f Kamenge. I t s share o f budget allocations declined from 10-14 percent in 2005 to 5-7 percent in 2006 and 2007. 3.52 Hospitals and the Fight against HIV/AIDS receive an excessive share of total Government and donor funding. During most o f the 2000s, autonomous hospitals received 36- 37 percent o f the current budget o f MSP, and the Kamenge hospital alone was allocated funds equivalent to 38-43 percent o f the hospital care current budget. T h i s i s not an optimal use o f scarce resources. Most o f the Burundian population lives in rural areas, and the main causes o f mortality and morbidity can be treated at lower levels o f care. In 2006, about 25 percent o f donor assistance to the health sector was allocated to HIV/AIDS, and most o f the NGOs also focus their activities on HIV/AIDS. The high priority given to H I V / A I D S i s questionable in a country with an adult prevalence rate which does not exceed 3 percent. 3.53 Budget allocations to the Health Ministry are dominated by salaries and other current expenditures. U n t i l 2005, current expenditures accounted for 9 1-96 percent o f budget allocations to MSP. About 35-46 percent o f current spending was for salaries; 27-46 percent for subsidies, which included a significant salary component; and 19-33 percent for goods and services. In 2006, access to HIPC funds enabled the ministry to increase the share o f capital expenditures (14 percent) and - within current expenditures - the share o f goods and services (36 percent). However, rates o f execution f e l l sharply for goods and services and for capital 65 expenditures. Health facilities charge for drugs and receive large donations from donors and NGOs for drugs, vaccines, and preservatives. These contributions are essential to improve the operational performance o f basic health services. Financial Management Issues 3.54 Improving the performance of the health sector will require a major effort on the part of the Government to increase budget allocations, improve rates of execution for capital expenditures and goods and services, and mobilize additional donor funding in line with appropriate strategic priorities for the sector. One o f the bottlenecks i s the lack o f financial management capacity within the Health Ministry. In 2006, MSP created the Direction Gknkrule des Ressources (DGR), which i s responsible for managing the ministry’s financial and human resources. However, budget planning and execution also depend on the performance o f Provincial Health Bureaus and provincial and local health facilities, which have limited capacity. 3.55 Budget planning i s poorly coordinated. Budget preparation begins when the Ministry receives the framework letter (Lettre de Cudrage) sent by the Finance Ministry to all the line ministries. The letter does not provide adequate guidance to spending ministries, as it does not include indicative budget ceilings for each sector and each ministry. All health departments, health programs, and Provincial Bureaus are expected to prepare their own budget proposals. These proposals are not based on realistic projections o f available resources, are seldom linked to clear sector strategic objectives, and do not prioritize objectives and expenditure programs. All the budget proposals are then reviewed by DGR, with the help o f the Ministry’s Planning Unit. The links with the sector’s strategic objectives are stronger since DGR and Planning have begun to prepare a health plan for 2007-2009. The central team has the skills needed for planning, but would need guidance and training on the costing o f health programs and projects. I n addition, the DGR i s overwhelmed by a wide variety o f administrative and budget execution tasks and cannot allocate sufficient time and resources to major strategy formulation and budget planning activities. 3.56 The fragmentation of the budget also i s a major problem. The planning o f HIPC- financed expenditures i s performed separately. The planning o f donor programs i s particularly difficult as central departments are not systematically informed o f projected and actual donor activities. As HIPC and donor funds are the two most important contributions to public health spending, i t i s virtually impossible for the Ministry to prepare realistic, comprehensive, and coherent budget proposals supporting adequate national priorities. The ongoing sector planning exercise, the forthcoming preparation o f a medium-term expenditure framework for the sector, and improved donor coordination in the context o f the preparation o f a possible SWAP should substantially improve the budget planning capacity o f the Ministry. 3.57 Budget execution and monitoring i s complicated by the lack o f capacity at the central and provincial levels. Until 2005, rates o f execution o f current expenditures remained high at 96-100 percent (although declining from 96 to 79 percent for goods and services). For domestically financed capital expenditures, rates o f execution declined from 102 percent in 2001 to 81-82 percent in 2004-2005. Poor and declining rates o f execution for goods and services and capital expenditures are largely due to cumbersome procurement procedures and the lack o f procurement capacity within MSP. 66 3.58 The poor disbursement performance o f HIPC funds, described above, led to a sharp decline in rates o f execution in 2006. In 2006, rates o f execution for all categories o f expenditures f e l l to 57 percent for current expenditures (including 36 percent for goods and services) and 4 percent for capital expenditures. T o improve i t s performance in the use o f HIPC funds, the Ministry now prepares a plan o f execution for HIPC-financed activities. However, it i s the totality o f the budget execution process (at the commitment stage), which needs to be better planned, monitored, simplified, and prioritized. 3.59 I n 2006, budget execution at the local level was further complicated by the decision to eliminate user fees for specific health services (women in delivery and children under 5). This i s a sound decision, which if well implemented will increase rates o f utilization and improve access to high-priority health services for the poorest segments o f the population. However, the decision was not based on a realistic assessment o f the full cost o f the measure, and in 2006, implementation involved multiple controls which overtaxed the capacity o f local facilities and were largely responsible for slow disbursement o f HIPC funds for health that year. Procedures were reviewed in 2007, with the Ministry o f Economy, Finance, and Development Cooperation releasing funds in advance and relying on the Health Ministry to control the use o f these resources. However, it i s s t i l l not clear whether budgeted funds will be sufficient to finance the full cost o f the measure. 3.60 Budget monitoringi s particularly weak. The Ministry does not have the staff necessary to supervise implementation o f large construction and rehabilitation projects, and does not monitor donor-financed programs. Primary health care facilities receive funds from a variety o f sources, including the payment o f salaries by the Ministdre de la Fonction Publique, the provision o f drugs and supplies by the Provincial Health Bureaus, and fees paid by patients. However, PHC facilities do not produce budgets and budget execution reports. T o assess the possibility o f significant leakage o f resources, the Government has launched a Public Expenditure Tracking survey for the sector. Intensive participation o f local communities in the management o f health facilities would also help improve the monitoring o f PHC management. So far, only one province has taken steps to involve Local Management Committees (Cornit& de Gestion) in the management o f resources available to PHC facilities. 67 Box 3.4: Main Recommandations in the Health Sector Major efforts should be made to improve health outcomes, which were severely affected by the political and economic crisis of the I990s, and have remained particularly dismal. Short term oriorities: Evaluate the budgetary implications o f the May 2006 decision abolishing user fees for certain medical services (children under 5 and women during delivery) in order to increase budget allocations, particularly those to the primary health care (PHC); 0 Improve the capacity o f health workers, and encourage the retention o f quali9ed medical personnel in rural areas; Establish a more meaningful and efective staffappraisal system (Le. a contract based system); 0 Reorganize the health sector into health districts; 0 Redeploy available human resources to provincial and local health facilities outside o f Bujumbura; and Improve the provision o f drugs and vaccines, and other critical goods and services, to district hospitals and PHC services. Medium to long term priorities: 0 Strengthen the financial management capacity o f the Ministry o f Health (at the central and provincial levels): One o f the main objectives of improving the financial management capacity o f MSP and extensive consultations with donors should be to protect high-priority programs, including drugs, vaccines and other medical supplies for district hospitals and PHC services.; 0 Intensib ongoing strategic planning activities in collaboration with its development partners. The process should be supported by the gradual introduction o f program budgeting (starting in 2009/2010), and by the development o f new donor coordination and financing instruments (S WAps) based on well-conceived sector strategies and program-budgets; and 0 Increase the role o f local communities in the management o f local health facilities. Public expenditure tracking surveys should assist the process and help monitor progress. 68 D. INFRASTRUCTURE Box 3.5: Key Sector Development Issues Burundi’s infrastructure has been virtually destroyed by seventeen years o f civil war. Neither the national road network, nor existing power and water supply facilities have been properly maintained during most o f that period. A major infrastructure rehabilitation program needs to be undertaken by the Government, with the help o f its developmentpartners. I n the case ofpower and water, the rehabilitation of existing infrastructure should be combined with substantial reforms in the management o f REGIDESO, the state enterprise responsible for producing and distributing power and water in urban centers. During the 2000s, the Government increased public spending on road maintenance, and the donor community invested heavily in the rehabilitation of the national road network. However, available evidence suggests that the present condition o f rural roads remains a major obstacle to the development of the agricultural sector. So far, very little has been done by the Government and by donors to rehabilitate the urban power and water systems managed by REGIDESO. A combination o f tarif adjustments, Government subsidies, and large donor programs w i l l be necessary to rehabilitate and extend the existing network and improve the financial viability o f the enterprise. The condition o f rural power and water systems managed by local governments needs to be reviewed, as a first step in the preparation o f a comprehensive rehabilitation program essential to improve access to basic water and power services for a growing share o f the rural population. 3.61 This report defines infrastructure sector as including road, water supply, and electricity supply subsectors. Table 3.14 presents the allocation o f public resources53among these subsectors and as compared to that in EAC comparators. Table 3.14: Public Spending on Roads, Power and Water Supply in Burundi and four other East African countries (YOGDP) Burundi Kenya Rwanda Uganda Tanzania 2001-04 2004-06 2001-04 2002-05 2003-04 2001-05 Total Spending on Roads 0.14 0.54 1.25 0.84 2.10 1.67 Investment 0.10 0.51 1.oo 0.73 1.66 0.94 O&M 0.04 0.03 0.26 0.10 0.45 0.72 Total Pub. Spending on 1.59 1.82 3.65 2.03 1.82 3.08 Power Investment 0.43 1.02 0.94 0.90 0.60 0.36 O&M 1.16 0.81 2.70 1.12 1.22 2.72 Total Pub. Spending on 0.70 0.77 0.61 1.08 1-05 0.48 Water & Sanitation Investment 0.20 0.33 0.33 0.74 0.61 0.3 1 O&M 0.50 0.44 0.28 0.33 0.44 0.17 Total Infrastructure 2.43 3.13 5.51 3.92 4.98 5.22 Investment 0.73 1.85 2.27 2.37 2.87 1.61 O&M 1.70 1.28 3.24 1.55 2.1 1 3.61 Source: World Bank estimates based on budget laws 2001-2006, and off-budget sources (FNR, and REGIDESO 2006); and Bricano-Gannandia (2007) for Kenya, Rwanda, Uganda and Tanzania. 53 Estimates of public spending in infrastructure displayed in Table 3.14 differ from those in Tables 2.4-2.6 because the former includes an estimation o f off-budget spending (REGIDESO and FNR) while the latter i s simply based on budget laws. 69 The Road Sector - Recent Performance and Public Spending 3.62 With 11,300 kilometers (kms) Table 3.15: Density of the Road Network in Burundi of roads, Burundi has one of the and other East Africa Countries largest road networks of the East Burundi Kenva Rwanda Tanzania Ueanda Africa Region, if one takes into Km o f account the size o f the country and roads(per 565.0 112.0 486.0 100.0 137.0 1000 of its population (Table 3.15). Local km2) governments are responsible for Km o f maintaining about 6,100 k m s o f rural roads(per 294.0 60.0 366.0 43.0 146.0 roads; the central Government i s 1000 o f inhabitants) responsible for the construction and maintenance o f about 5,200 k m s o f Source: World Bank estimates, based on Bricano-Gannedia and Foster (2007). national roads; only 1,200 k m s o f national roads are paved. 3.63 The road network and the road transport system of the country have been severely affected by the political and economic crisis of the 1990s. N o t only have the Government and the donor community stopped investing in the sector, but funds allocated to maintenance declined and became grossly inadequate. In addition, the crisis created insecurity; roadblocks and other illegal controls had a strong negative impact on transport operations and transport costs. 3.64 As a landlocked country, Burundi depends on the performance of the road, rail, and port systems of its neighbors. About two thirds o f Burundi’s international traffic uses the Tanzanian railroad and the port o f Dar es Salam. In 1996, most o f the other East African countries imposed an embargo on economic relations with Burundi. The measure had a profound impact on the country’s international transport system and external trade. 3.65 Significant efforts have been made by the Government over the past few years to rehabilitate the road network and improve maintenance. The Government created the Ofice des Routes (OdR), which i s responsible for rehabilitating and maintaining national roads, and the Fonds National Routier (FNR), an autonomous Road Fund, financed by the product o f fuel taxes and drivers’ licenses. As shown in Table 3.14 above, total public spending on roads increased significantly on average from 2001-04 to 2005-06. OdR, which has prepared a detailed program o f maintenance for national and urban roads, considers that the funds mobilized through FNR are now sufficient to finance more than two thirds o f the maintenance requirements o f the national road network. Table 3.14, however, also shows that despite the recent increase, Burundi’s Government s t i l l spends much less on the road sector than the four other low-income East African countries selected as comparators. 70 3.66 Recognizing the importance of Table 3.16: Donor Contributions to the Road Sector in an adequate road network for the Burundi and other Low-Income Countries recovery o f the Burundian economy, (in share of GDP) Average Annual Donor Contribution the donor community invested Low-Income Burundi SSA average massively in the rehabilitation of the countries road sector. As shown in Table 3.16, 1995-99 0.00 0.74 0.37 donor funding o f the road sector went up 2000-04 3.03 0.64 0.33 Source: World Bank estimates, based on OECD-DAC data from virtuallynothingi n the 1ggos to 3 percent o f GDP 2000-2004, close to 5 times the average in other low-income countries (LIC) and more than 9 times the average in sub-Saharan Africa. 3.67 This effort needs to be continued over the next few years to complete the ongoing national road rehabilitation process and extend it to other (non-classified) roads. A major problem i s the rehabilitation and maintenance o f rural roads, which i s essential to support rural development and stimulate the growth o f a market oriented agricultural sector. All Government spending on roads and virtually all the donor contributions finance national roads. Data collected by OdR do not cover rural roads, but farmers, merchants, and transporters complain that the present condition o f the rural road network i s a major obstacle to economic development. Power and Water - Recent Performance and Public Spending 3.68 With its abundant water resources, Burundi has the capacity to provide access to safe water to a growing percentage of its population and to increase substantially its power production through the development of small and medium-sized hydropower schemes. REGIDESO, a public enterprise, i s responsible for producing and distributing water and power to urban centers. The Direction Gknkrake de 1'Hydraulique et des Energies Rurales assumes the same functions for the rest o f the country. Sector regulation i s in the hands o f the Direction Gknkrale de 1 'Eau et des Mines, within the Ministry o f Energy and Mining. 3.69 REGIDESO used to be a well-managed institution, but its performance was severely affected by the crisis of the 1990s. Lack o f investment and inadequate maintenance led to a rapid deterioration o f an aging network. Technical losses and power outages increased. In theory, the power generation capacity o f REGIDESO includes 32 megawatts (MW) from small hydropower plants, 5.5 MW from a thermal unit, and about one third o f the production o f SINELAC, a regional enterprise owned and operated jointly by Rwanda, Burundi and the DRC. In practice, however, the company i s unable to meet more than 20 MW,, less than half o f the peak demand, estimated at 42 MW. T h i s i s due to two main factors: (a) the poor condition o f most o f REGIDESO's production and distribution facilities; and (b) the financial position o f the enterprise, whose earnings are not sufficient to pay for the purchase o f fuel for the thermal unit. REGIDESO's power and water rates are comparatively l o w and have not been adequately adjusted over the past few years. REGIDESO's power production declined from 167 Gigawatt Hour (GWh) in 2002 to 150 in 2006. 3.70 As shown in Table 3.17, access to water (45 percent of households) and electricity (2 percent) i s much lower in Burundi than in comparator countries. Access to water varies between 40 percent in rural areas and 75 percent in urban areas. Access to power does not exceed 71 2 percent; 62 percent o f the connections are in Bujumbura; most o f the energy used by Burundian households comes from wood and wood products. 3.71 So far, the Government has given a low priority to the rehabilitation of the power and water sectors. Table 3.14 above shows that domestically financed public expenditures (investments financed by the Government budget, and operation and maintenance (O&M) expenditures financed by REGIDESO) increased modestly during the 2000s but remain much lower than that for Burundi’s other East African comparators. 3.72 Inadequate Table 3.17: Power and Water Supply in Burundi and Four Other Government spending East African Countries was not offset by massive SSA inflows of donor Burundi Kenya Rwanda Tanzania Uganda Average assistance. As shown in Power -Households 2.0 16.0 6.2 11.4 8.6 27.2 Table 3.18 below, donor with access to contributions to power and electricity (%) water increased -Technical and 26.8 20.1 25.2 25.0 33.0 n.a. non-technical substantially in the 2000s, losses (“XI) but not enough to offset the -Number of 141.3 85.7 n.a. n.a. n.a. 56.4 lack o f adequate funding power outages (days) f r o m domestic sources. In Water 2005, the Government -Households 45.0 62.0 73.0 73.0 56.0 64.1 decided to undertake a with access to improved review o f REGIDESO’s water (%) tariffs; the review was not -Technical and 41.4 50.0 37.4 61.5 44.3 n.a. completed until M a y 2007 ~o’”~se~~~cal and its recommendations Source: World Bank estimates, based on Bricano-Garmedia and Foster (2007). have not been adopted by the Government or REGIDESO. Very little i s known about the condition o f water and power services in rural areas. Financial Management Issues 3.73 Budget planning i s Table 3.18: Donor Contributions to Power, Water Supply and performed by the Budget Sanitation (in shares of GDP) Directorate in the Ministry Average Annual Donor Contribution of Economy, Finance, and Sub- Low- Development Cooperation, Sahara Period Burundi Income in cooperation with the Vice Africa countries @SA) Ministry of Planning and the 1995-99 0.01 0.59 0.15 line ministries (Public Energy 2002-04 0 0.3 0.15 Works, and Mining and Water Supply and 1995-99 0.05 0.29 0.21 Energy). A major effort i s Sanitation 2000-04 0.24 0.23 0.2 required to evaluate the Source: World Bank estimates based on OECD-DAC data. situation o f the infrastructure sectors, notably rural roads and rural power and water; to develop sound sector strategies; and to identify budget priorities. 72 3.74 Rates o f budget execution have been generally high but declined seriously in 2006. Rates o f budget execution varied between 79 and 101 percent but fell t o 48 percent in 2006, mainly because o f unrealistic projections o f investment expenditures (rate o f execution o f the investment budget was only 35 percent). Box 3.6: Main Recommendationson the Infrastructuresector High priority should be given to the comprehensive rehabilitation o f the national road network, and to the overhaul of the urban power and water systems managed by the state-owned utility, REGIDESO. The creation of master plans in the infrastructure sector and the use of public-private partnerships to finance the sector should also be encouraged. Short term priorities: Roads Allocate adequate resources to road maintenance on the basis of the plan prepared by the Ofiice des Routes (OdR); e Assess the condition of rural roads and how their condition impacts the agricultural sector; Allocate appropriate resources to local governments for the maintenance of a priority network of regional and rural roads; Power and water Undertake the assessment o f the condition of rural power and water facilities, as a first step toward the development of a sound strategy for the sector; and Take measures to improve the financial viability and operational performance o f REGIDESO, through a combination of tariff adjustments and government subsidies. Medium to long term oriorities: Roads Explore, with other members of the East Afiican Community, ways to improve trafiic on regional and international routes linking Burundi with major seaports in Tanzania and Kenya. Power and water Create enabling environment for the private sector by developing the urban power sector, including through privatization of the state-owned utility; and Explore opportunities for regional cooperation to find alternative sources of energy (e.g., the Lake Kivu methane project in Rwanda). 3.75 The next chapter discusses the implementation o f the country’s public expenditure management (PEM) reform, focusing o n accountability, transparency, and effectiveness in the use o f domestic and foreign public resources. 73 CHAPTER 4. REVIEW OF PROGRESS IN PUBLIC EXPENDITURE AND BUDGET REFORMS 4.1 As noted in the preceding chapters, in order t o consolidate recent progress in reconstruction and effectively mobilize increased donor support, accelerate growth and reduce poverty, the Government critically needs to improve i t s public finance management. This must be accomplished through the orderly implementation o f its public expenditure management (PEM) reform, focusing o n accountability, transparency, and effectiveness in the use o f domestic and foreign public resources. 4.2 The first section o f this chapter gives an overview o f the PEM reform program. The second section reviews the budget preparation process; the third deals with budget execution; the fourth reviews audit and oversight; the fifth section reviews procurement procedures; and the sixth section discusses the conditions under which the introduction o f a medium-term expenditure framework (MTEF) might be possible in Burundi. A. OVERVIEW OF PUBLIC EXPENDITURE MANAGEMENT REFORMS 4.3 This review i s based mainly on existing studies o n PEM in Burundi carried out by several donors, particularly the W o r l d Bank (Country Financial Accountability Assessment in 2004; studies financed by PAGE); the IMF (Fiscal Affairs Department, A F R I T A C Central); the African Development Bank; the European Union; the U S Agency for International Development; and France. The mission and studies by these donors have already identified the main problems and made many recommendations for PEM reforms. With the assistance o f these donors, the Government has prepared a very ambitious PEM reform program, but this program reads more like a l i s t o f donor-driven recommendations than a realistic blueprint for reform. Very few measuresh-ecommendationsfrom that l i s t have been implemented to date because o f the lack o f full ownership by the Burundians, weak technical capacity, and delays and lack o f donor coordination in the provision o f technical assistance (TA). The implementation approach i s piecemeal because the recommendations for the PEM reforms are not prioritized and set in a strategic framework. This Overview section summarizes recent reviews on PEM progress.54 The Partnership Framework 4.4 A partnership agreement (Cadre de partenariat) was signed in M a y 2005 between the Government and i t s development partners.55 Within the framework o f the partnership agreement: 54 See (a) IMF “Mesures prioritaires pour 1 ‘ame‘lioration de la gestion budge‘taire.” of October 2007 by Jean-Luc HClis et al; (b) IMF “La Re‘forme de la gestion budge‘taire” o f May 2007 by Jean-Luc HBlis et al; (c) IMF “Amklioration de la gestion de la tre‘sorerie et strate‘gie de re‘forme dans le domaine desjnances publiques” of April 2007 by Jean-Pierre L e Bouder; and (d) various reports financed through the IDA’SPAGE. ’’ The original signatories of the partnership agreement include Belgium, France, African DevelopmentBank, the UNDP, the World Bank, and the European Commission. The IMF i s associated as a full member. 74 (a) donors disburse their budgetary assistance on a yearly basis, in one or several tranches, based on the progress made by the Government on maintaining a sound macroeconomic situation, the fight against poverty (particularly in social sectors), and improving public finance management; (b) the Government produces monthly reports o n budget execution, with an updated Table o f Government’s Financial Operations (TOFE) showing possible residual financing gaps; and (c) the Government reviews its public finance reform program and i t s report i s to be included in the general report o f the supreme audit institution, Cour des Comptes (CdC). T o achieve these objectives, the partnership framework i s assisted by a support unit, the Cellule d ’ a p p ~ ia ,~~ Technical Committee, and a Coordination Committee o f the partners. The Cellule d ’appui supported by the EU i s located in the Ministry o f Economy, Finance, and Development Cooperation (MEFDC) and i s in charge o f developing, monitoring, and updating the P E M reform program. The Technical Committee5’ supervises the coordination o f the different projects and TA involved in the P E M reform program. The Coordination Committee, composed o f the mission chiefs o f the donors concerned, organizes debriefing meetings with donors’ review missions and other important economic or public finance missions. I t s meetings are chaired by the Minister o f Economy, Finance, and Development Cooperation (Minister o f Finance). The Coordination Committee also meets as needed to review progress in implementation o f the P E M reform program; and to review projections and questions concerning budget support. Issues and Challenges 4.5 Because o f its late start, the effectiveness o f the partnership framework has been limited. Several commitments and actions agreed upon by the Government and the donors in the framework document have not been implemented. The documents needed by the partnership framework for its work have either not been prepared, or updated, or transmitted to the donors. These are the annual review o f PRSP and the update o f the P E M reform program.58 Support from the Government and donors to the Cellule d’appui has been insufficient or limited: the office equipment, allowances for the staff, and operating costs to be financed by donors have not been received by the Cellule d’appui until very recently. Also, the MEFDC did not make available enough office space for the Cellule d ’appui.The Cellule d ’appui i s weak partly because its status, terms o f reference, and work program are not clearly defined. The long-term TA experts to the Directorates o f Budget and Accounting have not been recruited by the partners. 