46312 FINAL SYNTHESIS REPORT A COMPARATIVE ANALYSIS OF DECENTRALISATION IN KENYA, TANZANIA AND UGANDA Jesper Steffensen Per Tidemand Harriet Naitore (Kenya Only) Emmanual Ssewankambo (Uganda Only) Eke Mwaipopo (Tanzania Only) August 2004 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda LIST OF CONTENTS LIST OF CONTENTS.............................................................................................ii ABBREVIATIONS.................................................................................................iv ACKNOWLEDGEMENTS ............................................................................... viii 1. INTRODUCTION ............................................................................................1 1.1 BACKGROUND TO THE STUDY..............................................................1 1.2 SPECIFIC OBJECTIVES AND PURPOSE..................................................2 1.3 SCOPE OF THE STUDY AND DIMENSIONS TO BE REVIEWED ........2 1.4 THE STUDY APPROACH AND THE TEAM ............................................7 1.5 THE SYNTHESIS REPORT.........................................................................8 1.6 COUNTRY PROFILES...............................................................................10 2. LEGAL AND POLICY ISSUES ................................................................11 2.1 INTRODUCTION .......................................................................................11 2.2 HISTORICAL BACKGROUND.................................................................11 2.3 CLARITY OF CURRENT LEGISLATION ...............................................13 2.4 DIVISION OF TASKS ACROSS LEVELS OF GOVERNMENT ............14 2.5 CURRENT LG REFORM EFFORTS.........................................................15 2.6 KEY LESSONS...........................................................................................17 2.7 CHALLENGES AND EMERGING ISSUES .............................................18 2.7.1 Uganda ................................................................................................18 2.7.2 Tanzania..............................................................................................19 2.7.3 Kenya ..................................................................................................19 3. LOCAL ADMINISTRATIVE AND POLITCAL STRUCTURES...21 3.1 OVERVIEW OF THE LG UNITS ..............................................................21 3.2 POLITICAL STRUCTURES ......................................................................22 3.3 HIERARCHY OF LGS (LOWER LEVELS OF LGS)...............................24 3.4 LG STATUTORY BODIES........................................................................25 3.5 CENTRAL GOVERNMENT OVERSIGHT FUNCTIONS.......................26 3.6 LOCAL GOVERNMENT ASSOCIATIONS .............................................27 3.7 KEY LESSONS AND CHALLENGES......................................................29 4. LOCAL GOVERNMENT FINANCE.......................................................31 4.1 COMPARISON OF THE SYSTEMS OF LG FINANCE...........................31 4.1.1 Objectives, Strategy and the Overall System of LG Finance.........31 4.1.2 LG Expenditures................................................................................35 4.1.3 LG Expenditure Autonomy and the Budget Approval Process ....36 4.1.5 The System of Inter-governmental Fiscal Transfers......................40 4.1.6 Performance Incentives.....................................................................42 4.1.7 LG Borrowing and Debt....................................................................44 4.1.8 Financial Management Issues...........................................................45 4.1.9 Institutional Framework ­ LG Finance...........................................47 4.2 COMMON LESSONS AND CHALLENGES............................................48 ii Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda 4.3 EMERGENG ISSUES AND NEED FOR FURTHER SUPPORT.............56 5. HUMAN RESOURCES................................................................................58 5.1 LEGAL AND POLICY FRAMEWORK ....................................................59 5.2 TRENDS IN LG AND CENTRAL GOVERNMENT EMPLOYMENT....62 5.3 HUMAN RESOURCE CAPACITY AT THE LG LEVEL ........................62 5.4 SERVICE CONDITIONS AND HRD PRACTICES IN LGS....................64 5.4.1 Remuneration.....................................................................................64 5.4.2 Wage Bill and Pay Reform................................................................64 5.4 3 Personnel Management Practices.....................................................65 5.5 LG