56 The Cellule d ’appui started in November 2006. ” The Technical Committee i s composed of donors’ representatives, managers of projects involved in PEM reform (e.g., PAGE), and members of the Cellule d’appui. 58 See next sub-section. 75 Recommendations 4.6 For the Partnership Framework Agreement to effectively help the Government in the preparation and orderly implementation o f the P E M reform program, the following actions are needed: 0 The Minister o f Finance, in cooperation with partners, needs to more clearly define the status, mission TOR, and work program o f the Cellule d ’appui. I t s operating costs should be agreed upon and made available in a timely manner. The partners should meet the specific TA needs o f the Cellule for the development and management o f PEM reforms; and for the preparation, monitoring, and updating o f the strategy and related action plan in the short and medium term in the area o f public finance. Moreover, the adoption o f the P E M reform program and i t s rationale at the level o f the Presidency and the Council o f Ministers, and the communication o f this to other ministries, should be envisaged. 0 Donors and the Government should fully implement their commitments toward the objectives, the TA, and the operating costs described in the framework document and its annexes. I n particular, the implementation by donors o f their commitment to align the formulation and execution modalities o f their budget support to the Government’s budget cycle i s essential for the preparation and execution o f a realistic budget and to facilitate smooth cash flow management. Need for a Strategy, Related Action Plan, and Coordination of the P E M Reforms and TA 4.7 Under the Partnership Framework, the MEFDC was to update, by July 15, 2005, the initial program for the modernization o f public financial management sent to the development partners on January 8, 2002. The update was to take into account existing or expected supports for the various projects in support o f PEM, and recent assessments o f the P E M by FAD and World Bank missions. The Cellule d’appui has yet to finalize a comprehensive P E M strategy and related action plan. I t prepared a road map (feuille de route) and an action plan presented to the Technical Committee o f the Partnership Framework in December 2006. Taking into account the comments received and a report by an AFRITAC mission, the Cellule d’appui prepared a more comprehensive draft strategy for P E M reforms and a related action plan and sent it to donors for comments. Comments were received only fkom FAD and World Bank. The current PEMFAR, prepared jointly with the authorities, aims to serve as the key input o f the revised strategy for P E M reforms o f the Ministry o f Economy, Finance, and Development Cooperation. 4.8 Some actions, undertaken in cooperation with the development partners, have translated into tangible results. In particular, most o f the reforms in public expenditure management have been financed and implemented through the IDA-financed PAGE project. The actions undertaken under this project have led to the (a) development o f the Budget and Accounting Plan o f the State (PNCE); (b) computerization o f the expenditure chain; (c) revision o f the Budget Laws and Regulations to incorporate the wage bill and expenditures financed by the HIPC resources into the accounting system; and (d) strengthening o f the Supreme Audit Institution (Cour des comptes). Similarly, an IMF-financed long-term TA expert in the Accounting Directorate i s helping to put the budget accounting in order. 76 4.9 These remarkable achievements are worth stressing. However, their impact will be short- lived without addressing the following weaknesses o f the current approach to P E M reforms: 0 The implementation and sustainability o f the reforms are limited by the lack o f strong ownership by the authorities, as illustrated by the recurrent mention o f the external origin o f the recommendations; the conclusions o f donors’ missions are often reached without the full involvement o f the Government’s services concerned; 0 As a result, there i s no real coordination or consistency between the actions proposed; their operationality i s limited by the absence o f a prioritized, systematic, and sequenced multi-year programming o f the actions o f the P E M reform program; The action plan does not take full account o f the limited technical capacity o f the MEFDC and the other entities involved, nor does it outline the concrete steps that would allow the implementation o f each action; 0 The lack o f internal and external communication programs to explain the rationale o f the reforms i s a serious barrier to the needed mobilization o f the staff concerned, the Government, and public opinion in support o f the orderly implementation o f reforms. 4.10 The Cellule d’appui has initiated a dialogue with the donors and a draft strategy and action plan o n the P E M reform program. I t needs, with the involvement o f the Minister o f Finance, to intensify the dialogue with the development partners, with a view to finalizing the P E M reform strategy and action plan by end June 2008, if needed, through the organization o f a mini-seminar in the end o f second quarter o f 2008 with the donors in the Partnership Framework. T o that end, all donors should submit their comments on the draft strategy and action plan to Cellule d ’appui in a timely manner. Recommendations 4.11 T o be more operational, the strategy and priority action plan have to be prepared by the Cellule d ’appui with the assistance o f all the development partners Concerned, with the Cellule d’appui and the Minister o f Finance taking full ownership. They should focus on priority objectives and identify the results expected in the medium term. The objectives should be detailed in concrete and realistic actions, taking into account the constraints and absorptive capacity o f the Government. For each action, the action plan will mention: the main entity in charge, the associated entities, the costs, the implementation schedule, the measurable impact, the financing (domestic and/or foreign), and the TA needed. If necessary, some actions could be subdivided into sub-actions representing steps or preconditions for the implementations o f the next action. The orderly and successful implementation o f the PEM reform program strategy and priority action plan will critically depend on two conditions: (a) the hll commitment and involvement o f the authorities, the Ministry o f Economy, Finance, and Development Cooperation in particular, in the formulatiodpreparation, management, and monitoring o f the priority action plan; and (b) the provision o f a sizeable TA program during a period o f a few years. The TA expert selected should have real situation experience, and be capable o f accompanying the Government in the reform process and transferring know-how. 77 4.12 The main recommendations are as follows: 0 Update and finalize the P E M reform strategy and related priority action plan for the short and medium terms, detailing short-term actions (2008 and 2009). 0 The Cellule d’appui, with the assistance o f the donors concerned, should carry out a critical review o f the priority action plan it has proposed to select a more limited core of priority actions. 0 The roles o f the partners’ Technical Committee and Coordination Committee must be strengthened to achieve more effective coordination among donors; provide timely support to the structures in charge o f the P E M reform program; and ensure adequate monitoring o f the program’s implementation. 4.13 In line with the above, the following six sections should be seen as contributions to the identification o f strategic priorities for the P E M reform agenda. However, these priorities will need to be subjected to the authorities’ review, and will not become operational unless the Government takes full ownership o f them. B. REFORM OF THE BUDGET FRAMEWORK Legal Basis o f the Budget 4.14 The Constitution o f October 28, 2001 sets the legal and institutional framework for the budget. I t defines the global framework for the state’s financial and administrative management, including management o f the State’s budget, and defines the basis and the rates o f taxes and duties.sg The Constitution defines the Budget L a w and the modalities for i t s adoption, which gives preponderance to the National Assembly over the Senate for the review and oversight. The main texts organizing the budget legal framework are in B o x 4.1. 4.15 Article 129 provides that the budget sets, for each year, the resources and expenditures o f the State. The annual process begins with the transmission o f the draft initial budget to the National Assembly at the beginning o f i t s October session (Article 139). The National Assembly votes to approve the budget. If the National Assembly makes no decision by December 3 1, the past year’s budget i s implemented by provisional twelfths. In that case, the Parliament (the National Assembly and the Senate) meets at the request o f the President, for a maximum o f 15 days, to consider the draft budget. I f the Parliament fails to adopt the budget within this timeframe, the budget i s adopted by decreeflaw by the Council o f Ministers (Article 140). 4.16 Article 130 o f the Constitution provides for the creation o f a Supreme Audit Institution (Cour des Cornptes) in charge o f reviewing and certifying the accounts o f all public entities. T h i s Court has the obligation to present to the National Assembly a report on the regularity o f the State’s general account, confirming whether the funds have been used in compliance with the 59 The 2001 Constitution is simply the latest manifestation o f the legal framework for PFM. The essential parts o f this framework go back to independence in 1962. 78 budget and established procedures. The text does not mention the periodicity o f such reports, but one can expect i t to be on a yearly basis. 4.17 The modified L a w o f March 19, 1964 s t i l l remains the backbone o f the Budget L a w in Burundi, although i t i s outdated in many respects, particularly (a) the split between the operating budget (ordinary budget, BO) and the investment budget (extraordinary investment budget, BEI); (b) the numerous exceptions which limit the scope o f the law, such as the multiplication o f Government accounts; and (c) the commitment o f resources to specific expenses. There i s n o organic law that clearly defines the objectives and the principles o f the budget laws. Furthermore, the decrees o f 1988 and 1989 providing the legal framework for “personalized administrations”60 and public entities do not allow a rational classification o f public entities. 4.18 Aware o f these limitations, the authorities have decided to revise the legal framework for the preparation and the execution o f the budget, with the assistance o f private consultants recruited by IDA-financed PAGE. The MEFDC i s fully involved in the revision process. A final draft o f the Budget Framework L a w (Organic Law) currently available has been adopted in March 2008 by the Council o f Ministers and will be submitted very soon to the Parliament. The broad principles o f the legal framework are as follows: 0 At the apex, the Constitution would set the broad principles o f public finance management. 0 Just below, the Budget Framework Law ( h i organique) would set the fundamental principles and rules for public finance. Complementary laws would set the rules for the management o f public finance in application o f the Framework Law. The most important o f those laws would be the L a w on Public Accounting, which would set the essential rules and procedures for the execution o f the budget. The main objectives o f that law, which would replace the public accounting law o f 1964, would be to: (a) clarify the roles and responsibilities o f the executive and the legislature in budget formulation and approval; (b) reaffirm internationally accepted budget management principles6’; (c) streamline the expenditure chain by clarifying and strengthening the internal control and internal audit framework; (d) clarify and strengthen external oversight; and (e) pave the way for future reforms, once the current system performs satisfactorily; i.e., moves toward performance budgeting and the devolution o f commitment authority to sectoral ministries. 0 Finally, various decrees would define the procedures and rules for application o f the Budget Framework L a w in more detail. These decrees would include, in particular: (a) a decree on the presentatiodscope and the preparation o f the Budget Law; and (b) the existing decree on the budgetary and accounting plan o f the State. 6o T h e law on the Cour des Comptes calls these entities rigiespersonalisies. 61 These are the principles of authoritativeness, annual basis, universality, unity, specificity, balance, accountability, transparency, stability or predictability and performance (see OECD Journal on budgeting, “The legal framework for budget systems, an international comparison”, Vol4, No 3,2004). 79 Box 4.1: Burundi - Main Texts Organizing the Budget Legal Framework Transition Constitution o f October 28, 2001 defining the respective prerogatives o f the Parliament and the Government concerning the budget. The L a w o f March 19, 1964, setting the rules o f public accounting for the State. This fundamental law sets the rules and the organization o f the preparation and execution o f the revenues and expenditures o f the State’s budget, and defines the prerogatives o f the different participants in the process. I t sets the principles o f a single Treasury account (STA), universality, yearly period, and non-reservation o f resources. The budget execution management system i s one without complementary period, starting January 1 and ending December 3 1. The Ministerial Ordinance No. 030/89 o f June 23, 1969 defining the rules for executing the Law o n the Public Accounting o f the State. I t sets the detailed rules for the execution o f the budget for (a) the public accountant, for inflows and outflows o f funds; (b) for credit managers; (c) for the Treasurer; and (d) for the Cashier o f Burundi. The Decree No. 1 4 7 1 o f December 10, 1971 introducing the budgetary control. The Decree No. 1/039 o f December 30, 1989 widening the scope o f the general budget to also encompass operations hors budget, annex budgets, autonomous budgets, and special accounts. The Decree No. 100/168 o f December 31,2004 approving the new State’s budget and accounting plan. Issues and Challenges o f the Proposed Reforms 4.19 Budget procedures traditionally pursue three objectives: (a) exerting control over public expenditure to ensure compliance with parliamentary authorization; (b) stabilizing the economy by adapting fiscal outturns to macro-economic circumstances; and (c) ensuring allocative efficiency in the delivery o f public services. The move toward a performance-based budget and the devolution o f commitment authority to line ministries i s inspired by reforms that have been taking place in OECD countries since the 1990s, in a movement referred to as the “new public management.” The rationale behind these reforms i s to increase the focus o f public expenditure management o n the third objective, the allocative efficiency o f public expenditure, in a context o f aging populations and diminishing tolerance for taxation. M o s t developing countries are characterized by significantly different conditions: with a young and growing population, governments need to increase tax revenues to finance essential services in order to meet the MDGs without excessive reliance on foreign aid. This does not mean that reforms inspired by the new public management movement are never appropriate in the context o f developing economies - there are circumstances where it i s also appropriate for these economies to focus on allocative efficiency. However, if new public management i s to be introduced in developing economies, it should be done for the right reasons, once a minimum o f preconditions are met and taking due account o f the gap between the demanding nature o f the change management process involved by these reforms and the limited level o f local capacity.62 For a similar argument, regarding emerging economies, refer to “From program to performance budgeting: the challenge for emerging economies,” Ministry of Finance Working Paper WP/03/169, Jack Diamond, June 2003. 80 4.20 The PEM reforms that took place in donor countries have had a strong influence o n the Burundi PEM agenda, both directly, through donor conditionality and technical assistance, and indirectly, because the authorities naturally considered the benefits o f implementing what was consistently presented as international best practice. This situation i s not unusual in developing countries. Yet i t would make little sense for Burundi t o focus primarily on the allocative efficiency o f i t s expenditure when i t s PEM system does not yet ensure either full compliance o f expenditure with parliamentary authorization or an adequate level o f control over the macro- economic impact o f i t s fiscal operations. 4.21 The preconditions that would need to be met before Burundi could hope to derive some benefits fi-om a move to performance budgeting would include at least the f ~ l l o w i n g : ~ ~ Implementation o f sound budgetary operating procedures, as provided for in the draft organic law and evidenced by the assessments provided by internal audit and inspection bodies (including Inspection Gdndrale de 1’Etat, IGE); A significantly higher degree o f consolidation o f public expenditure into the budget than i s currently the case (i.e., a marked increase in on-budget expenditure); A reasonable degree o f financial discipline, as evidenced by the credibility o f the budget (comparison o f budget estimates and outturns, both in aggregate and by sector); Reasonably accurate recording and reporting o f financial and physical data, as evidenced by the regular production o f consistent Government accounts and the opinions o f external auditors. 4.22 Until these conditions are met, introducing performance budgeting and devolving commitment authorizations would risk further undermining the PEM system rather than strengthening it. As Section I,on introducing the MTEF to Burundi, will make clear, this does not mean that the program-based approaches that were started at the sectoral level (in health, education, and the justice system) cannot produce useful results, as they should help improve the performance o f public service delivery in key sectors. Other tools exist to reinforce the performance outlook o f public expenditure in Burundi; for instance, an improvement o f the PIP process could provide a useful contribution. However, budget procedures should not be changed until the PEM system i s mature enough for this change not to affect fiscal discipline. 4.23 Once the preconditions outlined above have been met, if the Burundi authorities choose to focus o n improving the allocative efficiency o f public services by introducing performance budgeting, they may want to bear in mind the following considerations: Defining a meaningful program budget structure (i.e., one that provides accountability by linking the activities o f a given Government entity or entities with identifiable outputs 63 Similar preconditions were already identified in 1965 by the UN Manual for Program and Performance Budgeting, - based on the U S Planning, Programmingand Budgeting System (PPBS) experiment o f 1961 1971. Most o f the developing countries that adopted this approach had not met these preconditions and did not derive any benefits from it (see A. Waterson, “Development planning - Lessons from experience,” John Hopkins University Press, 1965). 81 and outcomes) i s an indispensable step - but it i s only the first step in a successfil move to performance budgeting; 0 As noted in Section I ,there i s a clear link between performance budgeting and the development o f a medium-term expenditure framework: as Government’s activities are n o longer defined in terms o f inputs but in terms o f outcomes, they need to be linked to the strategic objectives o f the Government derived from the PRSP. These cannot be merely annual, as most objectives take more than a year to achieve; 0 In order for performance budgeting to work, the authorities will need t o determine ways o f costing the activities identified in each program, using cost accounting methods. This will require donor cooperation in a context where most o f the investment expenditure i s financed by foreign aid flows and does not pass through the budget; 0 The full benefits o f performance budgeting only accrue when program managers are afforded extra budgetary flexibility and greater continuity in budget funding, and are subject to increased pressure to perform. I n the Burundi context, this may involve strengthening the capacity and responsibility o f the program manager (ordonnateur), reviewing the role and accountabilities o f the Contrde des de‘penses engage‘es and public accountants, and engaging donors to ensure a better predictability o f foreign aid flows; 0 Budget preparations cannot be done on the basis o f prior-year allocations and new spending measures approved by Parliament; but must be based o n realistic revenue projections by the DBC; Without an adequate management information system, including budget execution and outputs o f Government’s activity, the move to performance budgeting cannot really take place. 4.24 International experience on the challenges o f the move to performance budgeting indicates that i t should be treated as a major change management project, and that mere amendments o f the legal framework would be woefully insufficient t o implement this reform. In particular, lessons learned include the need for the following steps64: 0 Recognize a sense o f urgency to make a change: in the Burundi context, this has particular significance as the authorities currently need to deal with other urgent issues, including consolidation o f the peace agreement, the maintenance o f physical security, and the payment o f c i v i l servants; 0 Develop a powerful enough coalition to make the change: this would need to include constituents within the MEFDC, line ministries including the armed forces, Parliament and civil society, and the various ethnic groups recognized in the Constitution; 64Adapted from Kotter, J.P., “Why transformation efforts fail,” Harvard Business Review, Vol. 61 (March-April 1999, by Jack Diamond in “Performance Budgeting: managing the reform process,” Ministry of Finance Working Paper WPl03133, February 2003. 82 0 Identify an appropriate champion o f change in the Burundi context, who creates and communicates a common vision; empower, and give resources to an appropriate implementingagency; 0 Plan for early successes (easy wins) and build on them to sustain momentum for reform; 0 Institutionalize the changes in the law, accompanying it with appropriate levels o f training and targeted dissemination strategies. 4.25 As Schiavo-Campo and Tommasi (1999) put it, “introducing sophisticated and demanding performance budgeting systems (and particularly their culmination in detailed output budgeting) has been shown by the experience o f the last 50 years to be badly counterproductive when local capacity and ‘ownership’ are not conducive to their introduction. I n developing countries, this oversight has led to wasted resources and, in some cases, loss o f fiscal discipline. The reform lessons are that: (a) line-item budgeting and expenditure control must be firmly established before moving ahead with performance budgeting; and (b) complex reforms in a limited-capacity environment succeed only in disempowering the limited local capacity i t s e l f and reinforcing dependence on external ad~isory.”~’ M a i n recommendations for the budget legal framework are provided in B o x 4.2. Box 4.2: Budget Legal Framework-Main Recommendations I n the short term: 0 Use the current PEMsystem as a basis for clarification and strengthening; prepare and submit to the Parliament a budgetfiamework law setting sound basic principles ofpublic fnance management. Adopt separately a decree establishing the general rules of the State’s public accounting. I n the medium term: 0 Determine whether the authorities want to implement performance budgeting. 0 Identijj the prerequisites for the implementation ofpe$ormance budgeting, including ownership and local capacity. Identijj budget programs according to the strategic fiamework to fight poverty; and well-defined medium-term sectoral expenditure frameworks. Identijj the actions needed to modijj the accountability framework ofprogram managers and MEFDC representatives. I n the lone term, if the authorities decide to PO ahead with the reform: Test program budgeting and the devolution o f budget commitment in two or three priority line ministries. Use the above to revise the budgetframework law and the law on rules ofpublic accounting. Schiavo-Campo, S. and D. Tommasi. (Ed.), (1999) “Governance, Corruption and Public Financial Management,” Asian Development Bank: Manilla. 83 C. BUDGET COVERAGE AND PREPARATION Current Situation 4.26 The Ministry o f Economy, Finance, and Development Cooperation i s in charge o f the preparation and execution o f the budget. It i s composed o f three general directorates: the General Directorate o f Budget and Public Accounting (Direction gkndrale du budget et de la comptabilitk publique, DGBCP); the General Directorate o f Revenues (Direction gkndrale des recettes, DGR); and the General Directorate o f Public Procurement (Direction gdndrale des marchks publics, DGMP). Within the DGBCP, the Directorate o f Budget and Control (DBC) prepares the State’s general budget. The D B C has the following responsibilities: reception, analysis and preparation o f budget choices; preparation o f the budget, and the finance law; preparation o f delegations o f credit to be notified to credit managers o f line ministries; monitoring o f the State’s revenue collection; providing clearance for, and keep the books for expenditures at the stages o f commitment and liquidation; and monitoring o f expenditures at the stages o f payment order and payment. For the preparation o f the budget, D B C has two divisions: one for the ordinary budget (BO) (15 staff, specialized by sector), and one for the Investment Budget (BEI) (five staff, including two from the General Directorate o f Planning (DGP). 4.27 The line ministries do not have real structures for the preparation o f the budget; and basically rely on the financial controller (Comptable des dkpenses engagkes) seconded from the MEFDC. The Vice-Ministry o f Planning (VMOP) i s in charge o f preparing the macroeconomic framework document, and the Priority Public Investment Program (PIP). The Ministry o f Public Public Service, Labor and Social Security (MinistBre de la Fonction Publique, MFP) i s in charge o f managing civil servants, and the verificationshalidation o f their salaries (Office o f wages). 4.28 Some progress was made in 2006 for the preparation o f the budget. The VMOP prepared, with the assistance o f financial partners (World Bank), a 2006-08 macroeconomic framework. The 2007 finance law abolished several special funds except the National Road Fund (FNR). Some external financings for investment operations were recorded in an annex to the finance law and presented as financing source in the general balance table. 4.29 Important efforts were made so that the new budget nomenclatures o f the State’s Budget and Accounting Plan are classified in line with international norms. The consistent presentation o f the classification o f expenditure according to its nature (administrative, functional, by economic nature) and by source o f financing, should help improve the transparency o f the budget. In addition, a specific codification has been added to the PBCE for HIPC-financed expenditures and expenditures for poverty reduction to allow their identification, and monitoring through the information technology (IT) system. 84 Main Issues and Challenges Budget Coverage 4.30 The budget coverage i s not extensive. The 2007 Budget L a w covers the BO, the BEI, and the national road fund (FNR). But it does not cover all the foreign-financed investment operations.66 Moreover, there are several inconsistencies between the multi-year Priority Public Investment Program (PIP) and the BEI. For example, the 2007 budget includes appropriations for projects not listed in the PIP, which indicates that they are not priority projects. Furthermore, the BO and BE1are not consolidated at the level o f line ministries; and a very large proportion (more than 75 percent) o f investment expenditures are recorded as current expenditures. 4.31 Based on the I-PRSP o f 2003, a preliminary study was undertaken in 2005 to identify poverty reduction expenditures on the basis o f a functional classification. I t was expected that specific codes in the Plan budgdtaire et comptable de l ’ t a t (PBCE) would allow the identification o f functions and sub-functions corresponding to poverty reduction expenditures. However, the method adopted for the codification o f these expenditures, based on the six pillars o f the I-PRSP, proved too vague to identify the corresponding budget lines. The DBC has therefore decided to consider all budget lines as expenditures to fight poverty, except for the lines allocated to the armed forces. This approach could create confusion, and allow a favorable treatment o f expenditures not actually linked to the fight against poverty. The authorities are aware o f this risk, and plan to better identify poverty reduction expenditures in the preparation o f future budgets, using the full PRSP adopted by the Government in September 2006 and discussed at the IDA and IMF Boards in March 2007 (see B o x 4.3). Budget Presentation 4.32 The 2007 budget document i s not user-friendly and does not contain the information needed for a thorough analysis. The budget i s a single document consisting o f the Finance L a w and annexes with detailed selected revenues and expenditures. The Finance L a w includes a synthetic document in the f o r m o f a Table o f the Government’s Financial Operations (TOFE). Current expenditures are relatively detailed by departmentldirectorate and by division in each ministry, with a relatively important share (about 52 percent o f current expenditures) allocated to personnel expenditures. Support expenditures (charges communes) are detailed by ministry. However, the other expenditures are more globalized. The expenditures o f the FNR are presented in a very succinct form, but with n o synthesis o f the budget for each ministry and for the whole Government. 66Of the FBu 94 billion expected (Le., more than 30 percent of total budget resources), only 23.2 billion are recorded in the 2007 Finance Law. 85 - Box 4.3: The Full Poverty Reduction Strategy Paper PRSP, 2006-09 The PRSP finalized and published in September 2006 presents the vision o f the medium and long-term development o f Burundi, and sets objectives in the fight against poverty in line with the Millennium Development Goals (MDGs). The PRSP revolves around the following principles: (a) refocusing the State o n its key functions (Le., in service provision); (b) the consolidation and keeping o f peace and security; (c) capacity strengtheninghuilding; (d) energizing economic growth; (e) strengthening the community approach to economic and social development; (9 stressing the central role o f women in development; and (g) promotion o f a new partnership with donors. The sectoral programs and the macroeconomic framework are oriented toward four strategic priorities: the improvement o f governance and security; the promotion o f sustainable and equitable growth; the development o f human capital/resources; and the fight against HIV/AIDS. Box 4.4: Recommendations on the Coverage and the Presentation o f the Budget Law I n the short term: Consolidate the recurrent budget (BO) and extraordinary investment budget (BE4 accounts at the central bank. I n the functional classification: (a) identify functions and sub-jiunctions linked to the new priorities of the PRSP; and (b) detail, ifpossible, the sub-jiunctions concerning the priorities for poverty reduction selected by Burundi (e.g., fight against HIV/AIDS; craftsman sector; micro-projects, etc.) . Develop summary table o f the budget law for each ministry and for the Government; and also for HIPC-financed expenditures; and for the expenditures for the fight against poverty. Provide details o f non current expenditures, in particular the expenditures of the national road fund. I n the medium term: Establish, in cooperation with development partners, a system centralizing the information on all expected foreign financing; and align foreign financing with the budget process preparation cycle, so that the Budget Law takes into account all the resources (domestic and foreign-financed) and all operations financed under the Government, including the foreign- financed ones. Update, control, and improve the management and the computerization o f the payroll and the wage bill o f public servants; develop in the Budget Law a table o f public service roll by category and by ministry, complemented by the list o f positions filled and by information on the non-permanent manpower. Consolidate the ordinary budget (BO) and the investment budget (BE4 by ministry, under a single chapter, and using the same classification, in line with the new Organic Finance Law; and take into account in the recurrent budget expenditures for investment projects. 86 Budget Preparation 4.33 The budget preparation process i s weak at all levels. The Director General o f Planning (DGP) i s in charge o f preparing the macroeconomic framework forecasts. The model in use i s very old and has numerous variables that are difficult to adjust to take into account the social and other aspects o f the fight against poverty. The preparation and updating o f the macroeconomic framework also suffer from the absence o f statistical data, as n o national accounts have been produced since 1988.67 4.34 The M E F D C has neither the staff nor the tools for medium-tern forecasts o f revenues and expenditures. As a result, i t s long-term budget policy i s not based on medium-termprojections, including an assessment o f the impact o f current decisions on future expenditures. 4.35 Another major problem i s that the schedule for budget preparation always starts late, and thus does not give the divisions concerned the time to prepare effective and realistic forecasts. For example, the preparation o f the 2007 Finance Law was compressed into 2 months (August- September 2006); and the preparation o f the 2008 finance bill started only in July 2007, instead o f March 2007, as scheduled. The D B C has difficulty undertaking the preparation o f the supplementary budget at mid-year in parallel with the preparation o f the finance bill for the following year. For the current budget, the D B C usually proposes the same amount for the BO (salaries excluded) as the year before, adjusted by the inflation rate. 4.36 Further, the DBC’s capacities for budget preparation are very weak. The preparation process remains formal, often not well coordinated and disconnected from the macroeconomic forecasts. The D B C i s not sufficiently involved in revenue forecasts, and these forecasts are not reliable. 4.37 In addition, MEFDC’s Framework Letter to the line ministries, which i s supposed to set the parameters o f the draft budget, includes n o data or n o expenditure ceiling, and gives no indication o f the general parameters/developments that must be taken into account. The letter contains no table or figures in the text to help line ministriesprepare their budget proposals. 4.38 The link between the full PRSP and the annual budget i s not strong. For the preparation o f the 2007 budget, the Framework Letter issued by the MEFDC stressed that the priorities o f the budget should be expenditures in the social sectors, in particular in education, health, and the fight against HIV/AIDS; the repatriation and reinsertion o f refugees, displaced people, and victims o f natural disasters; as well as for expenditures in economic sectors such as agriculture and livestock and husbandry; mining; energy; and public works and equipment. However, some social services such as the Ministry at the Presidency in charge o f the fight against HIV/AIDS (which was recently merged with the Ministry o f Public Health) and the Ministry o f Education did not benefit from an increase in their budget allocations for 2007. As the B O and the BE1 are not consolidated by ministry, and as the data on foreign-financed operations in these social 67Technical assistance under the PAGE has started in 2007 with the objectives o f changing the model used; assisting the vice-Ministry o f Planning (VMOP) in the revision of the macroeconomic framework; preparing the overdue national accounts; and training all the staff concerned. AfDB i s also assisted VMOP in this area through TA. 87 sectors are not comprehensive, it i s not possible to assess the total resources (domestic and foreign-financed) allocated to these ministries, or to identify and monitor a large proportion o f Burundi’s poverty reduction operations. 4.39 The non-recording or partial recording o f foreign-financed and NGO-financed operations i s a serious barrier for authorities tasked with devising a meaningful strategy for the fight against poverty. 4.40 The success o f this approach will be dependent on the willingness o f development partners to channel their funds through the Burundi P E M system. This, in turn, will only be forthcoming when donors assess that the level o f fiduciary risk associated with this system i s acceptable. B o x 4.5 provides the main recommendations for budget preparation. The sequencing o f reforms i s therefore crucial to their chances o f success: First, the Government needs to demonstrate that i t has taken sufficient steps to improve i t s P E M performance, within the context o f the new Organic L a w (clearer budget procedures, simpler expenditure chain, reintegration o f a proportion o f off-budget expenditure into the budget, more effective internal controls, professional internal and external auditing, adequate accounting and reporting, strengthened accountability mechanisms); In parallel, donors should reinforce their coordination at the sectoral level and improve the information flow on external project financing, at both the planning and execution stages; opportunities for pooled funding at the sectoral level should be explored; the Government should review the PIP process to ensure that it i s consistent with the PRSP, the future MTEF, and the yearly budget, and to encourage donors to provide data on i t s execution; A third step would be for development partners to channel sectoral funds directly to the budget, once adequate fiduciary controls have been established; this step does not necessarily involve a move to performance budgeting, although it would be compatible with such a move were the Government to decide that it i s warranted. 88 Box 4.5: Budget Preparation- Main Recommendations I n the short term: Decide on a schedule for the preparation o f the budget, starting with the 2008 budget, based on a circular by the Minister of Finance, starting the preparation in February, following the availability of the macroeconomic framework. A draft schedule is presented in Annex 1. Improve budget credibility by requesting line ministries to prepare three scenarios o f budget every year based on: (a) the previous year allocation; (b) a 10percent increase ofpublic resources relative to the previous year; and (c) a 10 percent reduction in public resources relative to the previous year. Improve the content o f the MEFDC’s Framework Letter by including, in particular, expenditures ceilings by ministry; and also information on the macroeconomic forecasts for 3 years, and the general evolution trend ofpublic expenditures on the government economic and social strategy and the expected results, particularly in the fight against poverty. Theframework should be sent by end April at the latest. Establish a clearer link between the PRSP and the budget. And make sure that the links between the PRSP and the functional classifcation are reliable and operational, to allow the identification and monitoring o f the expenditures, to fight poverty during budget preparation and implementation/execution. Begin the M T E F through - (a) a multi-year macro framework; and (b) complete sectoral budget programs. Ensure that budget laws and payment laws (Loi de r2glement) enacted are published; facilitate access by civil society to information on these laws, in particular by posting them on the Government website. I n the medium term: Strengthen the capacity o f the DGP to prepare a reliable macroeconomic framework for the budget preparation. Develop, as afirst step, a medium-term budget Cfinancial)framework (Cadre budgitaire tr moyen terme, CBMT) which includes realistic forecasts o f Government revenues, and expenditures for each budget entity; and which allocates the resources according to the medium-term priorities defined by the Government. Strengthen the capacity o f the DBCfor budget preparation. ‘il unit in charge o f the synthesis of economic data and revenues forecasts ’’ could be establishedheated to: (a) ensure that economic constraints are taken into account in budget policy, shed economic light on the budget preferences of the Government; and ensure, in particular, that the macroeconomic framework and the revenue forecasts for the budget preparation are realistic; (b) coordinate, with the diferent interlocutors concerned, the work for the budget preparation, the supplemental budgets, the consolidation o f the B O and the BEI, and the synthesis of trade-offdecisions. 4.41 The following section will discuss issues related to budget execution. 89 D. BUDGET EXECUTION 4.42 This section will cover solely the budget expenditure channels. Accounting, audit (internal and external), oversight, and cash management will be treated in separate sections. Expenditure Channels Current Situation 4.43 Budget execution in Burundi i s organized as follows: 0 The MEFDC i s the only manager (ordonnateur) o f the budget, but it delegates that responsibility to the Ordonnateur trksorier du Burundi (OTBU), which i s in charge o f payment orders and keeps the books for the budget implementation. 0 T o implement the budget, the MEFDC uses credit managers: the Secretary General at the Presidency for the Presidency, the ministers for line ministries who can transfer that responsibility to their chefi de cabinet; all division chiefs reporting directly to the President. 0 Credit managers are assisted by financial controllers or comptables des dkpenses engage'es, accountants seconded by the MEFDC to assist in the administrative management o f their budget allocations. The DBC carries out the budget controls for commitment, and liquidation sub-phases o f the expenditures. The central bank i s the State's cashier (caissier de lJktat). 4.44 The execution o f expenditures i s done in four phases: commitment; liquidation; payment order; and payment for the BO as well as for the BEI. With the inception o f the computerization scheme (SIGEFI) in 2005, the four phases were integrated in the system o n the basis o f their dates.68 Main Problems and Issues 4.45 A substantial amount o f expenditures i s executed using exceptional procedures and other irregularities. This i s particularly the case for: (a) most foreign-financed expenditures, which are paid from accounts opened in commercial banks; (b) extra-budgetary accounts; and (c) The new system makes i t possible, in principle, to check the processing time, and the unpaid balance o f an expenditure b y comparing the dates: (a) in the commitment phase, the two dates are the dates o f processing, and validation (order form); (b) in the liquidation phase, the two dates are the dates o f the processing, and the validation o f the order form; and (c) in the payment order form, the two dates are the date o f the processing o f the certificate o f indebtedness (CI) and i t s validation. In theory, the validation o f the C I equates the transmission o f the C I to the central bank for payment. 90 expenditures made from Government’s bank accounts by non-autonomous public en ti tie^.^' The nature and the operating principles o f Government’s accounts are presented in B o x 4.6. 4.46 These exceptional procedures translate into the direct use o f public funds by ministries and ministerial entities without using the regular procedures; i.e., without control and accounting o f these expenditures by the MEFDC. They obscure the budget process, and reduce the relevance o f reports on budget execution, because most o f the operations financed through the exceptional procedures are not included in the budget reporting. B o x 4.7 discusses the complex nature o f budget execution process. Box 4.6: Government Bank Accounts The Government has accounts opened at the central bank, called OTBU accounts (1 101/01 for the recurrent budget, and 1102/01 for the investment budget). Budget revenues are deposited in these accounts, out o f which disbursements are made to pay budget expenditures and the debt. Besides these two accounts, the central bank has there are several other government accounts at the request o f the MF, for various operations, including recurrent expenditures. There i s another procedure by which the MF requests, by letter to the Governor o f central bank, the opening o f “Government accounts.” That letter i s followed by another one by which the MF informs the Governor o f the amount to be transferred from the 2 OTBU accounts to the “Government accounts,” and the names o f the persons with signature on these accounts. These “Government accounts” are usually opened for line ministries offices, provincial offices, special funds, public entities, specialized public offices, and for national counterparts to foreign-financed programs and projects. The beneficiaries enjoy a de facto autonomy. The operations financed from these accounts are not taken into account in the budget reporting or public accounting, even though these operations are public expenditures paid for by budgetary funds. The balances o f these accounts also are not taken into account in the State’s cash situation. On December 17, 2006, there were about 300 government accounts (80 accounts were closed in 2006, and about 45 in 2007). The balance on these accounts amounted to about FBu 13.4 billion on December 17, 2006, or about 5 percent o f total budget revenues. A committee established in 2005 in the M E F D C i s in charge o f gradually reducing the number o f government accounts that are not justified, with, as a first step, the regrouping o f all accounts for the same government entity into a single account. In an attempt to address the above issue, the Government recently suspended, by a provision in the 2008 Budget Law, authorizations for the opening o f extra-budgetary accounts, with the exception o f accounts for foreign-financed projects and those o f special funds listed in the budget, pending the issuance o f new legislative and regulatory provisions. 69 A payment i s derogatory to the rules when allocations are made available to entities without legal personality and financial autonomy. 91 Box 4.7: Complex Budget Execution Process In Burundi, the various phases o f budget execution are particularly long and onerous. The budget execution process i s the sole responsibility o f the Ministry o f Economy, Finance and Development Cooperation. There are three sets of controls at the commitment level, three at the liquidation level, three at the authorization phase, and one at the payment phase. These multiple controls at each stage o f budget execution do not improve overall control and have led to the development o f exceptional procedures to speed up the budget execution process. There i s evidence that the cumbersome process in place leads to delays in payments and thereby reduces budget execution rates, without improving overall control. On average, i t takes 60 days for a payment to be made after receipts are submitted. Donors have highlighted the need to rationalize and streamline control procedures, a recommendation which the Ministry o f Finance has adopted its reform strategy (Strategie sectorielle du Minisdre des Finances). 4.47 The formal expenditure process i s overburdened by a multiplicity o f compliance/regulatory controls, most o f which are redundant, with a questionable efficacy. I t could well be that the multiplicity o f controls encourages Government officials and suppliers to use exceptional procedures. The controls are concentrated in D B C and the Directorate o f Public Accounts (DC) and focus mainly on formal compliance. The central bank, as Burundi’s cashier, makes a full control o f the validity o f the transaction before payment. There i s n o manual o f procedures for budget implementation. 4.48 Several Government’s offices have special accounts which represent possibilities for them to spend above their appropriation ceilings. These accounts have n o justification, and do not allow active and effective cash management. The report o f the CdC for 2005 states that these account^'^ are difficult to control. 4.49 The Government does not have the wage bill under control. Even though the public service wage bill represents a sizeable share o f the State’s budget,” the r o l l o f public servants i s not updated regularly and includes many irregularitie~.’~ The software for the computerized management o f the payroll i s obsolete, and the MEFDC i s not in a position to effectively control the payment authorizations (liquidation) and the payment o f public servants’ wages.73 4.50 The authorities’ decision in M a y 2007 to increase the salaries o f civil servants across the board by 34 percent i s likely to make the effective management o f the budget and i t s balancing more difficult. This decision will further boost the already sizeable wage bill as a percentage o f 70 These special funds are the following: the “Fonds non encore affect? managed by the MEFDC; the “ taxe rimunkratoire sur attestation d ’impdt” jointly managed by the MEFDC and the Tax Director; the ‘tfonds de lutte contre la fraude”, and the “computerization tax” jointly managed by the MEFDC and the Customs Director; and the ‘tfonds d ’appui ci 1 ‘administration territoriale” jointly managed by the MEFDC and the Minister of the Interior. 7’ In 2006, the wage bill represented 24 percent of total expenditures, about 10 percent of GDP, and has been increasing since 2002. The payroll includes about 100,000 persons, a higher percentage of the population compared with neighboring countries. About 50 percent o f the personnel are in the army and the national police. 72 A recent audit of roll o f civil servants (army and police agents excluded) by the authorities has found 4.000 “ghost workers” on a total of about 46,000 civil servants. The transfer of the wage bill management from a private enterprise to the MEFDC i s underway with the assistance o f 73 PAGE, a WB-financed project, which i s also assisting in the development o f a new software program for the management of the public service data, which should in the medium-term facilitate the control of the wage bill. 92 total budget expenditures, and as a percentage o f the GDP. The financial adjustments needed to meet this increase will not be easy. 4.5 1 The certificates o f indebtedness go through numerous stages, including the physical transmission to various actors (tax administration for possible opposition, central bank). T h i s process translates into risks o f the loss o f the certificates, and a complicated accounting monitoring. Box 4.8: Budget Execution - Main Recommendations I n the short term: 0 Streamline budget execution procedures through the production o f clear budget execution manuals strictly limiting the use o f exceptional execution procedures, and train relevant staff within M E F D C and the line ministries on newprocedures; Reintegrate most o f the state’s expenditures into normal budget channels processed and accounted for through a stabilized information system (see below); Improve internal controls over payroll by introducing monthly checks o f staff lists as a basis for payroll data; and Introduce more eflcient accounting and budgetary procedures for the exhaustive recording, monitoring, and clearing o f arrears. I n the medium term: Define an action plan to strengthen the absorptive capacities o f ministries in order to improve rates of budget execution; Strengthen the internal audit unit of the MEFDC, concentrating on its efforts in the monitoring o f the budget, reviewing procedures, and improving budget management; Establish structures for the monitoring o f budget execution in the sector ministries, and formalize the interaction between M E F D C and the sector ministries on budget execution monitoring; Follow systematically delays in the execution o f expenses at every stage in budget execution procedures; and 0 Abolish useless transmission o f expenditure data (tax department, central bank). The tax administration could provide ahead o f time and regularly update a list of evidence to support any submission for payment. This list would be consulted by the service concerned before any payment is made. Another possibility is to provide the tax department with the detailed list of the payments scheduled and the name o f the corresponding suppliers. INTEGRATED E. THEINTERIM SYSTEM (SIGEFI) 74 FINANCIALMANAGEMENT Current Situation 4.52 Following the adoption in 2004 o f the new Budgetary and Accounting Plan o f the State (PBCE), the Ministry o f Economy, Finance, and Development Cooperation (MEFDC) initiated, 74 Source: “SIGEFI Implementation Review and Recommendation for Implementation of an Integrated Financial Management System (IFMIS),” November 2007 by Mr. Herbert Rwamibazi, IFMIS Expert. 93 with the assistance o f consultants (Soft Center) financed by IDA’S PAGE project, the consolidated computerization o f budget and accounting operations in the expenditure process through the SIGEFI. I n 2006, M E F D C received the final delivery o f the SIGEFI modules o n commitment (engagements), liquidation (liquidation), and authorization for payment (ordonnancement), but without the training o f the users, and without proper documentation. The MEFDC accepted the provisional delivery o f the module for payment baiements) at the same time, and SIGEFI started operations in the MEFDC. The system was not based o n a defined general architecture with the target o f integrating all public finance operations. The treatment o f the wage bill was done by the Centre National Informatique (CNI), a private enterprise. A manual interface has been devised to take into account in SIGEFI the data o n the liquidation o f personnel expenditures. 4.53 The implementation o f SIGEFI was undertaken as an interim measure to automate budget execution, but has faced several challenges, including (a) the absence o f a properly designed framework o f user requirements, functional processes, and defined scope; (b) the absence o f a team with experience and expertise in the design and implementation o f government financial management systems; and (c) the absence o f project management standards that would have created an environment o f escalating implementation scope over time. 4.54 F r o m an ICT perspective, the SIGEFI implementation was based o n the development o f a bespoke system designed using Visual Basic 6 and running on SQL Server 2008. An I F M I S expert recruited by the authorities to identify current implementation challenges, and recommendations o n the way forward, indicated in his report that the platform o n which the SIGEFI application was built three years ago i s n o w obsolete. The most serious concern raised by the expert was that the implementation o f SIGEFI by Soft Center was not matched with the presence/participation o f counterpart staff o f MEFDC, and after three and a h a l f years o f implementation, there has been n o knowledge and skills transfer to any internal staff o f MEFDC, necessitating that all modifications, reports, information, and coding required f i o m the system are only undertaken by the developers o f the system. 4.55 F r o m an accounting point o f view, the SIGEFI conceptual design was not based o n double entry accounting, but was designed primarily for single entry accounting. Critical accounting processes, such as documentation backing accounting entries and workflow for transactions, were not built into the system. The system was also missing controls necessary for effective management and execution o f the budget, including typical control triggers that would alert certain users and oversight teams when there i s a breach in payments beyond a certain level that have been effected without the proper authorization. 