RESTRUCTURING..............................................................................66 5.6 CAPACITY BUILDING IN LGS ...............................................................67 5.7 KEY LESSONS...........................................................................................68 5.8 CHALLENGES AND EMERGING ISSUES .............................................70 6. INSTITUTIONAL ARRANGEMENTS ..................................................72 6.1 INSTITUTIONAL ARRANGEMENTS .....................................................72 6.1.1 Institutions for Coordination............................................................73 6.1.2 Government­Donor Coordination...................................................74 6.1.3 Decentralisation and the Wider Reform Agenda............................74 6.1.4 Coordination of Capacity Building ..................................................75 6.1.5 Joint Programming, Reviews and Evaluations ...............................75 6.2 LESSSONS AND CHALLENGES.............................................................76 7. CONCLUSIONS AND AREAS IN NEED OF FURTHER REVIEW AND SUPPORT......................................................................................................77 7.1 INCENTIVES FOR LG SERVICE DELIVERY AND DEVELOPMENT 77 7.1.1 General Overview of the Incentives .................................................77 7.1.2 Kenya ..................................................................................................81 7.1.3 Tanzania..............................................................................................82 7.1.4 Uganda ................................................................................................84 7.1.5 The Political Context of the Reforms...............................................85 7.2 CONSIDERATIONS ON FUTURE REFORMS AND SUPPORT............86 7.2.1 Kenya ..................................................................................................87 7.2.2 Tanzania..............................................................................................90 7.2.3 Uganda ................................................................................................92 7.2.4 Emerging Issues and Areas for Further Cross Country Studies...93 8. LIST OF LITERATURE...............................................................................95 9. ANNEXES........................................................................................................96 9.1 LIST OF STAKEHOLDERS CONSULTED..............................................96 9.2 THE LG STRUCTURES IN THE THREE COUNTRIES....................98 9.3 PROCEEDINGS FROM THE CONFERENCE IN ARUSHA.............101 iii Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda ABBREVIATIONS ABP Area-based Programmes ALAT Association of Local Authorities of Tanzania ALGAK Association of Local Government Authorities of Kenya ALGE Association of Local Government Employers (Kenya) B Billion (109) CAG Controller & Auditor General CAO Chief Administrative Officer CAS Country Assistance Strategy CB Capacity Building CBF Common Basket Fund CBFSC Common Basket Fund Steering Committee CBG Capacity Building Grant CBO Community Based Organisations CBPD Capacity Building for Participatory Development CBS Central Bureau of Statistics CC County Council or City Council CD Council Director CFO Chief Finance Officer CFS Consolidated Fund Services CCM Chama Cha Mapinduzi CG Central Government CGP Capital Grants Programme CILOR Contribution in Lieu of Rates CR Constitutional Review CRT Council Reform Team CSD Civil Service Department CSD-LGP Civil Service Department ­ Local Government Board CSO Civil Society Organisations DC District Council DDC District Development Committee DDO District Development Officer DDP District Development Project (Five District Pilot for LGDP) DDP District Development Plan DEB District Education Board DEO District Education Officer DFID Department for International Development DFRD District Focus for Rural Development DHMB District Health Management Board DHO District Health Officer DID Directorate of Institutional Development DLG Directorate of Local Government Co-ordination DMO District Medical Officer DP Development Programme DPM Directorate for Personnel Management DPRSC Decentralisation Poverty Reduction Credit DPTC District Planning Technical Committee DRC Directorate of Regional Co-ordination DRDP District Rural Development Programme (Dutch funded) DSPFD Draft Strategy Paper on Fiscal Decentralisation in Uganda DTB District Tender Board DTS Development Transfer System EC European Commission iv Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda EDF European Development Fund EG Equalisation