4.56 In addition, critical interfaces with C N I for the payroll, the central bank, the revenue department for revenue information, and the debt department for debt data, have not been satisfactorily designed and built. The I F M I S expert identified potential security risks o n the SIGEFI network that could compromise data and information held on the system and the servers. The absence o f a disaster recovery plan for the data held within SIGEFI was further identified as a potential risk that could result in the loss o f all the data held o n the system. With the already scanty government data, the loss o f data o n the budget execution would be a severe b l o w to the reform processes currently taking place. 94 4.57 With respect to payroll processing, the interface requires the use o f transfer floppy disks by C N I and a detailed list o f public servants consistent with the diskettes. I t worked well from January to March 2007. Since April 2007, the personnel expenditures are no longer produced by SIGEFI because o f inconsistencies between the disks and the detailed roll/list o f civil servants transmitted to the MEFDC. For the month o f October 2007, C N I failed to produce any disks at all for the payroll information, and September 2007 figures were used as a basis for the payments to public servants. The problem with the non-integration o f the personnel expenditures since April 2007 i s the manifestation o f the lack o f an efficient management o f risks and incidents (planning and organization); o f development procedures (acquisition and implementation o f solutions); o f control o f services rendered by a third party (deployment and support); and o f the monitoring o f the process o f project implementation. It should be noted that the C M payroll data stands a risk o f manipulation with the continued use o f a disk to transfer payroll data to SIGEFI, a risk that could be eliminated by a direct interface o f the two systems. 4.58 At a cost o f US$45,000 for the design and implementation o f an interim system for budget execution, SIGEFI has achieved the objectives o f ensuring that the budget, once approved, can be tracked through the process o f commitment, liquidation, authorization for payment, and payment. However, significant deficiencies s t i l l exist in the system and need to be addressed to ensure the application’s stability in the interim to enable the government to produce information with proper classification and accounting treatment. Box 4.9: Interim Integrated Financial Management System - Main Recommendations I n the short term: Stabilization of SIGEFI: SIGEFI requires stabilization to enable the rectijkation o f certain functionality and improve the following: ’’ o Implementing recent recommendations with a view to stabilizing SIGEFI and putting in place some additional corrective functionality to enable its usefor the period 2008 - 2009; o Recruiting an Integrated Financial Management Information System (IFMIS) advisor to work with M E F D C on a work plan to ensure the stabilization process and creation o f a team that w i l l receive knowledgefi-om Soft Center and train the additional users of the system. The IFMIS advisor would also support the strengthening o f a SIGEFI Unit to be put in place to manage the current and future activities o f the system; and o Putting in place a support and maintenance agreement with Soft Center following completion o f any outstanding work, and with the team in place, ensure separation o f software developers and software users to reduce the potential risk offiaud that could resultfi-om having developers o f the system gaining access to the actual data, which would create the potential for modi5ing the current program and efecting unauthorized changes. I n the medium term: i it is necessary to replace SIGEFI with a new system better The Government w i l l need to decide f adapted to the Government’s needs or to adopt the existing integrated information system for financial management. ’’Implement recommendations contained in the report by the IFMIS Expert’s review o f SIGEFI. 95 4.59 In i t s decision to adopt a filly functioning IFMIS, the Government should take stock on relevant experience around the world. The World Bank has experience with implementation o f IFMIS in transition and developing economies, including in sub-Saharan Afiica. A Bank review in 2003 looked at 34 IFMIS projects across 6 regions and identified the following findings: On average, IFMIS projects took 7 years to complete, but this average reached over 9 years in Africa; 75 percent o f projects underwent component changes during the implementation; the least expensive systems (e.g., Albania, 366 users and Tajikistan, 503 users) cost around US$9 million; only 6 percent o f projects were assessed by Bank staff as highly likely to be sustainable, and 69 percent were assessed as likely sustainable. The main challenges identified were lack o f commitment to the project, ineffective project coordination, unrealistic budget forecasts, over-ambitious project designs, and insufficient attention to the need to streamline budget execution procedures before defining the IFMIS specifications. However, in cases where the implementation has been successful (e.g., Uganda, Tanzania, Kenya - despite connectivity issues, Sierra Leone, Ethiopia), i t was a significant driver in improving P E M outcomes and increasing donor confidence in the P E M system. Tanzania now has the highest PEFA (Public Expenditure Financial Accountability) scores within East Africa (18 percent A/B+ and 8 percent D/D+ compared to 8 percent A/B+ and 37 percent D/D+ for all sub-Saharan Africa). 4.60 The decision to proceed with an IFMIS project should be made by the authorities following a carefil consideration o f i t s pros and cons. Should they decide that the benefits o f the move to a IFMIS would outweigh i t s costs and express a strong commitment to an IFMIS project, the following considerations would need to be borne in mind: I t i s crucial to focus first on reforming the existing budget execution process in order to avoid defining IT specifications on the basis o f inefficient processes. The main focus o f the project should not be information technology, but capacity building: capacity should be identified at the outset and a clear capacity building plan should be embedded in the project design in order to increases the chances o f sustainability . Functional reform priorities should be determined in line with the P E M reform strategy that this report calls for - i t i s important to “think small” and avoid scope creep during the project, by building extremely strong project management discipline into the project design. The costs o f the various technical solutions available should be compared on the basis o f the total cost o f ownership (including hardware, connectivity, maintenance, training, etc.), as opposed to simply the software cost, and with capacity building and long-term sustainability in mind. The specific challenges o f the Burundi context should be taken into account at the initial stages o f the project (power supply, connectivity, computer literacy rates, etc.). 96 F. PUBLIC ACCOUNTING Current Situation 4.61 A new Plan Budgitaire et Comptable de l’Etat (PBCE) was implemented on January 1, 2005.76The key provisions are summarized in Box 4.8. Substantial progress has been achieved by the Government in the area o f public accounting, in particular: (a) all 2005 operations were processed in March 2006; (b) improved budget reporting; (c) establishment o f a general balance o f all accounts; (d) the production o f a provisional manual on budgetary and accounting procedures; and (e) the production o f quarterly TOFE. Box 4.10: The Key Principles of the PBCE Harmonization o f the budgetary and accounting nomenclatures. Creation o f different classifications allowing the analysis o f the budget, the monitoring o f i t s implementation, and the restitution o f the information as needed (administrative, fimctional, expenditures for the fight against poverty in six categories, by economic nature). Harmonization o f the PBCE nomenclatures with the nomenclatures in the 2007 Government Finance Statistics Manual 2001 (GSFM) and with those o f the TOFE to facilitate the preparation o f statistic data requested by the World Bank and IMF. Use o f double-entry system for public accounting by all public accountants. Automatic production o f the accounting data through SIGEFI. Main Problems and Challenges 4.62 The manual o f accounting procedures i s not yet finalized and i t s use i s limited because the data have not been regrouped by users’ category. The version in use, produced in 2005, i s provisional, pending the stabilization o f some procedures and the rationalization o f accounting methods. The current manual can be confusing as it mixes existing procedures, and the new procedures envisaged. 4.63 In addition, there are s t i l l outstanding problems concerning budget and accounting reports and their regular production. The conditions for the reliable reporting are not fully met. The CdC’s report on the 2005 budget found discrepancies, for both revenues and expenditures, between the data reported by the directorates o f the MEFDC and the data in the budget report presented by the MEFDC to the parliament (Loi de rsglement). 4.64 Following are some o f the problems: 0 The data on expenditures (committed, liquidated, paid) are incomplete because they do not include HIPC-financed expenditures; and most o f the foreign-financed operations and expenditures financed out o f Government’s accounts (extra-budgetary) are not reported in the public accounting. 0 There i s a significant delay between the time funds are cashed and their accounting because o f the delay in data processing. 76 The new PBCE was enacted by PresidentialDecree No. 100/186 o f 12/21/2004. 97 e The general accounts balance o f December 2006 included several discrepancies: the credit balances and the debit balances were not balanced at the beginning or at the end o f the period. These discrepancies are evidence that the accounting procedures are not filly mastered and that internal control i s weak or lacking. e Budget implementation reports are not produced at regular intervals, but only upon request (inside or outside the MEFDC). Box 4.11: Public Accounting - Main Recommendations I n the short term: e Produce monthly trial balances and regular reports on revenues and expenditures (including donor-fundedprojects) to allow verification and control77; e Record every operation in the SIGEFI at the various stages o f the expenditure process; and establish a system to record and report HIPC-financed expenditures. Establish within M E F D C a unit in charge of centralizing and analyzingfinancial and budgetary data on foreign-financed expenditures, which w i l l collaborate with sector ministries, aid coordination sector groups and the national aid coordination agencies. Financial execution reports on project expenditures should contain details on: (a) the funds received, disbursements, and transfers on bank accounts at the central bank and in commercial banks; and (b) direct payments by donors, as well as the classification o f the expenditures in accordance with State Accounting Plan. Line ministries could undertake a prior centralization o f these project financial data; and Define project financial reporting requirements in a letter to technical ministries, project coordinators and donors: The letter from M E F D C to all budget actors (line ministries, project managers, development partners) w i l l request them to transmit to the MEFDC on a regular basis reports on expenditure execution, detailing the nature o f the expenditures. F o r projects, the reports w i l l detail the disbursements received and the payments made and the source (counterparts fund, loans, or grants). I n the medium term: Institute a mandatory reporting systemfor special funds and special accounts; Complete, for each year, the necessary verifications and regularizations on the general balance of accounts before integrating it into the accounts for the following year; and Gradually develop, on the basis o f capacity and need, a unit within line ministries for collecting sectoral foreign-financed expenditures data. ” A provisional report should be produced soon after the end o f each month for the operations centralized in Bujumbura (representing the bulk o f the operations), to facilitate monitoringand planning. 98 G. CASH PLANNING AND MANAGEMENT78 Current Situation 4.65 The Direction de la trksorerie (DT), within the DGBCP i s in charge o f internal and external debt management, and o f the preparation o f monitoring o f the cash management plan. Until early 2007, the Cash Management Committee (CdT) was chaired by the Director o f DT and had the following members: the OTBU, the State’s cashier at the central bank, the Director o f Customs, the Director o f Taxes, the Director o f Administrative revenues and the Portfolio, and the DBC. The Committee used to prepare an annual cash management plan, monitored through weekly plans. The priority o f the DT was external debt management; and it lacked the staff for effective preparation o f cash planning and management. The main objective o f the cash planning was to determine the cash deficit for the reimbursement o f external debts; and debts to the central bank and their financing through central bank advances and/or treasury bonds guaranteed by central bank. 4.66 Some progress has been made by the authorities towards more effective cash management: 0 A commission was established in 2005 for the rationalization o f the Government’s accounts towards a single treasury account. Eighty accounts were closed in 2006 and 45 in 2007. 0 All the transit accounts o f the customs and tax directorates in the commercial banks have been closed. 0 The Minister o f Finance and the central bank have established a joint commission to better determine cash needs, and to organize the emissions o f Treasury bonds and obligations for the financing o f the cash needs, in replacement o f the former practice o f statutory advances by the central bank. The domestic debts have been audited, and a clearing plan i s being implemented with donor assistance. 4.67 In M a y 2007, the MEFDC established a Bureau de gestion de la Trksorerie de l’btat (BGTE) in the front office o f the MEFDC, under supervision o f the unit for monitoring reforms in the MEFDC, and with the technical and financial support o f a French TA. 4.68 An additional measure i s the closing by end 2007 o f all the sub-accounts o f the HIPC account, their consolidation into a single HIPC account, and i t s integration as a sub-account o f O T B U at the central bank. Main Problems and Challenges 4.69 The MEFDC faces the following challenges in preparing and implementing a cash plan as an effective budget management tool to reduce the risks o f building payment arrears: 0 The institutional set-up i s weak and not adapted to good accounting practices. The DT i s not the appropriate directorate to be in charge o f cash management, as most cash ’’Sources : IMF F,+D Report b y J.-L. Helis, et al., May 2007; and draft report “Mise en place d’un outil de gestion de la trisorerie de 1 ’Etat, ” b y J.-P. Le Bouder, July 2007. 99 activities are under the responsibility o f the OTBU within the Directorate o f Accounting (DC). The mission o f the BGTE in terms o f preparation, monitoring and updating the cash management plan i s not very clear/well-defined. I t s composition and i t s being under the supervision o f the office in charge o f the monitoring o f reforms in the MEFDC are not appropriate. 0 The state’s cash resources remain fragmented. In addition to the two main accounts in the central bank (BO and BEI), and despite the closing o f several accounts o f public entities, there are s t i l l close to 300 accounts in the name o f public entities. 0 Budget revenue projections are not comprehensive and do not include the amounts and schedules o f disbursement o f most foreign-financed budget supports. The accounts o f foreign-financed projects, including for the counterpart funds, are opened in commercial banks, adding to the fragmentation o f the State’s cash resources. Box 4.12: Cash Planning and Management - Main Recommendations I n the short term: Progressively reestablish the Treasury Single Account (TSA), by completing the ongoing consolidation of accounts of diferent public entities; Abolish accounts managed by entities that are not financially autonomous; and Produce, on a quarterly basis, a global cash management plan and a sectoral plan for each ministry and disseminate it publicly to enable each ministry to be aware o f the financial constraints of the entire Government. I n the medium term: Create a Cash Management Committee under the authority of the Minister of Economy, Finance, and Development Cooperation to supervise ongoing reforms and help the country move gradually toward a single treasury account. members to be appointed by the M E F D C should include the Governor of central bank and the vice Ministry o f Planning (VMOP) (see annex 2 for the details of the proposed institutional set-up); Establish a Groupe de gestion de la trtsorerie de l’ttat (GGTE) with a permanent technical secretariat consisting of a head, two deputies; and eleven non-permanent members, to assist the Cash Management Committee. Through dialogue within the Cadre de partenariat, (Partnership Framework) gradually transfer project accounts to the central bank in O T B U s sub-accounts; and obtain from donors the schedule and amount of their budget support, so it can be integrated into the budget. 4.70 Again, the success o f this approach i s contingent upon gradually building up the confidence o f donors in the fiduciary guarantees provided by the Burundi PEM system. The next section discusses internal audit and external oversight. 100 H. AUDIT AND OVERSIGHT Internal Audit Current Situation 4.71 Internal audits in Burundi are the responsibility o f the Inspection Gknkrale d’Etat (IGE), the control unit in the MEFDC (the Unit),” and internal control units in line ministries. Internal control units do not exist in all line ministries; and where they exist, their capacity and expertise are very weak. 4.72 IGE was created by decree No. 100/277 o f September 2006, under the oversight o f the MinistBre de la bonne gouvernance. IGE’s field o f investigation and its mission are wider than those o f the IGF (Inspection gknkrale desfinances): i t has a permanent mission to inspect and control the management and operations o f public services, state institutions, and enterprises or private associations under i t s control. The verifications by IGE cover all aspects o f management, and include the control for compliance with the laws and regulations, the verification o f the accounts, and the control o f the management. The new control unit in the MEFDC, for i t s part, i s now centered on the inspection and audit o f financial administrations. 4.73 IGE and the control unit in the M E F D C are to prepare, at the beginning o f each year, a draft annual verification program. They also have to produce semi-annual reports on the implementation o f their work program, and an annual report o f their activities. For both IGE and the control unit, their verification missions are not planned in advance and they use the cross- examination approach. In the case where embezzlement, fraudulent management, and malfeasance are established by the controls, the IGE’s and the unit’s final reports are transmitted to the Department o f the Public Prosecutor for criminal investigation. They inform the CdC o f the case o f gestion defait (unauthorized handling o f public funds). Main Problems and Challenges 4.74 IGE will be fully operational only in 2008. I n the summer o f 2007, it recruited 60 inspectors through international selection, with very competitive salaries.80 I t i s now housed in a furnished building large enough to accommodate the 60 inspectors. The effectiveness o f IGE will depend on taking the priority action o f defining the conditions for effective cooperation with the control unit, and with the other internal control units in line ministries. 4.75 The system o f internal control can be effective only if the recommendations and sanctions proposed are implemented, and if the activities o f the different control units are coordinated. According to the control units, the implementation o f the sanctions, particularly administrative (rkglementaire) sanctions, needs to be more systematic. ’’The control unit replaces and has the same prerogativesas the Inspection Ge‘nkraledes Finances (IGF) abolished in 2006. The salary for an inspector in IGE i s higher than the salary of a member of CdC. 101 4.76 It i s also important to ensure that the IGE and other internal audit bodies clarify their roles. The IGE could perform two types o f roles: an internal audit role and an inspection / investigation o f corruption role. As an internal auditor, the IGE would need to support the managers in line ministries by identifying weaknesses, analyzing their systemic causes, and proposing remedial actions. I t should not have an adversarial relationship with managers, but be supportive o f their risk management efforts. I t would define i t s workplan according to i t s assessment o f all kinds o f risks, including but not limited to corruption, but not on the basis o f specific corruption allegations or suspicions. The international standards that would apply to this activity would be that o f the Institute o f Internal Auditors. The techniques employed would be standard auditing techniques (questionnaires, interviews, sampling, financial analysis), as opposed to investigatory techniques, and the standard o f evidence accumulated would be sufficient for quality assurance purposes but would not amount to the level o f proof required by the Burundijustice system. 4.77 By contrast, the IGE could focus i t s resources on the investigation o f specific cases o f corruption. In that case, i t s workplan should be determined both by i t s corruption risk assessment and specific allegations or concerns brought to i t s attention. I t should have a formal process for reviewing and documenting allegations and determining and prioritizing follow-up actions. The minimum standards that would apply to this activity would be the Uniform Guidelines for Investigations approved by the Conference o f International Investigators at their 2003 Brussels Conference under the auspices o f the United Nations, as well as general due process principles. The techniques employed would be forensic and investigatory in nature, and would include cross- examination. The level o f evidence gathered should be sufficient to withstand scrutiny in a disciplinary proceeding or a court o f law. 4.78 The Burundi public sector needs for both o f these activities to be undertaken professionally, as they are complementary. However, these two types o f activities cannot easily be combined in the same unit, as the investigative function then tends to affect the relationship o f internal auditors with management and to prevent good cooperation and ownership o f the audit process by the administration, which i s critical to the implementation o f audit recommendations and hence to the effectiveness o f this function. Therefore, consideration might be given to separating the investigative function, vested in the IGE, from the internal audit function, vested in the ministerial audit and inspection bodies, including a reestablished Inspection Gine'rale des Finances (IGF). This option would help ensure a clear delineation o f responsibilities between IGE and other internal audit bodies, and help the latter establish themselves as useful management tools for the M E F D C and line ministries. 4.79 Another option would be to allow both the IGE and sectoral inspections to carry out both audit and investigative functions. It would reduce the clarity o f the separation o f roles between the IGE and other audit bodies and would risk preventing these bodies from developing a constructive relationship with managers. Therefore, should the Government decide to take this approach, care should be taken to ensure that both functions are clearly delineated within each internal audit entity and that clear criteria are determined for the intervention o f IGE on a matter within the responsibilities o f another internal audit body, such as IGF if it i s recreated. 102 Box 4.13: Internal Audit - Main Recommendations I n the short term: Expand the capacity o f ministerial audit units and the investigative role o f the Inspectorate General o f the State (IGE) for corruption cases; and Clarifv the respective roles of the I G E and internal audit units in line ministries, and separate the audit and investigation functions; consider specializing I G E on investigations o f corruption cases, while developing audit approaches in sectoral bodies. Alternatively, create separate sections on audit and investigations within inspectorates and define rules or criteria justifving the involvement o f I G E on a matter within the purview o f another inspectorate. I n the medium term: Ensure that both audit and investigative functions are performed professionally, in compliance with relevant international standards through adequate recruitment, training, quality assurance processes, peer review, and certification; Identifv capacity needs and constraints in the auditing and investigative profession and put in place technical assistance to address them systematically: and Define adequate follow-up and monitoring procedures for both investigative and audit work: and measure the buy-in o f managers on audit recommendations at the time o f issuance of the final reports (percent of agreed recommendations) to increase the chances of implementation. External Control and Budget Reporting to the Parliament (Loi de r2gZenzent) Current Situation 4.80 The Cour des comptes (CdC) was established by Article 178 o f the Constitution o f Burundi, which provides that the CdC “is in charge o f reviewing and certifying the accounts o f all public services. It also assists the Parliament in the control o f the implementation o f the finance law. The CdC presents to the Parliament a report o n the compliance o f the State’s general account with the existing laws and regulations, and confirmshtates whether the fundsh-esources have been used in compliance with existing/established procedures and with the budget approved by the Parliament. The CdC sends a copy o f its report to the Government., .” The CdC report i s a public document available to the press and c i v i l society. 4.81 In accordance with the L a w o f March 19, 1964 establishing/defining the regulatory framework o f the public accounting o f the State (Article 31), the Minister o f Finance has the obligation to send to the CdC the general account o f the State and the draft budget report to be presented to the Parliament (Loi de rgglement) within three months o f the end o f the fiscal year. Similarly, in accordance with the L a w No. 11002 o f March 3 1, 2004, which establishes the CdC and states i t s missions, organization, and functioning, and provided that the central bank’s accounts role as the state’s cashier, the operations o f OTBU and the certified accounts o f public services, with all supporting documents attached, are transmitted to the CdC by March 3 1 o f the 103 new fiscal year, for review (Article 69). These two documents and the State’s public accounts” are the basis for the CdC’s report on the compliance o f the State’s general account with the budgetingreporting laws and regulations.** Main Problems and Challenges 4.82 The CdC i s confronted with several challenges, the most important being the lack o f reliable public accounts. Despite ongoing important efforts by the Directorate o f Accounting to improve the quality o f the accounts it prepares, several problems remain, as reported by CdC in i t s analysis o f the FY2005 budget implementation report. These problems, listed in B o x 4.9 below, urgently need to be resolved. Another problem which deserves attention i s the recent challenge o f the Constitutional Court o f part o f the jurisdictional powers o f the CdCSg3 4.83 The lack o f an obligation for budget reporting for special funds and accounts does not allow the CdC to fully fulfill its responsibilities. There are numerous exceptions to the rule o f budget universality (i.e., the use o f a single Treasury account). These exceptions include the authorization for the commitment, liquidation, authorization to pay, and direct payment o f all or part o f the credits allocated through extrabudgetary special accounts; the delegation o f power to some credit managers for the direct payment o f some expenditures; and the collection o f specific revenues for extra-budgetary accounts. These exceptions make it difficult to review the State’s budget implementation o f both revenues and the expenditures. 4.84 The CdC also faces significant challenges in terms o f capacity and resources, which could impair its operational independence and effectiveness (ref. 1977 International Organization o f Supreme Audit Institutions (INTOSAI) Lima Declaration and I N T O S A I guidelines). As the country’s only Supreme Audit Institution, it should be guaranteed resources that are commensurate with i t s preeminence and at least equivalent to those o f the IGE. For filing reasons, the MEFDC sends to the CdC only the balances of accounts produced by each public accountant, with a consolidated balance. The supporting documents are not sent to the CdC but are at its disposal for consultation during its control processes. For 2005, the State’s general account document, titled “Report on the Implementation o f the 2005 Finance Law,” was sent to the CdC in May 2006, and the draft Loi de rsglement for FY2005 was sent on July 2 1,2006. 83 The Constitutional Court recently decided that the magistrates of the CdC do not have the power to try public accountants for their accounts and to require them to refund. The CdC magistrate can make a judgment on the accounts, but should refer the matter to the courts for the order to refund. This makes it uncertain which institution will ensure financial and budgetary discipline. 104 Box 4.14: Main Findings of the CdC’s report on FY2005 Budget Implementation Report and Loi de rgglement The mainfindings are the following: The data provided by the M E F D C in the budget implementation report, and in the Loi de rgglement are divergent (are not the same), and do not correspond with thoseprovided by the different technical services. Overspendings werefound. A difference o f about FBu 1.4 billion was found between the amount o f revenues for the gasoline special fund in the L o i de reglement, and the budget implementation report. Budget revenuesfiom foreign trade reported by the Customs Directorate are higher (by about FBu 395 million) than the amounts mentioned in the budget implementation report. 49.55 percent ofpersonnel expenditures (about 37.49percent o f the total recurrent expenditures) are executed outside the normal expenditure process. There is a difference of about Fbu 5.5 billion in the amounts o f expenditures for goods and services between the budget implementation law and the L o i de rkglement. The management o f mission allowances remains a problem: part is managed by the Secretary General of the Government, andparts are managed by the Minister o f Finance out o f the contingency funds and extrabudgetary special funds and accounts. The lack of afiamework for the management o f the State’s debt creates some dificulties in the accounting oDerations. Box 4.15: External Oversight - Main Recommendations I n the short term: Strengthen CdC through increased resources, technical assistance, and a twinning arrangement with an external similar organization (e.g., the CdC o f Belgium). strengthen the Cour des Comptes (CdC) through increased resources and technical assistance Expand the capacity o f ministerial audit units and the investigative role of the Inspectorate General o f the State (IGE) for corruption cases; Continue and finalize the ongoing collaboration with the Directorate o f Accounting on the administrative discharges o f accounts (apurement); and Develop cooperation and coordination between the CdC and the IGE, within the fiamework o f their different respective roles and institutional standing. I n the medium term: Enhance Government’s capacity to act upon reportsfiom the CdC and IGE, including sanctioning those involved in misappropriation and responding appropriately to information on poor policy performance; Enhance the capacity o f Parliament to review the accounts prepared by the CdC; Continue and finalize the ongoing collaboration with the Directorate o f Accounting on the administrative clearance of accounts (apurement); Develop, with the finance committee o f the Parliament and the Directorate o f Accounting, an approach to: (a) determine the modalities for the presentation o f the report on budget implementation (Loi de rgglement); and (b) review the regulatory rules in this area; Formalize/finalize with the DBC and the Directorate o f Accounting, a relevant nomenclature for justification documents, for budgetary operations, and for Treasury operations; and Increase the controls and ver8cations “in situ” of the public entities not submitted to public accounting rules. 105 I.PROCUREMENT REFORMS Current situation 4.85 The procurement reform has been on Government’s agenda since March 2001 when the Government organized the first workshop on this issue. In June 2003, the Government established a steering committee to carry out an assessment o f the national procurement system and propose a procurement reform action plan. I n 2004, the Bank issued a Country Procurement Issue Paper (CPIP). B o t h the Government and Bank assessments concluded that Burundi’s procurement system has a significant number o f deficiencies: (i) the legal and institutional framework i s outdated; (ii) the environment i s not conducive to effective procurement, as it i s plagued by political interference; and (iii)there i s a lack o f accountability, as well as non-compliance with existing procedures, either because the rules and regulations are not widely known or because o f a lack o f capacity to handle procurement in an efficient manner. 4.86 In June 2005, the steering committee reviewed and updated the Government’s action plan and merged it with the CPIP action plan. The integrated procurement reform action plan was built o n five components: (i)revision o f the institutional and legal framework with a view to separating the execution function from the regulatory function; (ii) modernization o f procurement procedures; (iii)strengthening o f procurement capacity; (iv) setting up an independent ex-post control system; and (v) enforcement o f adequate measures against corruption. 4.87 Finally, in February 2008, seven years after the launch o f the procurement reform, the Government enacted a new Procurement Code and institutional framework that complies with the OECDDAC recommendations on Public Procurement Reform. The Code will become effective by October 4, 2008, with the delay aimed at enabling the Government to set up the Procurement Regulatory Agency, the Directorate o f Public Procurement in the Ministry o f Economy, Finance, and Development Cooperation (MEFDC), and procurement units within each procuring entity, in time for these new entitles to participate in the budget preparation for 2009. Main issues and challenges 4.88 The Government has achieved an important milestone in enacting the new Procurement Code; however, the real challenge will be in its implementation, which i s the responsibility o f a steering committee put in place for that purpose o n M a r c h 19, 2008. Implementation o f the procurement reform i s a massive challenge in a country where the capacity o f civil service and public enterprises has been weakened by more than 12 years o f conflict and c i v i l unrest. Moreover, the country institutions remain fragile, and adoption and implementation o f reforms are often delayed or undermined by political interference. 4.89 Successful implementation o f the procurement reform will require Government commitment, pragmatism, and an ambitious capacity-building program. T o that end, the Government should: (i) publish the implementing measures o f the new Procurement Code, and establish the new institutions; (ii) adopt and disseminate modem tools (standard bidding documents, guidelines, manuals, procurement filing system, e-procurement) based on the new Procurement Code; and (iii) adopt and launch the implementation o f a robust procurement capacity-building program. 106 4.90 Publish the implementing measures o f the Procurement Code and establish the new institutions. T o adopt and publish the legal texts creating the Procurement Regulatory Agency (PRA), the Directorate o f Public Procurement in MEFDC and procurement u n i t s within line ministries and public enterprises i s the easy part o f this challenge; nevertheless, i t needs to be done in a timely manner so that procurement activities scheduled for 2009 are not delayed. 4.91 T h e drafts o f legal texts are available, and need to be reviewed and adopted by the Government. The P R A will play a key role in the implementation and monitoring o f the reform. It will be critical that the PRA have a small team o f dedicated and qualified staff, since the credibility and authority o f the agency will be a function o f i t s effectiveness in tackling the first issues brought to i t s attention. 4.92 Adopt and disseminate modern tools. The main change promoted by the procurement reform i s that more responsibility will be given to line ministries, public enterprises, and others public entities. The procuring entities should be prepared to fulfill their new responsibilities. In a country with a shortage o f staff with adequate s k i l l s and experience, it i s essential to make available tools such as standard bidding documents, guidelines and manuals that will assist the procuring entities in carrying out their duties. 4.93 Priority should be given to developing capacity for critical activities and procedures such as procurement planning, drafting o f technical specifications, drafting o f evaluation report, and procurement filing. Manuals explaining the basic principles and providing models and samples would be developed and disseminated. W i t h regard to the procurement method, attention should be focused on national competitive bidding and shopping. In order to strengthen transparency, standard documents should also include models aimed at informing bidders and citizens about bidding opportunities and the award o f contracts. Drafts o f most o f the standard bidding documents, manuals, and guidelines are already available. However, the new actors in the procurement process should not be swamped with too many documents, or documents difficult to use. The challenge i s to provide tools that are easy to understand and user friendly. 4.94 Adopt and implement a capacity program. The strengthening o f the procurement capacity program should become a continuous activity. The setting o f priorities, and the sequencing and contents o f the capacity program are critical, particularly in a context where the needs are huge, and human and financial resources are limited. An assessment o f the needs has been carried out by a specialized consulting firm. According to this assessment, n o fewer than 1200 staff should be involved in procurement activities. An assessment o f the local training schools and institutes that could be involved in providing procurement training should be undertaken immediately as activities aimed at strengthening procurement capacity should be carried out in Burundi. * Recommendations for Procurement Reforms 4.95 Beginning in 2009, the Procurement Regulatory Agency should become the focal point for the implementation o f the procurement reform. I t s actions should be based on the findings and results o f independent procurement audits, along with prior and ex-post reviews conducted by the Directorate o f Public Procurement. The main recommendations are as follows: 107 Box 4.16: Procurement Reforms - Main Recommendations I n the short term: Establishment o f the Procurement Regulatory Agency, the Directorate o f Public Procurement in MEFDC, and procurement units in the main procuring entities. These actions should be completed by the end ofAugust 2008, so that staff of procurement units can receive basic procurement training before the Procurement Code becomes effective in October 2008; Selection o f local training schools and institutes to be involved in the design and delivery o f a comprehensive procurement training program should be carried out by December 2008, in order to allow the program to be launched in September 2009 (see below). I n the medium term: Adoption and dissemination o f afirst set of tools. The Government, through the steering committee, should select a first set of procurement tools. This set should include: ( i) procurement planning and guidance to establish procurement plans; (ii) standard bidding documents for national competitive bidding and shopping procedures; (iii) technical specifications for the 20 or the 25 goods that are most frequently procured; (iv) review and update, i f necessary, o f the technical specifications and work plans concerning schools and health centers; (v) guidelines for drafting bid evaluation reports; and (vi) guidelines for procurement filing; Adoption and implementation o f training program for the procurement units. Immediate priority should be given to training o f trainers and training staff in the use of the first set ofprocurement tools and standard procurement documents. J. H O W CAN THE MTEF BEINTRODUCED IN BURUNDI? Overview 4.96 The MTEF i s a tool for sector-based medium-term budget planning; it has been introduced in a wide range o f developing countries, often with the support o f donor agencies, which view public expenditure management reforms as critical to attaining the MDGs. The major objective o f the MTEF i s to facilitate policy-based budgeting (Le., linking budget allocations to strategic policy objectives) within a sound fiscal framework. 4.97 The idea o f introducing an MTEF in Burundi dates from the P E M reform agenda supported by IDA’S Economic Management Support Grant (ERSG) approved by the Board on August 1st, 2006.84More specifically, this P E M reform includes many components, including (a) the introduction o f simple and effective procedures for budget execution and control, (b) better monitoring o f public expenditure as a result o f the development o f reliable and complete financial information; and (c) improved methods o f forecasting and treasury management. 84 The second tranche o f the resources under the ERSG was disbursed in December 2007. The operation expired thereafter in March 2008. The successor operation (ERSG2), under preparation, continues with the PEM reform agenda that started under ERSG. 108 4.98 The purpose o f this section i s to discuss the steps for the gradual introduction o f the MTEF in Burundi over the next few years. While there i s a political will to implement an MTEF, there are major constraints, especially in relation to the existing institutional relationships o f the budget process and the very weak (almost nonexistent in some agencies) capacities for implementing modem budget process in the public service. 4.99 When considering the introduction o f an MTEF in Burundi, it i s worth turning to past experience in other African states facing similar condition^.^^ Lessons learned includ9 the importance o f initial conditions and building budget basics; the dangers o f failing to manage expectations; the impact o f continued lack o f comprehensiveness o f the budget on the effectiveness o f the MTEF approach; and the effects o f failing to ensure that top-down ceilings are maintained. Another important point raised by this study i s that, although well-functioning performance or program budgets would strengthen the MTEF process, this process i s in no way conditioned by the existence o f program budgets. T h i s lesson should be borne in mind in the Burundi context. 4.100 It i s therefore necessary to be realistic about the pace at which an MTEF can be introduced and to be cautious in implementing reforms, especially in relation to (a) aligning the introduction o f new budget processes to the building the technical capacities o f relevant agencies; and (b) promoting a broader understanding o f the rationale for the MTEF within the public service and civil society. A dogmatic approach to MTEF implementation should be avoided; there i s no blueprint for the MTEF which can be applied uniformly to every country. Instead, the introduction o f MTEF reforms must be done in a pragmatic matter, with each step in the reform process carefully designed to ensure that it i s compatible with the technical and institutional capacities necessary for i t s effective implementation. 4.101 This section explains the motivation for introducing the MTEF, along with a discussion o f a few critical constraints and obstacles to its introduction. I t also outlines a phased strategy for i t s introduction. Motivation for the Introduction of an MTEF 4.102 The primary motivation for introducing the MTEF i s to improve budget resource allocation and to link this with the strategic policy objectives o f the Government, including those set out in national development strategies such as the PRSP. 4.103 Rational budget allocations, in which expenditures are prioritized to best meet the budget’s objectives, are not possible under the existing budget system for many reasons, among which: (a) the budget preparation process remains inadequate despite recent improvements: the budget i s not based on a sound macroeconomic framework and i s prepared in a short timeframe, with the preparation o f the supplementary budget occurring in parallel with the preparation o f the finance law for the following year; (b) line ministries do not have adequate structures or staff for 85 - See: “A Review o f Experience in Implementing Medium Term ExpenditureFrameworks in a PRSP Context A Synthesis o f Eight Country Studies” (commissioned by the Africa Policy Department o f the Department for InternationalDevelopment (DFID) U.K. in collaboration with the European Commission DG Development), 2003: 109 the preparation o f sectoral budgets; (c) the budget i s not comprehensive: although i t covers the recurrent budget (BO), the capital expenditures budget (BEI), and the national road fund (FNR), it does not cover all the foreign-financed investment operations; (d) the budget execution processes are overly complex: a substantial amount o f public expenditures are executed using exceptional procedures, and are not included in budget execution reports, which reduces the relevance o f those reports. The length and multiplicity o f compliance and regulatory controls has acted as an incentive for Government officials to use exceptional procedures; and (e) the separate determination o f current and capital budgets: the BO and BE1 are not consolidated at the level o f line ministries. 4.104 Line ministries are the most appropriate institutions for allocating budget resources within the sector in line with strategic policy objectives. Line ministries have the responsibility for formulating sector policies, but under the existing budget system the line ministries have only zero influence over sectoral budget allocations because o f the reasons mentioned above (e.g. the timing for budget preparation does not allow adequate arbitrage o f sectoral budget allocations); hence policy formulation i s divorced from budget allocations. The most important change to the budget process, with the introduction o f the MTEF will bring about i s the centralization o f responsibility for formulating sector budgets in the relevant line ministry. 4.105 Although the main motivation for reforming the budget allocation system i s to improve the allocative efficiency o f the budget, it will also provide additional benefits, not least because by linking budget allocations more clearly to policy objectives in a transparent manner, it will encourage donors to channel more o f their aid directly to the Government budget in the form o f budget support. 4.106 The P E M reforms mentioned above cannot be implemented in Burundi without a realistic budget. T o achieve this goal, the introduction o f an MTEF should be envisaged immediately. Indeed, the ultimate goal o f these reforms will be the simultaneous production o f a series o f MTEFs across the line ministries, allowing the harmonization o f inter- and intra-sectoral budgetary choices with the objectives o f the national Poverty Reduction Strategy, approved by the Government in September 2006. In effect, the complete MTEF approved by the Council o f Ministers, and based on program budgets prepared by the line ministries, will identify programs and requiredresources to accomplish the strategic objectives o f the Government. 4.107 Today, budget management and programming in Burundi are s t i l l carried out according to traditional processes; that is, line items have no apparent underlying objectives or strategy. However, efforts are underway, notably in education and health, to develop more coherent sectoral strategies with associated action plans, and to define the budgetary implications in the form o f program budgets or sectoral MTEFs. These initiatives are very important and should be encouraged, although they involve some risks. The main risk i s the fact that the resulting strategiedactions pladsectoral MTEFs will be based neither on a realistic forecast o f resources nor on the authorities’ decision on the inter-sectoral allocation o f resources. These risks could be minimized by a careful approach to implementing the MTEF in an appropriate way. Moreover, human capacities (in general, and in budget planning in particular) are so weak in the country that a prudent approach should be adopted to facilitate the appropriation and the continuity o f such an important sector-based tool for medium-term budget planning. 110 A Strategy for the Introduction of the MTEF 4.108 The following three-step approach i s suggested for the introduction o f an MTEF in Burundi: On the one hand, it i s the responsibility o f the Government as a whole to: (a) define the overall financial constraints o f the country; and (b) make adequate inter-sectoral budgetary choices consistent with national priorities, as spelled out in the Poverty Reduction and Growth Strategy o f the country. On the other hand, it i s desirable that an increasing number o f line ministries deepen their reflections on sectoral strategies and translate these into budgets consistent with sectoral programs. Ultimately, the progressive integration o f inter-sectoral budgetary choices (approved by the Council o f Ministers), and intra-sectoral budgetary choices proposed under budgets (prepared by line ministries), will lead to the setup o f a modern MTEF-based system o f budget programming. Table 4.1: Broad Phases of the Development of an MTEF Process Pre-MTEF Primary MTEF Intermediate Advanced MTEF process MTEF process process Macro- fiscal Establishment Based on a simole S i m d e domestic Inclusive o f all framework (top- andlor framework; focus model; stronger resources; more down) strengthening o f on domestic macro/fiscal sophisticated model; basic legal and revenues analysis; inclusion alternative scenarios institutional o f external finance may inform parameters, (e.g., financial decision making including ensuring programming or that a modem RMSM-X) Setting of sector Organic Budget Overall policy Some analysis o f Deeper/ more ceilingsftop down) L a w i s in place priorities, little/no inter-sectoral comprehensive analysis o f these priorities incl. basic analysis o f inter- priorities or o f policy priorities and sectoral priorities cross-sectoral cross-cutting issues issues Sector expenditure Limited sector Intra-sectoral Costing o f activities strategies (bottom expenditure broadly prioritized and stricter UP) strategies strategic/ prioritization programme mechanisms inform framework in place intra-sectoral strategic expenditure framework. - Source: World Bank Tajikistan: Programmatic Public Expenditure Review (June 22,2007). 4.109 Experience has shown that it took a number o f years to have sectoral program budgets completely integrated in the central MTEF in countries which embarked on reforming their budget programming. Table 4.1 shows the various stages to follow for the introduction o f a typical MTEF. For a country such as Burundi, where economic and financial management was severely disrupted by more than a decade o f political and socio-economical crisis, with devastating consequences on its institutional capacities, the completion o f the above three-step approach will be relatively longer. In light o f Table 4.1, there i s no doubt Burundi i s s t i l l in the pre-MTEF/primary MTEF stage. 111 4.1 10 I n the short term, the most urgent needs are to: (a) define the steps o f the process; @) undertake the training o f officials from the ministries of Economy and Finance (including Planning), and line ministries in the new techniques o f budget planning involved in the MTEF; and (c) create in the country an institutional capacity that will facilitate the internalization/ownershipof the MTEF. I t i s in this spirit that the Government and the World Bank propose to hire an international expert to initiate the above process and to define a realistic timetable for the gradual introduction o f program budgets in sectors and in the sectoral MTEFs. 4.11 1 I n sum, experience shows that there i s no need to be too dogmatic about MTEFs; in practice, they can take a variety o f forms aimed at strengthening the link between Government policies and budgetary allocations over time. At the same time, building such a link can take a significant amount o f time, and the steps should be taken right away. However, stakeholders should be careful o f overly ambitious expectations. Often, the beginnings o f a framework can be put in place, and the substance for the framework improved over time. 4.1 12 The introduction o f an MTEF in Burundi will first require: (a) an initial assessment o f the current situation; the methods o f budget planning/programming at the Ministry o f Economy, Finance and Development Cooperation, and sectoral/line ministries; and existing capacity in these institutions. Recommendations will follow for the next steps o f the process; and (b) the definition o f a realistic timetable and action plan for the preparation o f sectoral strategies and budgets for their eventual integration into a MTEF covering all Government spending, and for capacity building in relevant ministries and agencies. 112 CHAPTER 5. INCENTIVES AND PROFESSIONALIZATION IN PUBLIC ADMINISTRATION 5.1 A basic hypothesis underlying this PEMFAR study i s that while many of the public expenditure management and public expenditure allocation issues have already been studied and reforms agreed upon, most o f those reforms have not be carried out. Thus the issue i s one o f implementation, to which there are typically two types o f impediments - political will and capacity. The political will for reform i s a necessary condition for even beginning a public expenditure management study such as this one. This leaves the capacity o f the public administration, whose job it i s to implement reforms, as the principal constraint. Since “capacity” i s not just a matter o f skills but also o f incentives, which provide a motivation to act, a study o f incentives i s a proper part o f any PEMFAR. Similarly, the more professional the public administration is, in terms o f the qualifications o f staff, the way in which they are managed, and the quality o f advice given to the political leadership, the more likely the most appropriate reforms will be selected and implemented. 5.2 This chapter will examine the incentives paid to staff and the professionalization o f the public administration. Staffing costs, generally referred to as the wage bill, are covered separately in Chapter 2. Section A, below, discusses the issue o f staff incentives in the civil service, police, and military, whose salaries and allowances make up the salaries line item in the national budget; it does not cover staff in parastatals. The professionalization o f the public administration i s discussed in Section B. Section C concludes by proposing reform priorities. The key issues can be summarized as follows: Average salaries in the civil and uniformed services are high relative to per capita income; For some key skills, such as doctors, information technology (IT) professionals, and nurses, salaries are too l o w to attract and retain staff; Substantial restructuring and reengineering will be required if the delivery o f services i s to be improved and at the same time the wage bill reduced; The politicization o f appointments to top positions i s one of the main weaknesses o f the Burundian civil service; and Staff appraisal methods do not provide a sound basis for rewarding good performance. A. INCENTIVES 5.3 Although, on average, public service salaries are high relative to per capita income, they are low in absolute terms. Basic salaries range from less than US$14 a month for troops to around US$451 for a very few at the top o f the pay scale.86The average salary in the highest grade o f the highest category i s only about US$290 a month. Average salaries by category and grade for the civil service, teachers, military, police, and magistrates are shown in Table 5.1 and Annex 3. The analysis on salaries is limited to the civil service, military, and certain categories o f civil servants. A number o f public sector workers are paid more, including magistrates and employees in anti-corruption institutions. 113 Table 5.1: Average Salary by Category, M a y 2007 Monthly base Average monthly salary expenditure Category Numbers base salary (FBu) (FBu millions) Police* Agent 15,101 31,814 480 Brigadier 2,427 55,794 135 Officier 1,098 129,279 142 Total 18,626 758 Average 40,684 Militaire Troupes 21,829 26,914 587 Sous-officier 4,647 48,818 227 Officiers 1,632 121,441 198 Total 28,108 1,013 Average 36,023 Non-teachers A (Collaboration) 2,817 39,499 111 C (Direction) 972 6 1,825 60 E (Execution) 2,765 34,332 95 Indeterminate 6,691 42,863 287 Sous-contrats 7,834 nla nla Total 2 1,079 553 Average 41,758 Teachers A (Collaboration) 32,171 50,107 1,679 C (Direction) 1,218 89,268 113 E (Execution) 1,517 30,301 46 Sous-contrats 2,844 nla nla Total 37,750 1,838 Average 52,656 TOTAL STAFF 105.563 Note: The data for the police are for September 2007. Military and police staff numbers exclude civilians. Average salaries exclude staff that are kous-contruf ’ (contract workers) 5.4 Public service salaries are high compared with incomes elsewhere in the economy. The average sub-Saharan African civil servant’s salary i s five times the per capita income o f the country concerned. The average Burundi c i v i l servant’s basic salary (excluding allowances) i s more than four times the per capita income. This i s why it i s not difficult to recruit staff into the Burundi c i v i l service, at least to fill most positions. Teachers are paid more than six times the per capita income; the average for Africa i s 4.6 times.” 5.5 Some groups o f public servants are better off than others. Among the best-paid group o f public servants are magistrates, personnel o f the Inspection Gdndrale de I ’Etat, the Cour des Cornptes, and the Brigade Spdciale Anti-Corruption et Bonne Gouvernance.88After this category, the top grades in the police and military are the best paid, followed by teachers and finally the FP, whose basic salary for i t s top grade i s more than three times as high as the basic salary for the lowest grade in the FP. Teachers in the top grade are paid 45.6 percent more, on *’The calculations on average teacher and civil service salaries include only the basic salary and exclude allowances. Also excluded in the calculations are the salaries o f teacher and non-teacher staff who are contract workers. 88Political personnel in the Presidency, Vice-presidency at the General Secretariat o f the Government do not appear in the pay lists and have their own budget. Their salaries are not included in the analysis. 114 average, than c i v i l servants in the top grade o f the FP. The bottom grades are paid about the same in all four groups. 5.6 Average salaries have not kept pace with inflation. Between 1989 and 2001, while salaries increased by 43 percent (there were three adjustments to the salary index), inflation rose by 374 percent. Since then the index has been increased from 134 to 154, although there have been n o rises in this decade apart from 2001 and 2006. 5.7 Allowances are not as high as in some countries. Burundi’s allowances are not among the highest in Africa, and they are transparent. Allowances in general add about 28 percent to the basic c i v i l service salary; the housing allowance alone adds 25 percent to the basic salary. Allowances make up a higher proportion o f gross military salaries, especially for the lower ranks. For example, the take-home pay o f the troops i s roughly double the basic salary. Officers receive allowances depending o n their jobs: pilots and doctors each receive an allowance o f FBu 20,000 each month, and an engineer receives FBu 10,000. Allowances exceed the basic salaries in some countries, especially for the more senior staff. For example, allowances are many multiples o f basic salaries for the top grades in Zanzibar. Some countries are n o w folding most allowances into salaries, as during the first phase o f the pay reform in Tanzania. 5.8 Salary increases are across the board and do not differentiate among s k i l l categories. When salaries are increased by the Government, the same percentage increase i s given to all staff. Increases are not differentiated by scarcity o f skills and experience. The only staff who tend to be treated differently are IT staff, some doctors, and others with particularly scarce skills. These are covered under the sous contrut category or are paid by donors. Sous contrut includes both the poorest and best-paid public sector employees, reflecting the fact that these employees are among the lowest and highest qualified. In almost every country that has reformed i t s pay structure, higher increases in salary have been given to what are often referred to as key technical and professional staff. This has been the case in Zambia, Kenya, Tanzania, Uganda, and Mozambique, for example. Underlying this policy i s the need to decompress salaries; that i s to say, to widen the gap between the top and the bottom. 5.9 The compression ratio in Burundi’s public service - the average salary in the top grade divided by the average salary in the bottom grade - i s 6:1, which i s quite low by international standards. The ratio o f the highest to the lowest i s about 18:l. Mozambique aimed to increase i t s compression ratio to 20: 1 by giving much higher increases in salaries to key technical and professional and managerial staff. Some countries may have gone too far in this direction: both Kenya and M a l a w i n o w have extremely high compression ratios, although these ratios may be somewhat misleading. The compression ratio may also be misleading in Burundi’s case. 5.10 The main problem i s not so much that the top grades are not paid enough, but that those with certain key skills are paid less than they can earn elsewhere. I n many countries, managerial staff members, such as permanent secretaries or directors general, are quite w e l l paid, but the top technical staff on whom they rely are not. This i s why the focus o f pay reform i s often o n key technical and professional staff rather than on managers. Facing a similar situation, Rwanda adopted a policy o f training technical staff abroad, in order to overcome the shortage o f 115 key staff. Pay reform, as w e l l as increased training for key technical and professional staff, especially in IT and medicine, are necessary in Burundi. 5.1 1 There i s a large difference between the salaries paid for public servants sous-stutut and those who are not governed by the statut des fonctionnaires. The category o f well-paid staff includes the police, military, and also magistrates and staff from IGE and the Audit Court, among others. 5.12 The need to pay higher salaries to key technical and professional staff will grow. W e were told by some staff we interviewed that there i s n o need to differentiate among different skills, since the labor market outside the public sector i s so weak that the civil service has n o problems in filling vacancies. Indeed, lists o f those to be interviewed for vacant positions can be seen o n the walls o f the MinistBre de la Fonction Publique, du Travail, et de la Securite Sociale (MFP, Ministry of Public Administration, Labor and Social Security): most lists have ten or so names. However, there i s almost certainly a salary issue for the kinds o f staff that are, and will continue to be, needed to lead the reform efforts. Increasing salaries above public service levels for staff in public service positions i s going o n right now. Some donors pay doctors up to ten times as much as the average monthly salary o f US$60 paid by the Government to medical doctors, and some o f these medical doctors are foreigners from countries such as Egypt, Cuba, and Nigeria. Teachers, too, can receive top-ups to their salaries from parent groups and NGOs, although donor support, from countries such as Belgium and France, i s usually in the form o f paying o f f the arrears in teachers’ pay. Some donors are paying the salaries o f senior officials at levels as high as four times the Government’s rate.” 5.13 Some counties are now introducing what i s called a senior executive service (SES) as a way to pay the salaries needed to attract, motivate, and retain s t a f f with higher than civil service salary levels (SESs are common in developed countries). The top 140 or so staff in Sierra Leone’s c i v i l service are becoming members o f its SES, with salaries well above the top rates for c i v i l servants. Liberia i s targeting top professional staff rather than the heads o f departments and managers for its SES. One o f the principal objectives o f Sierra Leone’s SES i s to eliminate the need for project implementation units (PIUs), which have proliferated there. Thus the staff n o w in PlUs would be folded into the civil service but paid the same salaries as they are being paid in PIUs. Donors, including the W o r l d Bank in the case o f Liberia, have agreed t o provide financial support for the establishment o f both o f these SESs. The salary issue i s going to become more serious as the economy and, in particular, the private sector develops. Within three years o f the end o f Sierra Leone’s conflict, professionals in the parastatals, NGO, donor, and private sector were being paid five to ten times as much as those in the public sector. Further, there i s a class o f employees whom i t w i l l prove difficult to attract: those who have opportunities in other countries. Doctors, lawyers, accountants, engineers, economists, and other well-qualified professionals can seek work in other developing and developed countries, where salaries are many multiples o f those in Burundi. For example, doctors in neighboring Rwanda’s public service are paid more than ten times as much as doctors in Burundi’s. Conflicts exacerbate this problem by driving out professionals who then become accustomed to higher salaries in the diaspora. 116 B. PROFESSIONALIZATION 5.14 The politicization o f appointments to top positions, which i s common in sub- Saharan Africa, has weakened leadership and management effectiveness. In Burundi, members o f the Minster’s Cabinet, directors-general, and directors are all appointed by the President. The politicization o f appointments i s particularly common in post-conflict countries, where the Government has to be sure o f having loyalists they can count o n at the top o f the public administration, and presidents have to reward key constituents in the peace process. The cost associated with this politicization i s that advisors are unlikely to “speak truth to power” when they are dependent o n those in power for their jobs. The effectiveness o f the public service depends partly o n h o w well qualified staff are to do their jobs and h o w w e l l they are managed. In addition, managers may not have the skills and experience needed to implement the agreed policies. B o t h factors tend to reduce the confidence staff have in their managers, a problem exacerbated by the rapid turnover o f people in these key positions. 5.15 Some officials have stated that two years into this Government’s administration,the time has come to professionalize the top o f the public administration. A proposal has been prepared to appoint a professional civil servant, a secretary-general immediately below the minister, to head the public administration. The minister’s cabinet would continue to provide advice o f a policy and political nature.” Under the current proposal, the secretary-general would be appointed f r o m within the public administration, although the appointment would be made by the President. The implementation o f this proposal would be a big step in the direction o f establishing a more professional c i v i l service. The Ministry o f Foreign Affairs already has a secretary-general, and it would be worthwhile to assess this experience before considering extending it t o other ministries. 5.16 Appointments within the public service are generally based on qualifications. Civil servants entering the lowest grade o f the civil service have to have four years o f secondary education, while those entering two grades above need a university education. This reflects the fact that the point o f entry in the grading structure i s determined by the level o f education. The police and military have similar requirements. There is, however, some anecdotal evidence suggesting that some staff are hired o n the basis o f family and political connections. 5.17 Training budgets are low. There i s little scope for improving the quality o f staff through training funded from the national budget. There has been n o budget for training in the c i v i l service as a whole for some years, apart from orientation training for new entrants. However, some training i s funded by the donors through their projects. 5.18 The Minist2re de fa Fonction Pubfique has limited responsibilities in the management o f human resources. It i s responsible for managing human resources o f the public administration. The Director General for Public Administration supervises four directorates: (a) management o f careers; (b) management o f salaries; (c) recruitment and control o f staffing; and T h i s separation of responsibilities differs from Westminster-style governments, such as in the Anglophone countries in sub-Saharan Africa, where the secretary-general (permanent secretary) would be both the head of administration and the principal policy advisor to the minister. 117 (d) regional offices. T h i s Director General also serves on the Recruitment Commission. Although the Ministry has had input into the redesign o f the salary grading system, i t has had n o input into overall salary policy, including the proposed 34 percent increase in benefits for 2007, which was finally shelved and postponed to 2008. 5.19 M F P has delegated partial responsibility for the management o f human resources to the Ministry o f Education and the Ministry of Health. All other staffing in the public administration has been frozen. Staff selected by ministries should, according to the law, be ratified by the Recruitment Commission that was created in 2007. However, this Commission does not meet, although there are plans for it to do so in the near future. The management o f staff by the ministries to which responsibility has been devolved i s suffering from a series o f problems. There i s evidence o f staff who are hired but not added to the payroll and wait months or even a year to be paid.g' Likewise, staff who retire, die, or move to another j o b are not removed from the payroll. M F P believes that there are some 2,000 ghosts in education for these reasons (see Chapter 2 for more details). The results o f the 2008 census o f civil servants should help to eliminate this problem. M F P would like to re-centralize the management o f civil service human resources. 5.20 Where they exist, human resources departments in the ministries are not as well organized as they could be?2For example, the Health Ministry's human resources department i s divided into three sections, dealing with statutory staff, contract staff, and records, respectively. Proposals have been made to manage the staff in a more functional manner, with sections responsible for planning, training, pay, personnel relations, security, and health in the workplace. In many countries, the equivalent o f the M F P would be responsible for managing human resources overall, including setting human resources policy, with the selection o f staff being made by the ministries and ratified by a public service commission (PSC), which would also set the standards for recruitment, promotions, discipline, and dismissals. The Recruitment Commission has not been operational, since n o staff have been recruited other than new teachers. I t would be important for authorities to ensure that the commission first becomes operational and assumes, in the medium and long term, the functions o f a real public service commission. For instance, in Kenya, the functions o f the PSC include: (a) recruitment for the civil service and the local authorities; (b) promotions and acting appointments; (c) disciplinary matters o f the c i v i l service and local authorities; and (d) civil service examinations and occupation tests. M F P would work with the MEFDC on salary and wage bill issues, and be responsible for training policy and institutions. In the context o f the proposed Cadres Organiques reforms, M F P believes that with the introduction o f a fully functioning integrated human resource management system, it could be beneficial to decentralize some o f these human resource management functions to the line ministries, while retaining overall control. ''These delays arise, in part, because new staff need to produce a dossier administrat$ and the verification o f relevant documentation can take several months. 92A human resources advisor exists in all ministries that do not have a human resources department, which i s the case for many ministries. 118 5.21 The way in which the staff appraisal system functions does not provide a sound basis for rewarding good performance (See Box 5.1). T h e staff appraisal system c o u l d b e the basis for building a high-quality p u b l i c administration, since it w o u l d encourage the best performers to rise m o r e quickly. E a c h year staff members are graded Clite, tr& bon, bon o r insufflsunt. Those graded &lite m o v e up three echelons in their grade, those t r b bon two, and bon one, w h i l e a succession o f insufflsunt grades m a y result in dismissal. However, 75 percent o f the staff are given a trks bon r a t i n g and only 2 percent insufflsunt, reflecting a bureaucratic culture that does not l i k e to penalize poor performance. In addition, p o l i t i c a l appointees are automatically graded as Clite. Box 5.1: Grading Structure in the Civil Service Progression up the grading structure only slowly rewards good performance. The civil service salary grading structure (Statut de 1998) includes many steps, and the progression up the scale i s systematic and slow. Staff enters at one o f the three “categories” (see tables 1 and 2 in Annex 3). University graduates enter at the bottom o f C: Direction; those who have completed secondary education and three years o f college enter category A: Collaboration; those with at least four years o f secondary education j o i n category E: Execution. Each category has three grades (there are four for teachers). Each grade has a number o f echelons, the total number o f echelons being 140. If someone i s graded t r i s bon, he or she moves up two echelons; if elite, they move up three echelons. If someone receives a trks bon score for each o f the six years, he or she would move up a grade. Well over half the staff receives a tris bon score for their annual appraisal. Those receiving lower scores would take longer to obtain a grade promotion. Staff can move up categories only if they achieve higher educational qualifications. An indice i s associated with each echelon; the higher the echelon and the higher the grade and category, the higher the score. Salaries are determined by multiplying an overall salary index (point d ’indice), which stands at 154.1, by the categorylgradelechelon score for the person concerned. To take an example, if a person i s i n category collaboration, grade 11, echelon 14, the echelon indice i s 286, so the salary w i l l be 154.1 times 286 or FBu 44,072 a month as a basic salary; if the person were higher up i n the same category, say at grade I11echelon 12, that person would receive 154.1 times 427 or FBu 65,800. I t would take 20 years o f t r i s bon assessments to rise f i o m the bottom to the top o f a category. During that time, without the overall salary indice being increased (and it i s usually only increased to take account o f inflation, and imperfectly at that), the person’s salary would increase by about four times. Progression up the grading structure i s similar for other public services, although the categories and grades are different. The range o f salaries i s similar too. A new grading structure (Statut de 2007) with eight grades and 146 echelons was approved in 2006 but has yet to be implemented. 5.22 L o w funding o f other recurrent expenses makes staff less effective. S t a f f members do not have the tools they need: f e w computers can lead to manual record keeping, and there are insufficient h n d s for fuel and the maintenance o f equipment, among other deficiencies. As discussed in Chapter 2, this p r o b l e m i s at the heart o f the wage bill issue. T h e high and r i s i n g wage bill i s c r o w d i n g out the funds needed to enable staff members to b e productive. 5.23 There may b e a need to restructure the civil service to improve i t s effectiveness and efficiency. M a n y believe that ministries - o r even the public administration as a w h o l e - needs to b e restructured. In other countries, responsibility for leading a restructuring lies with the MFP or i t s equivalent. So-called functional reviews are among the m o r e popular r e f o r m activities in developing countries. T h i s i s a multi-step process that aims to restructure the p u b l i c 119 administration so that it can be more effective (Le., deliver the required results) and efficient (i.e., deliver the results with the least possible amount o f human and financial resources): Redefine the role o f the central public administration v i s - h i s local government, parastatals, and the private sector very often, to focus o n policy formulation and monitoring the implementation o f policy and evaluating i t s impact. Define medium-term expenditure priorities in accordance with national plans and the PRSP. For each ministry, clarify its mission, core, and non-core activities in line with the above priorities. Identify overlapping responsibilities and opportunities to merge ministries and agencies. Examine ministerial structures in line with their missions and core and non-core responsibilities. Redefine structures in each ministry. Redefine staffing in accordance with new structures. Prepare j o b descriptions. Match staff t o jobs. H i r e and reallocate staff members accordingly. Re-engineering of some key public services focused on service delivery could bring ~ high returns at little cost. Most public service reform programs have a “quick wins” component. Ethiopia’s reform program i s built around a combination o f decentralization o f service delivery and process re-engineering; the results o f this program have been impressive, with huge reductions in the time taken to register businesses, carry out land transactions, import and export, and even get married. Burundi could follow the example o f other countries and very quickly and cheaply achieve improvements in public services simply by re-engineering processes at the interface between citizens and businesses and public servants. C. REFORM PRIORITIES 5.25 Short-term priorities. At this stage, the most urgent measures are to: (a) ensure that the wage bill over 2008-2010 does not rise as a percentage o f GDP by eliminating ghost workers, reviewing non-wage benefits, and ensuring there are n o new across-the-board wage increases in the public administration in the revised budget; and (b) put in place institutional arrangements to bring and keep the wage bill under control. T h i s second priority would require: 0 The MEFDC and MFP to work more closely together o n salary and staffing issues. a Establishment o f clear staffing targets for line ministries that are in line with a lower wage bill. a MFP to monitor implementation o f agreed staffing ceilings. a Setting up a human resources management information system (covered under PAGE). 120 0 Ensuring that the zero recruitment policy i s followed, except for education, health, and justice. 0 Consideration o f the possibility o f adopting a partial replacement policy, so that one out o f every two or three retiring public servants i s replaced, instead o f the current policy o f replacing all departing staff member^.'^ 5.26 As noted in Chapter 2, the Government i s undertaking a series of important measures to improve wage bill management in 2008. This includes (a) using the 2008 census to create a new payroll list; (b) the development o f an integrated human resources management system, with an interface with SIGEFI, which i s scheduled t o be ready by June 2008; and (c) Cadres Organiques; i.e., the development o f a medium-term staffing plan and an analysis o f public administration structures, with the a i m o f improving human resource management and service delivery. All three reforms are covered under the W o r l d Bank's PAGE. 5.27 Medium-termpriorities. These include the need to review (a) the structure o f the public service; (b) salary levels throughout the public service; and (c) the grading structure and the basis for appointments and promotions; and (d) to restore the training budget. B o x 5.2 spells out these recommendations in more detail. 5.28 The first step i s to put in place a joint M F P N E F D C reform team to prepare and ensure the implementation o f this program o f reform and capacity building. The loss o f control over staffing and the wage bill means that MEFDC and MFP have to work together on the reforms and not run separate reform programs. This team should include staff members o f the very highest caliber, who would work full-time on the reforms. 5.29 The Cadres Organiques o f the MFP intends to address a number o f these medium-term priorities. I t aims to: (a) review the structure and organization o f public service; (b) define the optimal number o f staff members for each structure; and (c) implement a medium-term plan o f staffing needs. Other important priorities include: harmonization o f salaries in the public service; revision o f the promotion structure and the basis for appointments and promotions; creation o f human resource departments within each ministry; funding for technical training; analysis o f different options for encouraging voluntary early retirement (plan de reconversion); establishment o f an office o f pensions and professional risks for civil servants (ONPR); redynamization and strengthening o f the administrative bureau o f the MFP in charge o f c i v i l service reform (Bureau pour 1'Amdlioration des Services de 1 'Administration Publique, ASAP), by providing i t with sufficient funding, technical staff members, and technical assistance. T h i s policy i s currently under discussion in France, where it i s planned that one out o f every three public servants 93 will be replaced, instead o f the current policy o f replacing all departing staff. 121 Box 5.2: Medium-TermPriorities Review the structure of the uublic service 0 I s the balance between a small civil service and relatively large numbers of teachers eficient and effective? 0 Are teachers and health workers in the right number in the rightplaces? Begin a process o f functional reviews of all ministries. Review salarv levels throuphout the uublic service Are the relative salaries among the four services (civil service, teachers, military, and police), with the police best paid and the civil service worst at the top, appropriate? 0 Which professional skill groups are most dificult to recruit, motivate, and retain? 0 Could some allowances be folded into basic salaries? Are all allowances necessary? Review the pradina structure and the basis for amointments and uromotions Permit performing staff members to move up more quickly. Put in place a more meaningfir1 and effective staff appraisal system based on performance as well as skills. 0 Reduce the incidence ofpolitical appointments at the top. Convert the recruitment commission into a public service commission, acting as the custodian of a professionalized civil service. Restore the trainina budpet 0 Restore budgetary support for in-service training. 0 Build in-country training capacity. 122 123 94 95 Annex 1: Proposed Schedule for the Preparation of the Budget Law Steps Deadline Responsibility Observation 1. Presentation o f the macroeconomic March DGP, DBC, BRB, D B C framework and the tax framework 2. Preparation o f the Minister in charge April MEFDC o f Finance (MF)’s circular, consistent with the macroframework and the tax frame work 3. Distribution o f the circular with the April MEFDC budget outlines (Guidelines, and expenditures limits) 4. Preparation o f the draft budgets May-June Budget offices, projects and programs managers, revenue departments 5. Budget conferenceshegotiations July MEFDC, DGP, MFP, Public Expenditure with line ministries DGBCP, DGR, Budget Review inputs are divisions, project and introduced program managers 6. Preparation and submission o f the August DBC IMF program preliminary draft o f the budget law to negotiations, and the office o f the MF World Bank feedback on the drafi budget 7. Presentation o f the preliminary draft September MEFDC o f the budget law to the Council o f Ministers 8. Draft budget law sent to Parliament October MEFDC 9. Consideration and approval o f the October - National Assembly and budget law December Senate 10. Enactment o f the budget law December President 94 Frp, “La rbforme de la gestion budghtaire D, May 2007, by Jean-Luc Helis et al. 95 For a fiscal year o f January 1 to December 3 1. 124 Annex 2: Cash Management and MTEF Proposals - Annex 2 A: Cash Management Proposals for an Institutional Set-up and Responsibilities The following proposals were discussed with the authorities, and approved by the MF. They should help for the preparation, implementation, monitoring and updating o f a cash management plan used as a tool for the management o f the budget. The Cash Management Committee (CMC) Under the direct authority o f the MF: 0 Members: Governor BRB; MEFDC; DGR; DGBCP; DGMP; the Coordinator o f the “Cellule d ’appui“; OTBU; the Chief o f the Permanent Secretariat o f the GGTE. 0 Mission: o Review/analyze the implementation report by GGTE o f the cash management o f the month past or ending o Review/analyze and adopt the starting revolving quarterly cash management plan presented by the GGTE with the first month detailed by week o Using the report, analyses and proposals o f the GGTE, make the trade-off decisions required by the situation and notify the decisions (e.g., revised commitment ceilings) to all budget users concerned (ministries; other public institutions) o Adopt the minutes o f each meeting at the end o f the meeting 0 The CMC assisted by the Permanent Technical Secretariat (STP) o f the GGTE, located in the office o f the Cellule d ’appui. 0 Meetings o f the CMC: at least once a month, at the end o f the month or the first days o f the next month for a past month. The Cash Management Grouu (GGTE) 0 Members: o Permanent members: the 3 members o f the permanent technical secretariat o f the GGTE o Non-permanent members: OTBU, state’s cashier at the BRB; the staff in charge o f statistics for the customs; taxes; administrative revenues; the staff in charge o f external debt and the HIPC program at the Directorate o f “Trborerie”; the staff in charge o f BO in the Directorate o f Budget (total o f 11) 0 Mission: o Facilitate the timely availability o f the relevant data to the STP o Review and finalization o f the documentation (tables, analysis, proposals) prepared by the STP for the week ending; for the current month 125 0 Review, and finalization o f the documentation (projections, targets, adjustments) for the next week, or month, or revolving quarter o Analysis o f performance, slippages, and proposals for adjustment, trade-off to the CMC 0 Meetings: o Once a week, at the end o f the week Chaired by the Coordinator o f the Cellule d’appui, assisted by the DGBCP for an interim period and later on by the DGBCP. The Permanent Technical Secretariat (STPI 0 Members: o The coordinator o T w o assistants 0 Requirement to be a member: o Be appointed by the MF following recommendation by the DGR and the DGBCP o Work exclusively for the STP with no other assignment 0 Mission: 0 The STP o f the GGTE assist the GGTE and the C M C 0 Collect the data needed to prepare and monitor the implementation o f the Cash Management plan 0 Prepare, monitor and update the Cash Management Plans (CMP): provisional C P M for the FY; revolving quarterly CMP; monthly CMP; weekly CMP 0 Prepare the tables for the implementations results for a past period (week; month; quarter); and analyze the results 0 Prepare, adjust the CMP for the new period (week, current month, current revolving quarter) 0 Analyze the results/performances, and their impact on future periods; and prepare recommendations (adjustment, trade-off, new commitment ceilings.. .) for the GGTE and the C M C 0 Finalize the analysis and recommendations discussed by the GGTE for presentation to the C M C Location: The STP i s located in the premises o f the ‘‘Cellule d ’uppui” 0 Assistance to the STP: The STP and the GGTE are assisted by the French Technical Assistance project to the MEFDC, and the French Advisor to the MEFDC. The assistance i s in the form of: bonus linked to results, office equipment and operating costs, training, and also advice, comments. Substitution for the preparation and monitoring o f the CMP i s forbidden. I t should be clear, from the beginning, that it i s the responsibility 126 o f Burundians to prepare, monitor, analyze the results and developments, make proposals and decisions to update the CMP, make the necessary adjustments and trade-off. Directorate in charge o f Cash Management planning, The present “Direction de la Trborerie ” i s mostly concerned with external and internal debt management. The responsibility for cash management planning should be transferred to the “Direction de la Cornptabilite‘“ where O T B U i s located. 127 Annex 2 B. Introduction o f an MTEF in Burundi: proposed role o f the internationalexpert The role o f the international expert will be to evaluate the approach o f sectoral budget programming currently utilized by the Burundian authorities and to define an action plan with a realistic timetable with a view to: (i)gradually improving intersectoral budgetary choices; (ii) optimizing the composition o f public expenditures in line ministries in the context o f sectoral budgets; ( iii)strengthening the capacities o f ministries and institutions involved; and (iv) launching the process o f MTEF with that a i m o f achieving this ultimate goal in to achieve this goal within 4-6 years. Progressive improvement of intersectoral budgetary choices: The key elements o f a more strategic budget planning/programming are the following: (a) the introduction o f the multi-year concept in programming revenue and expenditure; (b) the realistic multi-year projections o f government resources (tax and non-tax revenue, and external resources), based o n a suitable macroeconomic framework and a detailed analysis o f the programs o f major donors; (c) the identification o f basic scenarios (low, base, high) that demonstrates budget choices according to alternative strategies; (d) the identification o f rigidities (acquired rights, irreversible decisions, etc.), which limit budgetary choices. These include for instance, pre-decided quasi-fixed expenditures (e.g. wages and salaries), and ongoing programs and projects that cannot be interrupted; and (e) an initial test o f the allocation o f available and projected Government resources in line with the poverty reduction strategy. nthis context, the proposed international expert will: I Assess the methods employed in the ministries o f Finance and Planning t o determine the growth prospects o f the economy (macroeconomic framework) and derive projections o f government revenues. Assess Government and donors forecasts and availability o f foreign resources (budget supports and projects), which represent nearly 50 percent o f resources available for the budget. Based on the above diagnosis, the proposed expert will make recommendations aimed at establishing the MTEF o n a more permanent basis. With respect to the sectoral allocation o f resources, the international expertise will, in line with the findings o f this PEMFAR report (specifically on the intra- and inter- sectoral allocation as well as on the PEM section), discuss with the ministries o f Finance and Planning some o f the main factors that can and must influence this allocation. These discussions and thoughts will ultimately enable the authorities to introduce multi-year objectives o f public expenditure for each o f the major sectors and ministries. Such objectives will guide the efforts o f these ministries and sectors to maximize the efficiency o f public expenditure in their respective sectors in the context o f program-budgets. n this area, the tasks o f the international Optimization o f intra-sectoral budgetary choices: I expert entail: Assess the progress made by the social sectors (education and health) in the definition o f their sectoral strategies, and the costing (budgetary implications) o f these strategies. On this basis, the expert will define with these departments a proposed timetable for the preparation o f sectoral MTEF or program-budgets for these two sectors. 128 Investigate the status o f sectoral strategies in other priority sectors (transport, energy and water, agriculture and environment), as well as examining the feasibility o f the progressive development o f sectoral MTEF / program-budgets by the associated ministries and agencies. Finally, on the basis o f all o f these discussions, the international expert will propose a comprehensive timetable for the development o f sectoral MTEF gradually covering the whole o f public expenditure, thus preparing the development o f central MTEF through the integration o f the budgets o f various sectors. Capacity Building : The greatly enhanced role o f line ministries in the budget process will not be possible without a major strengthening o f their capacities for budget planning; including capacities for costing expenditure proposals, linking activities to policy objectives, cost benefit appraisal o f projects, and the long term forecasting o f demand for services. Line ministries will need to be restructured to allocate many more staff to budget planning functions. They will also require substantial technical assistance in budget planning, especially in critical sectors such as the social sectors, and this should be a priority for the TA programs o f donors. Currently, the capabilities o f the ministries o f Finance, Planning and sectorayline ministries are extremely limited and largely absorbed by other priority activities. assess the capacity o f these institutions in budget In this context, the international expert will: (i) programming and financial management, as well as in elaborating sectoral and national strategies, and translating these into budget objectives; and (ii) recommend the technical assistance necessary to strengthen the capacity in these institutions and thereby facilitate the gradual improvement o f the budget planning process. 129 Annex 3: Wage bill and Salaries Annex 3. Table 1: Civil Service Base Salaries b y Category and Grade (monthly salary, as o f May 2007) Category Grade Moyenne Valeur Base salary (FBu) Numbers indiciaire d'indice Indeterminee I1 286 154 44,085 1308 I2 138 154 21,261 74 Average 42,863 A A1 202 154 31,182 1788 A2 328 154 50,582 885 A3 484 154 74,65 1 144 Average 39,499 c1 325 154 50,150 649 c2 52 1 154 80,3 17 28 1 c3 769 154 118,503 42 Average 61,825 El 135 154 20,810 2 E2 204 154 31,369 1981 E3 293 154 45,136 650 E indetermine 168 154 25,812 132 Average 34,332 Source: MFP. Annex 3. Table 2: Teachers' Base Salaries b y Category and Grade (monthly, as o f May 2007) Category Grade Moyenne Valeur Base Salary (FBu) Numbers indiciare d'indice A A indetermine 1 287 154 44,274 3689 A indetermine 2 246 154 37,893 5800 A2 324 154 49,864 18424 A3 5 24 154 80,733 3162 A4 719 154 110,838 1096 Average 50,107 C C indetermine 456 154 70,304 58 c2 53 1 154 8 1,896 860 c3 778 154 119,882 243 c4 1,122 154 172,973 57 Average 89,268 E E2 198 154 30,449 666 E3 28 1 154 43,334 110 E4 412 154 63,526 34 E indetermine 172 154 26,536 707 Average 30,301 Source: MFP. 130 Annex 3. Table 3: Military Base Salaries by Rank (monthly, as o f M a y 2007) Salary (Fbu) Numbers Troupes Recrue 3,785 0 2eme classe 16,113 250 lere classe 22,191 12,529 Caporal 30,429 6,270 Caporal chef 41,241 2,780 Moyenne 22,752 TOTAL 2 1,829 Sous-officiers Caporal Candidat 17,477 0 Sergent 3 1,639 1,500 l e r Sergent 40,682 1,364 l e r Sergent M a j o r 52,452 835 Adjudant 67,470 414 Adjudant chef 87,358 3 10 Adjutant major 112,044 224 Moyenne 58,446 TOTAL 4,647 Officiers Adjudant Candidat Officier 36,900 2 Sous-Lieutenant CO 38,413 382 Sous-Lieutenant Stagiaire 45,778 Sous-Lieutenant N o m m e 53,102 Lieutenant 68,245 746 Capitaine 89,281 27 1 Commandant 116,335 180 Major 150,557 160 Lieutenant-Colonel 194,315 131 Colonel 246,432 122 General de Brigade 313,504 14 General M a j o r 398,388 6 Lieutenant General 504,486 2 Moyenne 173,5 18 TOTAL 1,632 Civils Travailleur Civil 8,822 da Travailleur Civil A3 19,980 da Travailleur Civil Militarise Cat A 19,980 da Travailleur Civil Militarise Cat B 29,914 da Travailleur Civil Militarise Cat C 40,386 da Moyenne 27,565 da TOTAL 28,108 Source: M i n i s t r y o f Defense 131 Annex 3. Table 4: Police Base Salaries by Rank (Monthly, as o f May 2007) Rank Numbers Monthly base salary (Fbu) Agents AP3 0 16,133 AP2 0 19,252 AP1 3348 22,395 APP3 208 25,536 APP2 1790 28,677 APP 1 2669 31,818 APC3 4298 34,959 APC2 354 38,100 APC 1 2434 41,141 Sub-total 15,101 Average 28,668 Brigadiers BP3 11 3 1,639 BP2 503 36,160 BP 1 788 40,682 BPP3 134 46,567 BPP2 240 52,452 BPP 1 153 59,961 BPC3 151 67,470 BPC2 140 87,358 BPCl 307 112,044 Sub-total 2,427 Average 59,370 Officers CP 4 398,388 OPC 1 13 3 13,504 OPC2 158 246,432 OPC3 104 194,3 15 OPP 1 123 150,557 OPP2 133 116,315 OPP3 253 89,281 OP 1 232 68,245 OP2 77 60,673 OP3 1 53,102 Sub-total 1098 TOTAL 18,626 Average 169,08 1 Source: National Police 132 Annex 3. Table 5: Judiciary Salaries by Rank (monthly, as o f May 2007) Housing Function Grade Base (50%) Function Transport Risk Bonus Total Stand in Judge Court o f G14 33,396 16,698 0 20,000 15,000 30,000 115,094 Residence G13 42,026 21,013 0 20,000 15,000 30,000 128,039 Judge Court o f Residence Vice-president of the Court of G12 52,888 26,444 10,000 20,000 15,000 30,000 154,332 Residence President o f the Court o f G11 55,991 27,996 15,000 20,000 15,000 30,000 163,987 Residence Judge Court of Bankruptcy, G10 62,6 18 31,309 0 80,000 34,000 60,000 267,927 Substitute of the Public Prosecutor Adviser, Court of Appeal, First GO9 74,579 0 20,000 80,000 34,000 60,000 0 Substitute General, Court of Appeal Vice-president of the Court o f 80,000 34,000 60,000 268,579 GO8 88,825 37,290 25,000 Appeal, First Substitute General Court of Appeal President of the Court o f GO7 110,793 44,413 0 80,000 45,000 65,000 325,115 Appeal, Prosecutor General Court o f Appeal Adviser to the Supreme Court, GO6 132,956 55,397 30,000 80,000 45,000 65,000 345,206 Substitute General Supreme Court Vice-president of the Supreme GO5 158,285 66,478 35,000 80,000 45,000 65,000 408,353 Court, First Substitute General Supreme Court President of the Supreme GO4 196,520 150,000 45,000 80,000 50,000 70,000 449,763 Court, Prosecutor General of the Republic GO3 234,059 150,000 45,000 80,000 50,000 70,000 591,520 Out o f Category Stand in Judge Court o f GO2 288,768 150,000 50,000 80,000 50,000 70,000 629,059 Residence GO1 409,624 150,000 45,000 80,000 50,000 70,000 688,768 Judge Court o f Residence 133 Annex 4: Macroeconomic and Government Financial Operation Tables Annex 4. Table 1: Key Macroeconomic Indicators (in FBu billions, unless otherwise indicated), 2001-2006 2001 2002 2003 2004 2005 2006 Nominal GDP 1/ 550.0 584.6 644.7 748.6 861.0 986.6 Real GDP (1996=100) 277.1 289.5 285.9 299.8 302.4 318.0 GDP deflator (1996=100) 198.5 201.9 225.5 249.7 284.7 310.3 Real GDP growth (%) 2.1 4.5 -1.2 4.9 0.9 5.2 Population growth (%) 4.8 3.7 2.9 2.0 2.0 2.0 Industrial production index (1989=100) 83.4 94.0 93.0 91.8 95.2 Consumer price index (CPI) 40 1 395 438 474 537 55 1 Inflation (annual average) 9.3 -1.4 10.7 8.4 13.2 2.7 Budget deficit (commitment basis, %) -5.2 -1.3 -6.2 -5.2 -6.0 -2.1 Excluding grants -7.2 -5.6 -13.7 -19.2 -16.8 -20.4 Exchange-rate FBU/US$ (annual 830 93 1 1,083 1,101 1,082 1,029 average) External debt 888 1,174 1,371 1,458 1,238 1,313 Domestic debt 82.9 97.3 118.8 200.5 193.1 231.3 Debt service 32.4 34.0 49.7 75.1 74.0 93.6 External debt 17.9 20.2 32.2 46.8 35.4 59.1 Domestic debt 14.5 13.8 17.5 28.3 38.6 34.5 Debt service ratio (%) 57.0 62.5 66.1 73.3 47.0 53.3 Current-account deficit -29.1 -3.9 -8.1 -35.7 -10.0 -136.2 % o f GDP -5.3 -0.7 -1.3 -4.8 -1.2 -13.8 Overall balance o f payments -27.0 -10.0 -10.9 33.1 30.6 18.2 Yo o f GDP -4.9 -1.7 -1.7 4.4 3.6 1.8 Sources: Burundi authorities. 134 Annex 4. Table 2: Composition of Real GDP 1996 prices (in FBu billions), 2001-2006 2001 2002 2003 2004 2005 2006 Primary sector 141.3 152.6 139.2 146.0 136.8 143.7 Food crops 112.5 117.0 113.3 113.5 112.3 112.3 Export crops 10.6 16.9 7.5 12.3 3.3 9.1 Secondary sector 36.2 37.0 40.0 43.3 46.5 48.7 Agroindustry 9.1 9.1 10.1 10.0 10.6 11.0 Construction 8.9 9.2 10.5 12.9 14.5 15.7 Mines and energy 2.2 2.3 2.0 2.0 2.5 2.3 Services 78.9 81.3 85.4 90.4 98.3 104.4 Government 50.7 52.2 44.1 58.2 63.5 67.6 Transport & communications 10.6 11.0 11.4 12.1 13.1 14.0 Commerce 10.1 10.4 10.9 11.5 12.7 13.5 Other services 7.5 7.7 8.0 8.6 9.0 9.4 GDP at factor prices 256.4 270.9 264.6 279.7 28 1.6 296.8 Indirect taxes (minus subsidies) 20.7 18.6 21.3 20.1 20.8 21.2 GDP at market prices 277.1 289.5 285.9 299.8 302.4 318.0 Source: Ministry o f Planning. 1/ N o National Accounts have been produced since 1998, although the Ministry o f Planning provides estimates. Annex 4. Table 3: Composition of Real GDP by Expenditure 1996 prices (in FBu billions), 2001-2006 2001 2002 2003 2004 2005 2006 Consumption 592.9 627.3 698.0 802.0 963.8 1,093.2 Private 483.7 518.5 561.8 638.5 758.8 803.2 Government 109.2 108.8 136.2 163.5 205.0 290.0 Gross fixed capital formation 34.2 55.8 69.1 84.1 133.8 245.0 Changes in stock 0.0 3.0 0.0 0.2 0.0 0.0 Exports 37.7 36.0 48.7 60.2 76.1 121.5 Imports 114.9 137.8 171.1 198.0 312.6 473.5 GDP at market prices 549.9 584.2 644.7 748.5 861.0 986.2 (change in percent) Consumption 2.0 1.2 -3.3 3.9 -1.3 -2.0 Private 2.5 1.9 -5.8 2.8 2.0 -7.9 Government -0.9 -2.0 8.9 8.6 -14.4 19 Gross fixed capital formation -1.4 15.6 12.8 19.1 4.0 74.6 Changes in stock 0.0 0.5 0.0 0.0 0.0 0 Exports -4.8 -5.6 38.9 14.5 -0.9 13.7 Imports -4.8 -14 5.5 15.4 -5.7 11.4 GDP at market prices 2.1 4.5 -1.2 4.4 0.9 5.1 Source: Ministry o f Planning 1/ N o National Accounts have been produced since 1998, although the Ministry o f Planning provides estimates. 135 Annex 4. Table 4: Structure of the Public Debt (in FBu billions, end year), 2001-2006 2001 2002 2003 2004 2005 2006 Domestic debt 82.9 97.3 118.8 200.5 193.1 23 1.3 Treasury certificates 31.3 32.1 31.0 32.7 25.6 14.2 Central bank advances 42.5 54.5 81.1 121.7 150.4 208.5 Other 9.1 10.6 6.6 46.2 17.0 8.5 External debt 888.3 1,174.3 1,370.5 1,457.8 1,238.2 1,3 13.3 Direct 737.3 975.4 1,142.2 1,229.6 1,075.7 1,137.9 Indirect 150.9 198.8 228.3 228.2 162.5 175.4 Total 971.2 1,271.5 1,489.3 1,658.3 1,431.3 1,544.6 Source: Ministry o f Finance. Annex 4. Table 5: InternationalAid Commitments and Disbursements (in US$ millions), 2001-2005 2001 2002 2003 2004 2005 Disbursements ODA total (net) 137.2 171.5 227.4 361.5 365.0 Grants 145.0 152.2 203.1 288.3 328.4 Commitments ODA total 69.5 99.0 121.0 23 1.8 223.0 Grants 69.5 99.0 121.0 182.9 217.0 Disbursements (% o f commitments ODA total (net) 197.3 173.2 187.9 156.0 163.7 Grants 208.6 153.7 167.8 157.6 151.4 Source: OECD, Geographical Distribution o f Financial Flows to Developing Countries. Annex 4. Table 6: Net Official DevelopmentAssistance (in US$ millions), 2001-2005 2001 2002 2003 2004 2005 Bilateral 54.8 84.7 121.2 185.8 180.9 us 4.9 21.2 49.0 43.8 54.7 Belgium 4.7 7.1 4.4 34.8 14.5 France 7.5 16.5 15.9 25.1 21.3 Netherlands 11.7 9.6 12.7 23.3 22.9 Multilateral 82.4 86.8 106.2 175.8 184.1 EU 62.6 36.4 47.8 69.1 85.4 IDA 2.2 25.0 27.6 43.3 43.1 IMF -6.2 -2.8 39.1 21.1 UNDP 4.5 5.4 6.0 8.9 6.5 Total 137.2 171.5 227.4 361.5 365.0 Grants 145.0 152.2 203.1 288.3 328.4 Source: OECD, Geographical Distribution o f Financial Flows to Developing Countries. 136 Annex 4. Table 7: Government FinancialOperations (in FBu billions), 2001-2006 2001 2002 2003 2004 2005 2006 Domestic revenue 110.2 118.5 136.1 147.0 172.0 178.3 Tax revenue 103.2 104.9 120.6 133.6 158.8 162.2 Income tax 28.5 29.4 32.4 35.7 41.8 43.8 Taxes o n goods and services 48.7 51.9 58.4 67.6 78.3 82.9 Taxes o n international trade 21.7 23.2 29.5 30.0 38.4 33.2 Other tax revenue 4.3 0.4 0.3 0.3 0.3 2.3 Nontax revenue 7.0 13.6 15.5 13.4 13.2 16.1 Expenditure and net lending 149.8 151.5 224.7 291.1 316.5 379.2 Current expenditure 118.6 119.5 141.8 163.1 200.7 226.9 Salaries 40.1 45.9 53.8 58.6 72.6 96.6 Civilian 21.6 23.6 30.8 34.8 41.9 58.5 Military 18.5 22.3 23.0 23.8 24.0 23.0 National Police 6.7 15.1 Goods and services 44.1 38.5 47.4 53.6 65.8 62.2 Civilian 18.4 19 23.4 28.1 26.7 27.1 Military 25.7 19.5 24.0 25.5 29.6 23.0 National Police 9.5 12.1 Transfers and subsidies 15.9 16.4 15.8 26.4 30.3 40.6 Interest payments 18.5 18.7 24.8 24.5 32.0 27.5 Domestic 9.6 7.7 13.3 14 19.4 16.1 Foreign 8.9 11 11.5 10.5 12.6 11.4 Exceptional expenditure 0.0 0.0 0.0 10.7 33.2 35.0 DDR 0 0 0 7.4 8.7 35.0 Elections 0 0 0 3.3 24.5 0.0 Capital expenditure 35.1 33.4 84.9 119.9 84 119.3 Domestic resources 18.5 6.4 26.5 36.2 19.5 37.8 External resources 16.6 27 58.4 83.7 64.5 81.5 N e t lending -3.9 -1.4 -2 -2.6 -1.4 -2.0 Primary balance 1/ 83.5 11.3 -7.4 -38.5 -49.4 -93.9 Primary balance including exceptional expenditure -4.5 12.7 -5.4 -25.2 -14.8 -56.9 Overall balance (commitment basis) Including grants -28.4 -7.8 -40.1 -39.0 -51.6 -20.3 Excluding grants -39.6 -33.0 -88.6 -144.1 -144.5 -200.9 Change in arrears (reduction -) 21.3 9.4 -2.2 -58.5 -10.8 -21.6 External (current interest) 18.5 6.6 4.2 -49 -10.7 -0.2 Domestic 2.8 2.8 -6.4 -9.5 -0.1 -2 1.4 Overall balance (cash basis) Including grants -7.1 1.6 -42.3 -97.5 -62.4 -41.9 Excluding-grants -18.3 -23.6 -90.8 -202.6 -155.3 -222.5 Sources: BRB, IMF. 137 Annex 5: Public Expenditure Statistics Source: Ministry o f Finance, SIGEFI. Excludes debt service, externally financed recurrent and capital expenditures. 0 In FBu billions and percentages. Actuals o n a commitment basis. Annex 5. Table 1: a): Total Public Expenditure - 2001 Revised Budget ~~~ ~ Current expenditures Goods and Transfers & Capital Sectors Wage bill services Subsidies TOTAL expenditures TOTAL I n FBu billions Education 12.53 2.1 5.15 19.78 0.89 20.66 Health 1.57 0.92 0.94 3.43 0.3 3.73 Defense 18.79 14.7 0.04 33.53 0.53 34.06 Agriculture 0.27 0.15 0.43 0.85 1.13 1.98 Energy and mines 0.11 0.06 0 0.17 0.3 1 0.48 Transport & Telecom 0.03 0.08 0.1 0.2 1 0.35 0.55 Public works 0.18 0.57 0 0.76 0.45 1.21 Other 6.67 8.7 4.92 20.28 4.98 25.27 TOTAL 40.15 27.28 11.58 79.01 8.93 87.94 I n percentages Education 63.3 10.6 26.0 100.0 4.3 100 Health 45.8 26.8 27.4 100.0 8.0 100 Defense 56.0 43.8 7.5 100.0 1.6 100 Agriculture 31.8 17.6 2.2 100.0 57.1 100 Energy and mines 64.7 35.3 0.0 100.0 64.6 100 Transport & Telecom 14.3 38.1 0.5 100.0 63.6 100 Public works 23.7 75.0 0.0 100.0 37.2 100 Other 32.9 42.9 24.9 100.0 19.7 100 138 - Annex 5. Table 1: b): Total Public Expenditure 2001 (Actuals on a Commitment Basis) Current expenditures Transfers Goods and & Capital Sectors Wage bill services Subsidies TOTAL expenditures TOTAL I n FBu billions Education 12.85 2.04 1.26 16.14 0.86 17.01 Health 1.56 0.89 0.94 3.38 0.3 3.69 Defense 18.46 14.99 0.04 33.48 0.53 34.01 Agriculture 0.26 0.13 0.4 0.8 1.12 1.91 Energy and mines 0.11 0.06 0 0.17 0.3 1 0.48 Transport & Telecom 0.03 0.08 0.1 0.2 0.34 0.54 Public works 0.18 0.58 0 0.76 0.37 1.13 Other 6.62 7.26 3.04 16.93 2.55 19.48 TOTAL 40.06 26.03 5.77 71.86 6.38 78.24 I n percentages Education 79.6 12.6 7.8 100.0 5.1 100 Health 46.2 26.3 27.8 100.0 8.1 100 Defense 55.1 44.8 0.1 100.0 1.6 100 Agriculture 32.5 16.3 50.0 100.0 58.6 100 Energy and mines 64.7 35.3 0.0 100.0 64.6 100 Transport & Telecom 15.0 40.0 50.0 100.0 63.0 100 Public works 23.7 76.3 0.0 100.0 32.7 100 Other 39.1 42.9 18.0 100.0 13.1 100 - Annex 5. Table 1: c): Total Public Expenditure 2002 (Revised Budget) Current expenditures Transfers Goods and & Capital Sectors Wage bill services Subsidies TOTAL expenditures TOTAL I n FBu billions Education 13.19 2.19 6.01 21.39 1.23 22.62 Health 1.62 1.03 1.14 3.79 0.17 3.95 Defense 23.70 16.75 0.12 40.57 0.80 41.37 Agriculture 0.28 0.18 0.56 1.01 1.03 2.04 Energy and mines 0.12 0.06 0.00 0.18 0.39 0.57 Transport & Telecom 0.03 0.10 0.11 0.24 0.33 0.57 Public works 0.09 0.86 0.08 1.04 0.89 1.93 Other 8.19 11.42 5.79 25.39 3.04 28.44 TOTAL 47.22 32.59 13.81 93.62 7.87 101.49 I n percentages Education 61.7 10.2 28.1 100.0 5.4 100 Health 42.7 27.2 30.1 100.0 4.3 100 Defense 58.4 41.3 0.3 100.0 1.9 100 Agriculture 27.7 17.8 55.4 100.0 50.5 100 Energy and mines 66.7 33.3 0.0 100.0 68.4 100 Transport & Telecom 12.5 41.7 45.8 100.0 57.9 100 Public works 8.7 82.7 7.7 100.0 46.1 100 Other 32.3 45.0 22.8 100.0 10.7 100 139 - Annex 5. Table d): Total Public Expenditure 2002 (Actuals on a Commitment Basis) Current expenditures Transfers Goods and & Capital Sectors Wage b i l l services Subsidies TOTAL expenditures TOTAL I n FBu billions Education 13.60 2.38 5.63 21.61 0.72 22.33 Health 1.57 1.02 1.08 3.68 0.16 3.84 Defense 22.30 15.21 0.12 37.63 0.73 38.36 Agriculture 0.26 0.18 0.56 1.oo 0.98 1.98 Energy and mines 0.12 0.09 0.00 0.20 0.1 1 0.32 Transport & Telecom 0.03 0.12 0.11 0.25 0.3 1 0.56 Public works 0.09 0.82 0.00 0.91 0.57 1.49 Other 8.09 9.07 5.50 22.66 2.78 25.45 TOTAL 46.07 28.89 13.00 87.95 6.37 94.32 I n percentages Education 62.9 11.0 26.1 100.0 3.2 100 Health 42.7 27.7 29.3 100.0 4.2 100 Defense 59.3 40.4 0.3 100.0 1.9 100 Agriculture 26.0 18.0 56.0 100.0 49.5 100 Energy and m i n e s 60.0 45.0 0.0 100.0 34.4 100 Transport & Telecom 12.0 48.0 44.0 100.0 55.4 100 Public works 9.9 90.1 0.0 100.0 38.3 100 Other 35.7 40.0 24.3 100.0 10.9 100 - Annex 5. Table 1 e): Total Public Expenditure 2003 (Revised Budget) Current expenditures Transfers Goods and & Capital Sectors Wage bill services Subsidies TOTAL expenditures TOTAL I n FBu billions Education 19.01 2.43 6.73 28.17 0.41 28.57 Health 1.63 1.04 1.17 3.84 0.17 4.00 Defense 24.00 6.56 0.12 40.68 0.39 41.06 Agriculture 0.28 0.21 2.09 2.58 0.93 3.51 Energy and m i n e s 0.12 0.07 0.00 0.20 0.23 0.43 Transport & Telecom 0.03 0.10 0.12 0.25 0.01 0.26 Public works 0.10 0.79 0.33 1.23 0.53 1.76 Other 10.06 1.96 6.67 28.69 2.50 3 1.20 TOTAL 55.23 33.16 17.24 105.63 5.16 110.79 I n percentages Education 67.5 8.6 23.9 100.0 1.4 100 Health 42.4 27.1 30.5 100.0 4.3 100 Defense 59.0 40.7 0.3 100.0 0.9 100 Agriculture 10.9 8.1 81.0 100.0 26.5 100 Energy and mines 60.0 35.0 0.0 100.0 53.5 100 Transport & Telecom 12.0 40.0 48.0 100.0 3.8 100 Public works 8.1 64.2 26.8 100.0 30.1 100 Other 35.1 41:7 23.2 100.0 8.0 100 140 - Annex 5. Table 1 f): Total Public Expenditure 2003 (Actual on a Commitment Basis) Current expenditures Goods and Transfers & Capital Sectors Wage bill services Subsidies TOTAL expenditures TOTAL I n FBu billions Education 18.39 2.43 6.73 27.55 0.36 27.91 Health 1.64 1.04 1.16 3.84 0.11 3.95 Defense 22.99 16.53 0.12 39.64 0.39 40.03 Agriculture 0.26 0.14 0.59 0.99 0.88 1.87 Energy and mines 0.12 0.06 0.00 0.18 0.22 0.40 Transport & Telecom 0.03 0.10 0.12 0.25 0.01 0.26 Public works 0.11 0.76 0.33 1.20 0.29 1.49 Other 10.22 11.66 6.64 28.52 2.31 30.84 TOTAL 53.76 32.72 15.70 102.19 4.56 106.75 I n percentages Education 66.8 8.8 24.4 100.0 1.3 100 Health 42.7 27.1 30.2 100.0 2.8 100 Defense 58.0 41.7 0.3 100.0 1.o 100 Agriculture 26.3 14.1 59.6 100.0 47.1 100 Energy and mines 66.7 33.3 0.0 100.0 55.0 100 Transport & Telecom 12.0 40.0 48.0 100.0 3.8 100 Public works 9.2 63.3 27.5 100.0 19.5 100 Other 35.8 40.9 23.3 100.0 7.5 100 141 - Annex 5. Table 1 g): Total Public Expenditure 2004 (Revised Budget) Current expenditures Transfers Goods and & Capital Sectors Wage bill services Subsidies TOTAL expenditures TOTAL I n FBu billions Education 22.44 2.72 8.68 33.83 0.58 34.41 Health 1.75 1.59 1.48 4.82 0.49 5.31 Defense 25.32 18.16 0.06 43.53 0.80 44.33 Agriculture 0.38 0.22 2.12 2.72 1.08 3.80 Energy and mines 0.13 0.13 0.00 0.26 0.33 0.59 Transport & Telecom 0.03 0.10 0.14 0.27 0.40 0.68 Public works 0.11 0.93 0.38 1.43 0.78 2.22 Other 11.65 14.00 7.68 33.33 3.20 36.53 TOTAL 61.80 37.85 20.54 120.19 7.66 127.85 I n percentages Education 66.3 8.0 25.7 100.0 1.7 100 Health 36.3 33.0 30.7 100.0 9.2 100 Defense 58.2 41.7 0.1 100.0 1.8 100 Agriculture 14.0 8.1 77.9 100.0 28.4 100 Energy and mines 50.0 50.0 0.0 100.0 55.9 100 Transport & Telecom 11.1 37.0 51.9 100.0 58.8 100 Public works 7.7 65 .O 26.6 100.0 35.1 100 Other 35.0 42.0 23.0 100.0 8.8 100 - Annex 5. Table 1 g): Total Public Expenditure 2004 (Actuals on a commitment basis) Current expenditures Transfers Goods and & Capital Sectors Wage bill services Subsidies TOTAL expenditures TOTAL I n FBu billions Education 21.06 2.55 8.68 32.28 0.55 32.83 Health 1.77 1.40 1.46 4.62 0.05 4.67 Defense 23.78 18.10 0.06 41.94 0.80 42.74 Agriculture 0.29 0.18 0.62 1.09 1.04 2.14 Energy and mines 0.13 0.12 0.00 0.25 0.33 0.58 Transport & Telecom 0.03 0.10 0.14 0.27 0.40 0.67 Public works 0.11 0.91 0.37 1.38 0.69 2.07 Other 11.44 12.89 7.54 31.87 3 -40 35.27 TOTAL 58.60 36.26 18.85 113.71 7.26 120.97 I n percentages Education 65.2 7.9 26.9 100.0 1.7 100 Health 38.3 30.3 31.6 100.0 1.1 100 Defense 56.7 43.2 0.1 100.0 1.9 100 Agriculture 26.6 16.5 56.9 100.0 48.6 100 Energy and mines 52.0 48.0 0.0 100.0 56.9 100 Transport & Telecom 11.1 37.0 51.9 100.0 59.7 100 Public works 8.0 65.9 26.8 100.0 33.3 100 Other 35.9 40.4 23.7 100.0 9.6 100 Source: Ministry o f Finance, SIGEFI. 142 - Annex 5. Table 1 h): Total Public Expenditure 2005 (Original Budget) Current expenditures Transfers Goods and & Capital Sectors Wage bill services Subsidies TOTAL expenditures TOTAL I n FBu billions Education 26.44 1.06 13.74 41.24 0.87 42.1 1 Health 1.78 0.96 2.3 1 5.06 0.44 5.49 Defense 25.07 17.21 1.27 43.54 0.85 44.39 Agriculture 0.38 0.22 0.68 1.28 1.oo 2.28 Energy and mines 0.13 0.04 0.10 0.27 0.36 0.63 Transport & Telecom 0.03 0.20 0.20 0.44 0.33 0.77 Public works 0.12 1.31 0.19 1.62 0.72 2.34 Other 21.94 19.63 11.38 52.96 3.44 56.39 TOTAL 75.89 40.64 29.88 146.40 8.00 154.41 I n percentages Education 64.1 2.6 33.3 100.0 2.1 100 Health 35.2 19.0 45.7 100.0 8.0 100 Defense 57.6 39.5 2.9 100.0 1.9 100 Agriculture 29.7 17.2 53.1 100.0 43.9 100 Energy and mines 48.1 14.8 37.0 100.0 57.1 100 Transport & Telecom 6.8 45.5 45.5 100.0 42.9 100 Public works 7.4 80.9 11.7 100.0 30.8 100 Other 41.4 37.1 21.5 100.0 6.1 100 - Annex 5. Table 1 i):Total Public Expenditure 2005 (Actuals on a Commitment Basis) Current expenditures Transfers Goods and & Capital Sectors Wage bill services Subsidies TOTAL expenditures TOTAL I n FBu billions Education 28.07 1.02 13.73 42.83 0.63 43.46 Health 1.83 0.75 2.29 4.87 0.36 5.23 Defense 23.98 17.20 1.25 42.42 0.85 43.27 Agriculture 0.33 0.18 0.68 1.20 0.98 2.18 Energy and mines 0.12 0.04 0.10 0.26 0.30 0.57 Transport & Telecom 0.03 0.20 0.20 0.43 0.32 0.75 Public works 0.11 1.32 0.19 1.62 0.56 2.19 Other 18.17 18.76 11.21 48.14 3.09 5 1.22 TOTAL 72.64 39.48 29.65 141.77 7.09 148.86 I n percentages Education 65.5 2.4 32.1 100.0 1.4 100 Health 37.6 15.4 47.0 100.0 6.9 100 Defense 56.5 40.5 2.9 100.0 2.0 100 Agriculture 27.5 15.0 56.7 100.0 45.0 100 Energy and mines 46.2 15.4 38.5 100.0 52.6 100 Transport & Telecom 7.0 46.5 46.5 100.0 42.7 100 Public works 6.8 81.5 11.7 100.0 25.6 100 Other 37.7 39.0 23.3 100.0 6.0 100 143 - Annex 5. Table 1 j): Total Public Expenditure 2006 (Revised Budget) Current expenditures Transfers Goods and & Capital Sectors Wage bill services Subsidies TOTAL expenditures TOTAL I n FBu billions Education 41.02 1.28 18.43 60.74 19.45 80.19 Health 2.76 10.85 3.37 16.99 2.67 19.65 Defense 23.16 23.41 1.26 47.83 0.85 48.68 Agriculture 0.63 0.37 0.77 1.77 2.52 4.29 Energy and mines 0.15 0.05 0.42 0.61 7.77 8.39 Transport & Telecom 0.04 0.23 0.27 0.53 0.4 1 0.94 Public works 0.16 1.51 0.22 1.89 3.73 5.62 Other 30.2 3 1.48 15.84 77.52 7.72 85.23 TOTAL 98.11 69.18 40.58 207.87 45.13 252.99 I n percentages Education 67.5 2.1 30.3 100.0 24.3 100 Health 16.2 63.9 19.8 100.0 13.6 100 Defense 48.4 48.9 2.6 100.0 1.7 100 Agriculture 35.6 20.9 43.5 100.0 58.7 100 Energy and m i n e s 24.6 8.2 68.9 100.0 92.6 100 Transport & Telecom 7.5 43.4 50.9 100.0 43.6 100 Public works 8.5 79.9 11.6 100.0 66.4 100 Other 39.0 40.6 20.4 100.0 9.1 100 - Annex 5. Table 1 k): Total Public Expenditure 2006 (Actuals on a Commitment Basis) Current expenditures Transfers Goods and & Capital Sectors Wage bill services Subsidies TOTAL expenditures TOTAL I n FBu billions Education 39.08 1.16 17.87 58.12 8.32 66.43 Health 2.69 3.86 3.20 9.76 0.10 9.85 Defense 22.94 22.91 1.25 47.09 0.84 47.93 Agriculture 0.46 0.22 0.73 1.42 1.80 3.22 Energy and mines 0.15 0.05 0.42 0.61 1.46 2.07 Transport & Telecom 0.04 0.23 0.27 0.53 0.03 0.56 Public works 0.17 1.48 0.22 1.87 2.67 4.54 Other 27.59 26.14 14.65 68.39 5.35 73.74 TOTAL 93.11 56.05 38.62 187.78 20.56 208.34 I n percentages Education 67.2 2.0 30.7 100.0 12.5 100 Health 27.6 39.5 32.8 100.0 1.o 100 Defense 48.7 48.7 2.7 100.0 1.8 100 Agriculture 32.4 15.5 51.4 100.0 55.9 100 Energy and mines 24.6 8.2 68.9 100.0 70.5 100 Transport & Telecom 7.5 43.4 50.9 100.0 5.4 100 Public works 9.1 79.1 11.8 100.0 58.8 100 Other 40.3 38.2 21.4 100.0 7.3 100 Source: Ministry o f Finance, SIGEFI. 144 - Annex 5. Table 1 1): Total Public Expenditure 2007 (Original Budget) Current expenditures Goods and Transfers & Capital Sectors Wage bill services Subsidies TOTAL expenditures TOTAL I n FBu billions Education 53.16 1.49 22.04 76.69 10.85 87.54 Health 4.65 3.93 4.42 13.00 2.97 15.98 Defense 26.69 24.09 1.48 52.26 1.31 53.57 Agriculture 0.99 0.37 0.89 2.26 4.13 6.39 Energy and mines 0.16 0.05 0.54 0.74 3.92 4.66 Transport & Telecom 0.04 0.26 0.29 0.60 0.89 1.49 Public works 0.18 4.17 0.24 4.59 3.93 8.52 Other 37.47 3 1.93 16.85 86.25 8.98 95.24 TOTAL 123.34 66.29 46.76 236.39 36.98 273.38 Education 69.3 1.9 28.7 100.0 12.4 100 Health 35.8 30.2 34.0 100.0 18.6 100 Defense 51.1 46.1 2.8 100.0 2.4 100 Agriculture 43.8 16.4 39.4 100.0 64.6 100 Energy and mines 21.6 6.8 73.0 100.0 84.1 100 Transport & Telecom 6.7 43.3 48.3 100.0 59.7 100 Public works 3.9 90.8 5.2 100.0 46.1 100 Other 43.4 37.0 19.5 100.0 9.4 100 Annex 5. Table 1 m): Functional Distribution of Domestically-FinancedPublic Expenditures, 2001- 2006 (Commitments, excluding debt service) 2001 2002 2003 2004 2005 2006 Average (in billions o f constant 2001 FBu) General public services (excl security and debt repayments) 23.5 25.5 27.3 29.5 27.9 41.5 29.2 Security spending 36.5 40.2 37.6 37.3 42.8 50.9 40.9 Defense 34.0 37.7 35.3 34.8 30.2 32.4 34.0 Interior and Public Security 2.5 2.5 2.4 2.5 12.6 18.6 6.8 Priority ministries 28.6 30.0 31.6 35.2 37.9 59.0 37.1 Education 20.9 21.9 24.6 26.7 30.3 44.8 28.2 Health 3.7 3.8 3.5 4.1 3.6 782 4.3 Agriculture 1.9 1.9 1.6 1.7 1.5 2.2 1.8 Infrastructure 2.1 2.3 1.9 2.7 2.4 4.8 2.7 Other expenditure 2.0 2.3 2.4 2.7 2.6 3.1 2.5 Total 90.7 98.0 98.9 104.7 111.2 154.6 109.7 (inpercentage o f total public expenditures) General public services (excl security and debt repayments) 25.9 26.0 27.6 28.2 25.1 26.9 26.6 Security spending 40.3 41.0 38.0 35.6 38.4 32.9 37.7 Defense 37.5 38.5 35.6 33.2 27.1 20.9 32.1 Interior and Public Security 2.8 2.5 2.4 2.4 11.3 12.0 5.6 Priority ministries 31.6 30.6 31.9 33.6 34.1 38.2 33.3 Education 23.0 22.4 24.8 25.5 27.2 29.0 25.3 Health 4.1 3.9 3.5 3.9 3.3 4.6 3.9 Agriculture 2.1 2.0 1.7 1.7 1.4 1.4 1.7 Infrastructure 2.4 2.4 1.9 2.6 2.2 3.1 2.4 Other expenditure 2.2 2.4 2.4 2.6 2.4 2.0 2.3 Total 100 100 100 100 100 100 100.0 145 Annex 5. Table 1 n): Domestically Financed Public Expenditure by Economic Classification, 2002 Original Revised Executed % % Revised- Original- Executed Executed Current expenditure 110.4 110.7 109.8 99.2 99.4 Salaries and benefits 47.4 47.2 46.1 97.6 97.2 Goods and services 32.4 32.6 28.9 88.6 89.1 Transfers and subsidies 13.4 13.8 13.0 94.1 97.2 Interest payments (due) 17.3 17.1 21.9 127.9 126.5 Domestic debt 8.3 7.7 12.5 162.9 151.4 Foreign debt 9.0 9.4 9.3 99.2 103.5 Capital expenditure (domestic resources) 6.6 7.9 6.4 80.9 96.0 TOTAL 117.1 118.6 116.2 98.0 99.2 Note: Excludes Fonds Speciaux, net lending, contingency line and foreign-financed expenditure Annex 5. Table 1 0 ) : Domestically Financed Public Expenditure by Economic Classification, 2003 Original Revised Executed % O Y Revised- Original- Executed Executed Current expenditure 129.1 128.4 129.8 101.1 100.5 Salaries and benefits 55.3 55.2 53.8 97.3 97.2 Goods and services 34.9 33.2 32.7 98.7 93.8 Transfers and subsidies 15.8 17.2 15.7 91.1 99.4 Interest payments (due) 23.1 22.7 22.3 98.0 96.6 Domestic debt 11.6 11.3 13.3 118.2 114.9 Foreign debt 11.4 11.4 8.9 78.1 78.1 Capital expenditure (domestic resources) 4.9 5.2 4.6 88.4 93.9 TOTAL 133.9 133.5 134.3 100.6 100.3 Note: Excludes Fonds Speciaux, net lending, contingency line and foreign-financed expenditure 146 Annex 5. Table 1 p): Domestically Financed Public Expenditure by Economic Classification, 2004 Original Revised Executed YO YO Revised- Original- Executed Executed Current expenditure 146.9 152.8 144.2 94.4 98.2 Salaries and benefits 58.9 61.8 58.6 94.8 99.4 Goods and services 40.0 37.8 36.3 95.8 90.6 Transfers and subsidies 18.6 20.5 18.9 91.8 101.4 Interest payments (due) 29.4 32.6 30.5 93.4 103.8 Domestic debt 14.3 17.6 14.7 83.7 103.1 Foreign debt 15.1 15.1 15.8 104.8 104.3 Capital expenditure (domestic resources) 7.2 7.7 7.3 94.7 100.9 TOTAL 154.1 160.5 151.5 94.4 98.3 Note: Excludes Fonds Speciaux, net lending, contingency line and foreign-financed expenditure Annex 5. Table 1 q): Domestically Financed Public Expenditure by Economic Classification, 2005 Original Executed % Original- Executed Current expenditure 185.5 170.8 92.0 Salaries and benefits 75.9 72.6 95.7 Goods and services 40.6 39.5 97.2 Transfers and subsidies 32.5 30.3 93.0 Interest payments (due) 36.5 28.4 77.8 Domestic debt 17.3 12.3 71.0 Foreign debt 19.2 16.1 83.9 Capital expenditure (domestic resources) 7.5 7.1 94.5 Contingency line 5.3 5.8 108.7 TOTAL 193.0 177.9 92.1 Note: Excludes Fonds Speciaux, net lending, contingency line and foreign-financed expenditure Annex 5. Table 1 r): Domestically Financed Public Expenditure by Economic Classification, 2006 Original Revised Executed % % Revised- Original- Executed Executed Current expenditure 218.1 236.0 209.6 88.8 96.1 Salaries and benefits 90.1 98.1 93.9 95.7 104.2 Goods and services 66.4 69.2 55.7 80.5 83.8 Transfers and subsidies 35.1 41.2 39.5 95.8 112.3 Interest payments (due) 26.4 27.5 20.2 73.6 76.6 Domestic debt 11.3 11.3 10.1 89.5 89.5 Foreign debt 15.1 16.2 10.1 62.4 67.0 Capital expenditure (domestic 44.8 45.1 20.6 resources) 45.6 45.9 TOTAL 262.9 281.1 230.2 81.9 87.6 Note: Excludes Fonds Speciaux, net lending, contingency line and foreign-financed expenditure. 147 Annex 6: Summary list of equations for the small model Economic Growth + 0.837*LF-POP - 0.003*ttrade + 0.267*IP-GDP gr-GDPpc = C-1 (1) - 0.525*PubWS-GDP + 0.1 lO*PubTR-GDP - 0.293*IntPaym-GDP + 0.42*PubOT-GDP + 0.567*PubCAP-GDP - 0.056*TaxRev-GDP + 0.095*NtaxRev_GDP + 0.079*FoGr-GDP GDPPc = (1 + p-GDPpc)*GDPpc.l Private Investment IP-GDP = C-3 + 0.015*gr-GDPpc + 0.637*GDPpc + 0.063*PubCAP-GDP - 0.0023*Fdebt-GDP - 0.015*(G-GDP - R-GDP) Total Expenditure PubGTOT = PubWS-GDP + PubTR-GDP + IntPaym-GDP (4) + PubOT-GDP + PubCAP-GDP, Public Investment PubCAP-GDP = C-5 + 4.729*log(GDPpc) + 0.085*FoGr-GDP - 0.216*urban + 7.985*Open - 0.167*IntPaym-GDP + 2.297*log(FdebtWGDP) - 0.15 * [log(Fdebt-GDP)] 2. Total Government Resources PubRTOT = TaxRev-GDP + NtaxRev-GDP + FoGr-GDP FDebt-GDP = FDebt-GDP.l - ADDebt-GDP + PubGTOT - PubRTOT or FoGr-GDP = - AFDebt-GDP.l - ADDebt-GDP + PubGTOT - TaxRev-GDP - NtaxRev-GDP Interest payments IntPaym-GDP = Dint-rate*DDebt-GDP-l + Fint-rate*FDebt-GDP-l Poverty links gr-POVHRAV = -9.3*(1 - Gini)3*@-GDPpc g-POVHGEP = 0.39 - 4.93*(1 - GINI)*g-GDPpc + 0,6*log(GDPpc)*AGINI There are therefore two budget closure rules: equation (7a), which corresponds to the “borrowing” mode, and equation (7b), which corresponds to the “aid” mode. 148 Annex 7: Structure of a Small Model linking Public Expenditure, Growth, and Poverty A \\ Public Wages -$ investment Non- interest I /and salaries I expenditure on goods/services Tax expenditure Transfers and reven (1es Subsidies Interest Non-Tax Overall deficit (before revenues \ grants) Payments I t A I I Overall deficit (after grants) 149 Annex 8 :Tablel: Baseline Scenario of Simulations (2006-2015) 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Real GDP per capita (% change) 2.1 1.3 1.4 1.6 1.8 1.7 1.7 1.7 1.6 1.4 Poverty rate (Ravallion's adjusted formula) 66.9 64.5 62.0 59.3 56.4 53.8 51.2 48.8 46.6 44.8 Government Sector (% of GDP) Total resources (including grants) 36.8 39.2 39.1 39.0 38.9 38.8 38.6 38.6 38.4 38.2 Total revenues 18.8 20.2 20.0 20.0 20.0 20.0 20.1 20.1 20.1 20.1 Tax revenue 17.2 18.3 18.1 18.2 18.2 18.2 18.2 18.2 18.2 18.2 Nontax revenue 1.6 1.9 1.9 1.8 1.8 1.8 1.8 1.8 1.8 1.8 Foreign aid (grants) 18.0 19.0 19.1 18.9 18.9 18.8 18.5 18.5 18.4 18.1 Total expenditure 38.5 40.9 40.8 40.7 40.6 40.5 40.3 40.3 40.1 39.9 Current expenditure 22.8 25.0 24.9 24.6 24.5 24.5 24.3 24.3 24.3 24.3 Wages and salaries 10.0 11.8 11.8 11.5 11.2 11.2 11.0 11.0 11.0 11.0 Transfers 4.2 4.6 4.5 4.6 4.7 4.7 4.7 4.7 4.7 4.7 Other current expenditure 8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 Capital expenditure 12.8 12.9 13.0 13.1 13.2 13.0 13.0 13.0 12.8 12.5 Interest payments 2.9 2.9 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 Overall fiscal balance (including grants) -1.7 -1.7 -1.7 -1.7 -1.7 -1.7 -1.7 -1.7 -1.7 -1.7 Memorandum items Private investment (% o f GDP) 8.0 8.0 8.0 8.0 8.0 8.0 8.0 8.0 8.0 8.0 Aid (% o f total revenue) 95.8 94.1 95.3 94.6 94.6 93.7 92.3 92.2 91.5 90.2 Total A i d (% o f capital expenditure) 140.9 146.8 46.8 44.9 43.8 144.2 142.7 142.8 143.4 144.4 External debt (% o f GDP) 162.2 163.9 65.6 67.3 69.0 170.7 172.4 174.1 175.8 177.5 Domestic debt (% o f GDP) 24.6 24.6 24.6 24.6 24.6 24.6 24.6 24.6 24.6 24.6 Alternative Poverty rates Income per capita growth elasticity o f -0.5 66.9 66.5 66.0 65.5 64.9 64.4 63.8 63.3 62.8 62.3 Income per capita growth elasticity o f -1.0 66.9 66.0 65.1 64.1 62.9 61.9 60.8 59.8 58.8 58.0 Income per capita growth elasticity o f -1.5 66.9 65.6 64.2 62.7 61.0 59.5 58.0 56.5 55.1 54.0 150 Annex 8. Table2: Higher Total Public Investment, 2007-15 (Deviation from the Baseline scenario) 2007 2008 2009 2010 2011 2012 2013 2014 2015 Real GDP per capita (% change) 0.0 2.7 2.7 2.7 2.7 2.7 2.0 1.3 0.7 Poverty rate (Ravallion's adjusted formula) 0.0 -4.8 -8.8 -12.1 -14.8 -16.9 -18.0 -18.4 -18.2 Government Sector (YO of GDP) Total resources (including grants) 0.0 4.0 4.0 4.0 4.0 4.0 3.0 2.0 1.o Total revenues 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Tax revenue 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Nontax revenue 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Foreign aid (grants) 0.0 4.0 4.0 4.0 4.0 4.0 3.0 2.0 1.o Total expenditure 0.0 4.0 4.0 4.0 4.0 4.0 3.0 2.0 1.o Current expenditure 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Wages and salaries 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Transfers 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Other current expenditure 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Capital expenditure 0.0 4.0 4.0 4.0 4.0 4.0 3.0 2.0 1.o Interest payments 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Overall fiscal balance (including grants) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Memorandum items Private investment (% o f GDP) 0.0 0.3 0.3 0.3 0.3 0.3 0.2 0.2 0.1 Aid (% o f total revenue) 0.0 20.0 20.0 20.0 20.0 19.9 15.0 10.0 5.0 Total Aid (% o f capital expenditure) 0.0 -1 1.0 -10.5 -10.2 -10.4 -10.1 -8.1 -5.9 -3.3 External debt (% o f GDP) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Domestic debt (% o f GDP) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Alternative Poverty rates Income per capita growth elasticity o f -0.5 0.0 -0.9 -1.7 -2.6 -3.4 -4.2 -4.7 -5.1 -5.2 Income per capita growth elasticity o f -1.0 0.0 -1.8 -3.4 -5.0 -6.4 -7.8 -8.7 -9.3 -9.5 Income per capita growth elasticity o f -1.5 0.0 -2.6 -5.0 -7.2 -9.2 -10.9 -12.1 -12.7 -12.8 151 Annex 8. Table3: Reallocation of Spending to Higher Total Public Investment, 2007-2015 2007 2008 2009 2010 2011 2012 2013 2014 2015 Real GDP per capita (% change) 0.0 0.0 1.0 1.3 1.8 2.2 2.6 2.7 2.8 Poverty rate (based on Ravallion’s formula) 0.0 0.0 -1.7 -3.8 -6.2 -8.8 -11.4 -13.8 -15.8 Government Sector (YO of GDP) Total resources (including grants) 0.0 0.0 -1.0 -1.7 -2.7 -3.5 -4.4 -4.3 -4.2 Total revenues 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Tax revenue 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Nontax revenue 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Foreign aid (grants) 0.0 0.0 -1.0 -1.7 -2.7 -3.5 -4.4 -4.3 -4.2 Total expenditure 0.0 0.0 -0.9 -1.7 -2.7 -3.5 -4.4 -4.3 -4.2 Current expenditure 0.0 0.0 -1.5 -2.2 -3.2 -4.0 -5.0 -5.0 -5.0 Wages and salaries 0.0 0.0 -1.5 -2.2 -3.2 -4.0 -5.0 -5.0 -5.0 Transfers 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Other current expenditure 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Capital expenditure 0.0 0.0 0.5 0.5 0.5 0.5 0.6 0.7 0.9 Interest payments 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Overall fiscal balance (including grants) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Memorandum items Private investment (% o f GDP) 0.0 0.0 0.0 0.1 0.1 0.1 0.1 0.1 0.1 Aid (% o f total revenue) 0.0 0.0 -4.8 -8.6 -13.5 -17.3 -22.1 -21.4 -20.7 Total Aid (“ho f public investment) 0.0 0.0 -12.5 -17.9 -25.5 -31.4 -38.9 -39.4 -40.3 External debt (% of GDP) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Domestic debt (% o f GDP) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Alternative Poverty rates Income per capita growth elasticity o f -0.5 0.0 0.0 -0.3 -0.8 -1.3 -2.0 -2.8 -3.6 -4.4 Income per capita growth elasticity o f -1.0 0.0 0.0 -0.7 -1.5 -2.6 -3.8 -5.2 -6.7 -8.0 Income per capita growth elasticity o f -1.5 0.0 0.0 -1.0 -2.2 -3.7 -5.4 -7.4 -9.2 -1 1.0 152 Annex 8. Table 4a: Partial Debt Relief -Reallocation of savings to Public Investment, 2007-15 (Deviation f r o m the Baseline scenario) 2007 2008 2009 2010 2011 2012 2013 2014 2015 Real GDP per capita (% change) 0.0 0.1 0.6 0.6 0.6 0.6 0.6 0.7 0.7 Poverty rate (based o n Ravallion's formula) 0.0 -0.1 -1.0 -1.9 -2.7 -3.5 -4.2 -4.8 -5.4 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Government Sector (% o f GDP) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Total resources (including grants) 0.0 0.0 0.0 0.1 0.1 0.1 0.2 0.2 0.2 Total revenues 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 T a x revenue 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Nontax revenue 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Foreign aid (grants) 0.0 0.0 0.0 0.1 0.1 0.1 0.2 0.2 0.2 Total expenditure 0.0 0.0 0.0 0.1 0.1 0.1 0.2 0.2 0.2 Current expenditure 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Wages and salaries 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Transfers 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Other current expenditure 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Capital expenditure 0.0 0.0 0.6 0.6 0.6 0.7 0.7 0.7 0.8 Interest payments 0.0 0.0 -0.6 -0.6 -0.6 -0.6 -0.6 -0.6 -0.6 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Overall fiscal balance (including grants) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Memorandum items 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Private investment (% o f GDP) 0.0 0.2 0.2 0.2 0.2 0.2 0.2 0.3 0.3 Aid (% o f total revenue) 0.0 0.0 0.2 0.3 0.5 0.6 0.8 0.9 1.1 Total Aid (% o f public investment) 0.0 0.0 -5.9 -5.9 -6.1 -6.1 -6.3 -6.4 -6.7 External debt (% o f GDP) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Domestic debt (% o f GDP) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Alternative Poverty rates Income per capita growth elasticity o f - 0.5 0.0 0.0 -0.2 -0.4 -0.6 -0.8 -1.0 -1.2 -1.4 Income per capita growth elasticity o f - 1.o 0.0 0.0 -0.4 -0.8 -1.1 -1.5 -1.8 -2.2 -2.5 Income per capita growth elasticity o f - 1.5 0.0 -0.1 -0.6 -1.1 -1.6 -2.1 -2.6 -3.1 -3.5 153 Annex 8. Table 4b: Total Debt Relief -Reallocation of savings to Public Investment, 2007-15 (Deviation f r o m the Baseline scenario) 2007 2008 2009 2010 2011 2012 2013 2014 2015 Real GDP per capita (% change) 0.0 0.1 1.1 1.2 1.2 1.2 1.3 1.3 1.4 Poverty rate (based o n Ravallion's formula) 0.0 -0.2 -2.1 -3.8 -5.4 -6.8 -8.0 -9.2 -10.2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Government Sector (YO of GDP) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Total resources (including grants) 0.0 0.0 0.1 0.1 0.2 0.2 0.3 0.4 0.4 Total revenues 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Tax revenue 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Nontax revenue 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Foreign aid (grants) 0.0 0.0 0.1 0.1 0.2 0.2 0.3 0.4 0.4 Total expenditure 0.0 0.0 0.1 0.1 0.2 0.2 0.3 0.4 0.4 Current expenditure 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Wages and salaries 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Transfers 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Other current expenditure 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Capital expenditure 0.0 0.0 1.2 1.2 1.3 1.3 1.4 1.5 1.5 Interest payments 0.0 0.0 -1.1 -1.1 -1.1 -1.1 -1.1 -1.1 -1.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Overall fiscal balance (including grants) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Memorandum items 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Private investment (% o f GDP) 0.0 0.4 0.5 0.5 0.5 0.5 0.5 0.5 0.5 Aid (% o f total revenue) 0.0 0.0 0.3 0.6 0.9 1.2 1.5 1.9 2.2 Total Aid (% o f public investment) 0.0 0.0 - 1 1.4 -1 1.4 -11.7 -1 1.7 -1 1.9 -12.2 -12.7 External debt (% o f GDP) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Domestic debt (% o f GDP) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Alternative Poverty rates Income per capita growth elasticity o f -0.5 0.0 0.0 -0.4 -0.8 -1.2 -1.5 -1.9 -2.3 -2.7 Income per capita growth elasticity o f -1.0 0.0 -0.1 -0.8 -1.5 -2.2 -2.9 -3.6 -4.3 -5.0 Income per capita growth elasticity o f -1.5 0.0 -0.1 -1.2 -2.2 -3.2 -4.2 -5.1 -6.0 -6.9 154 Annex 8. Table 5: Gradual Wage Reduction, 2007-15 (Deviation f r o m the Baseline scenario) 2007 2008 2009 2010 2011 2012 2013 2014 2015 Real GDP per capita (% change) 0.0 0.0 1 .o 1.3 1.8 2.2 2.6 2.7 2.8 Poverty rate (based o n Ravallion's formula) 0.0 0.0 -1.7 -3.8 -6.2 -8.8 -11.4 -13.8 -15.8 Government Sector (YO of GDP) Total resources (including grants) 0.0 0.0 -1.0 -1.7 -2.7 -3.5 -4.4 -4.3 -4.2 Total revenues 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 T a x revenue 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Nontax revenue 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Foreign aid (grants) 0.0 0.0 -1.0 -1.7 -2.7 -3.5 -4.4 -4.3 -4.2 Total expenditure 0.0 0.0 -0.9 -1.7 -2.7 -3.5 -4.4 -4.3 -4.2 Current expenditure 0.0 0.0 -1.5 -2.2 -3.2 -4.0 -5.0 -5.0 -5.0 Wages and salaries 0.0 0.0 -1.5 -2.2 -3.2 -4.0 -5.0 -5.0 -5.0 Transfers 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Other current expenditure 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Capital expenditure 0.0 0.0 0.5 0.5 0.5 0.5 0.6 0.7 0.9 Interest payments 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Overall fiscal balance (including grants) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Memorandum items Private investment (% o f GDP) 0.0 0.0 0.0 0.1 0.1 0.1 0.1 0.1 0.1 Aid (% o f total revenue) 0.0 0.0 -4.8 -8.6 -13.5 -17.3 -22.1 -21.4 -20.7 Total Aid (% o f public investment) 0.0 0.0 -12.5 -17.9 -25.5 -31.4 -38.9 -39.4 -40.3 External debt (% o f GDP) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Domestic debt (% o f GDP) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Alternative Poverty rates Income per capita growth elasticity o f -0.5 0.0 0.0 -0.3 -0.8 -1.3 -2.0 -2.8 -3.6 -4.4 Income per capita growth elasticity o f -1.0 0.0 0.0 -0.7 -1.5 -2.6 -3.8 -5.2 -6.7 -8.0 Income per capita growth elasticity o f -1.5 0.0 0.0 -1.0 -2.2 -3.7 -5.4 -7.4 -9.2 -1 1.0 155 i Q\ 3\ 2 M 0 0 3 0 3 N c;' 00 2 c;' Q\ E 0 0 N 0 0 N 0 0 N e 4 e n 3 3 u L, u L, L, E 2 E 3 2 E E E E E 5 u” 5 i ; 2 3 2 3 3 B 3 9 E E E E b B a m e, B e e 0 0 s V s s Y 0 U 8 2 fi 2 2 Q\ Q\ 0 3 m 0 3 3 N k 0 0 N E 2 3 El CL 3 N .3 CI Y E -4 2 I 3 u 5 9 E- $j 9 E E E E 5 E 2 b a b a M a v) 0 M 8 s s Y 2 0 4 0 3 $ 8 3 I 3 3 3 3 u n 0 0 c) d 3 3 -4 2 0 2 0 0 0 hl hl 3 3 B 2 2 Y 8 E 5 E B B 0 s 2 3 3 3 0 0 0 2 0 2 0 2 0 0 0 0 N w hl I I a E 4 ig A -L 2 a u I ts; .I 4 Ir 8 k cu 0 E .s c 4 0) B Y v1 .I v1 4 z . a I d F - 4 3 x 3 3 N I a .E c) 4 Q\ 0 3 0 Q\ 3 3 ? W 0 0 N 0 hl 0 N J u 3 E E E c, E w u 2 3 3\ 3 3 N Q\ 0 0 c\1 Q\ 0 0 r 4 x 3 3 N BIBLIOGRAPHY Benitez, D., Estache, A., D-M. Niyungeko (2007), “Overcoming Obstacles to Growth in Burundi”. This report prepared in the context o f Burundi-Sources o f Rural Growth (P082767). Bricefio-Garmendia, C. (2007). “Fiscal Costs o f Infrastructure Provision: A Practitioner’s Guide”, The World Bank. Briceiio-Garmendia, C. and V. Foster, (2007). “More Fiscal Resources for Infrastructure? Evidence from East Africa”, The World Bank Bruns, B., Mingat, A,, Rakotomalala, R. (2003). Achieving Universal Primary Education by 2015 - a Chancefor Every Child. The World Bank: Washington D.C. Castro-Leal, F., Dayton, J., Demery, L.,Mehra, K. 1999. Public Social Spending in Africa: D o the Poor Benefit? The World Bank Research Observer 14 (1): 49-72. Chambas, G. and Laporte, B. 2007. Introduction de la TVA au Burundi. Etude Prdalable de 1’Impact Budgdtaire, Economique et Social. Draft. Devarajan, S., Swaroop, V. and Zhou, H-f. 1996. The Composition o f Public Expenditure and Economic Growth. Journal o f Monetary Economics 37: 3 13-344. Clements, B., Bhattacharya, R., Nguyen, T.Q., 2003. External Debt, Public Investment, and Growth in Low-Income Countries. (IMF Working Paper No.03/249). Washington, D.C. Dinh, H. 1999. Fiscal Solvency and Sustainability in Economic Management. The World Bank (Working Paper No. 2213): Washington, D.C. Gupta, S., Clements, B., Baldacci, E., and Mulas-Granados, C. Fiscal Policy, Expenditure Composition, and Growth in Low-Income Countries. Journal o f International Money and Finance 24: 44 1-463. Helis, J-L, L e Bouder, J-P, Ngudma, So,Tendelli, S. 2006. L a Reforme de la Gestion Budgktaire. I M F , Department of Fiscal Affairs. Washington, D.C. International Monetary Fund. 2006. Burundi - Selected Issues and Statistical Appendix, IMF Country Report No. 06/307, Washington, D.C. Mlambo, K and Oshikoya T.W. 2001. “Macroeconomic Factors and Investment in Africa”. f African Economies 10: 12-47. Journal o OECD (2005). Geographical Distribution of Financial Flows to Aid Recipients. OECD: Paris. Regideso, Rapport Annuel d’activites, Exercice 2005 et 2006. 167 Republic o f Burundi, 200 1-2007, Fixation du Budget Gkndral de la Rdpublique du Burundi Pour 1’Exercice, Cabinet du PrCsident. Republic o f Burundi, 2001-2006, Fixation du Budget Gdnkral Rdvisd de la Rdpublique du Burundi Pour 1 ’Exercice, Cabinet du PrCsident. Republic o f Burundi. 2007a. Programme d ’Investissements Publics Pour la Rdduction de la Pauvretd au Burundi. PIP 2007-2009. Ministkre de l a Planification du DCveloppement et de la Reconstruction. Republic o f Burundi. 2007b. Programme d’Actions Prioritaires de mise en a m v e du CSLP 2007- 201 0. Table Ronde des Partenaires, May 24-25. Republic o f Burundi. 2007c. Etude Approfondie de 1’Assiette Fiscal Pre-DeJinitiJ: Projet d’ Appui a l a Gestion Economique. SOFRECO. Republic o f Burundi. 2006. Programme d ’Investissernents Publics Pour la Rdduction de la Pauvretd au Burundi. PIP 2006-2008. Ministbre de l a Planification du DCveloppement et de l a Reconstruction. Republic o f Burundi. 2005a. Programme d’lnvestissements Publics Pour la Rdduction de la Pauvretd au Burundi. PIP 2005-2007. Ministke de la Planification du DCveloppement et de la Reconstruction. Republic o f Burundi. 2005b. Situation de la Gestion Courante de la Reforme Structurelle des Socidtds ci Participation Publique. Exercice 2005. Deuxikme Vice-prdsidence, Service ChargC des Entreprises Publiques. Schiavo-Campo, S. and D. Tommasi. (Ed.). 1999. Governance, Corruption and Public Financial Management. Asian Development Bank: Manilla. Sommeiller, E., and Q. Wodon 2007: Impact de 1’Education Primaire gratuite sur la Scolarisation au Burundi, World Bank, Mimeo. f the Public Sector. W.W. Norton & Company: New Stiglitz, J.E. 2000. Economics o YorWLondon, Tchaidze, R. 2007. Quasi-Fiscal Deficits in NonfinancialEnterprises. IMF (Working Paper WP/07/ 10). Washington D.C. UN/AIDS. 2006. 2006 Report on the Global AIDS Epidemic. Geneva. World Bank. 2007a. L e Systkme Educatif Burundais: Diagnostic et Perspectives pour une Nouvelle Politique Educative dans l e Contexte de 1’Education Primaire Gratuite pour Tous. The World Bank. (Working Paper No. 109). Washington D.C. -_-__--- ,2007b. “Burundi Joint IDA-IMF StaffAdvisory Note on the Poverty Reduction Strategy Paper. ” Washington, D.C. Processed. 168 -------- . 2007c. Sources of Rural Growth in Burundi: A Macroeconomic Perspective. Draft. Washington, D.C. -------- . 2005a. “Enhanced Heavily Indebted Poor Countries (HIPC) Debt Initiative.” Washington, D.C. Processed. ____-___ , 2005b. “Burundi Interim Strategy Note.” Washington, D.C. Processed. -------- 2004a. “Burundi Country Financial Accountability Assessment.” Washington, D.C. , Processed. -------- .2004b. Zambia: Public Expenditure Management and Financial Accountability Review. The World Bank (Country Studies Series): Washington D.C. -------- . 2007. Tajikistan: Programmatic Public Expenditure Review, The World Bank: Washington D.C. Processed. -------- .2008. Haiti: Public Expenditure Management and Financial Accountability Review. The World Bank: Washington D.C. Processed. -------- . 1998. Assessing Aid: What Works, What Doesn ’t and Why. Oxford University Press: New York. ____-__ . 1992. “Burundi Public Expenditure Review”. Washington, D.C. 169