Grant ESDP Education Sector Development Fund ESW Economic Sector Work EU The European Union FA Final Accounts FDS Fiscal Decentralisation Strategy FDTF Fiscal Decentralisation Task Force FMCB Financial Management Control Board FMS Financial Management System FY Finance Year GDP Gross Domestic Product GER Gross Enrolment Rate GoK Government of Kenya GoT Government of Tanzania GoU Government of Uganda GPT Graduated Personal Tax G-Tax Graduated Tax GTI. Government Training Institute HIPC Highly Indebted Poor Countries HLG Higher Level Government HOD Head of Department HPPG Harmonized Participatory Planning Guidelines HRM Human Resource Management IEC Information, Education and Communication IFMS Integrated Financial Management System IOB Interim Oversight Board IPF Indicative Planning Figure JGDR Joint Government Donor Review JICA Japan International Cooperation Agency KCC Kampala City Council KCRC Kenya Constitution Review Commission KLGRP Kenya Local Government Reform Programme KRA Kenya Revenue Authority KRB Kenya Roads Board KUPP Kenya Urban Poverty Project KUTIP Kenya Urban Transport Infrastructure Project LADP Local Authority Development Plan LAIFOMs Local Authority Integrated Financial Operations Management System LAs Local Authorities LASC Local Authority Services Charge LASDAP Local Authority Service Delivery Plan LATF Local Authority Transfer Funds LC Local Council LDG Local Development Grant (from LGDP) LG Local Government LGA Local Government Authority LGBC Local Government Budget Committee LGDP Local Government Development Programme (in MoLG) LGEA Local Government Expenditure Analysis LGFC Local Government Finance Commission LGLA Local Authority Loans Fund LGLP Local Government Loans Board LGR Local Government Revenues LGROC Local Government Releases and Operations Committee v Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda LGRP Local Government Reform Programme LGRT Local Government Reform Team LGSC Local Government Service Commission LGSP Local Government Support Programme LLG Lower Levels of Local Governments LM Line Ministry LRECC Local Revenue Enhancement Coordination Committee M & E Monitoring and Evaluation M Million (106) M&E Monitoring & Evaluation MC Municipal Council MFPED Ministry of Finance Planning and Economic Development MIS Management Information System MLG Ministry of Local Government MMO Manpower Management Officer MOEC Ministry of Education and Culture MoES Ministry of Education and Sports MOF Ministry of Finance MoH Ministry of Health MoLG Ministry of Local Government MoPS Ministry of Public Service MRPW Ministry of Roads and Public Works MTEF Medium-Term Expenditure Framework MTP Medium Term Plan MTR Mid-Term Review NA National Assessment NAO National Audit Office NAT National Assessment Team NCC Nairobi City Council NET Net Enrolment Rate NGO Non-Governmental Organisation NRM National Resistance Movement NUSAF Northern Uganda Social Action Fund O&M Operation and Maintenance O&OD Opportunities and Obstacles to Development OAG Office of the Auditor General OC Other Charges OPM Office of the Prime Minister PAC Public Accounts Committee PADEP Participatory Agricultural Development and Empowerment Project PAF Poverty Action Fund PCU Project Coordination Unit PE Personal Emoluments PEAP Poverty Eradication Action Plan PEDP Primary Education Development Programme PMA Programme for Modernisation of Agriculture PMO Prime Minister's Office PO-OPSM President's Office ­ Office of Public Service Management PO-RALG President's Office - Regional Administration and Local Government PRA Participatory Rural Assessment PRCF Poverty Reduction Co-Financing Fund PROLoGs Poverty Reduction through Optimising Local Governance Systems PRSC Poverty Reduction Support Credit PRSP Poverty Reduction Strategy Paper PSRP Public Sector Reform Programme vi Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda PSC Public Service Commission RAS Regional Administrative Secretary RDE The Royal Danish Embassy RGOZ Revolutionary Government of Zanzibar RGS Recurrent Grant System RIPS Rural Integrated Project Support Programme RMLF Road Maintenance Levy Fund RNE Royal Netherlands Embassy ROM Result Oriented Management RS Regional Secretariat RTB Recurrent Transfer Budget RTC Review Technical Committee RTS Recurrent Transfer System RWSS Rural Water Supply and Sanitation Project SASE Selective Accelerated Salary Enhancement SDP Support to Decentralisation Programme (UNCDF) SDU Support to Decentralisation Unit SFG School Facility Grant SNV Netherlands Development Organisation SSP Social Support Programme (under TASAF) SWAP Sector-Wide Approach SWG Sector Working Group TA Technical Assistance TASAF Tanzania Social Action Fund Project TB Tender Boards TC Town Council ToRs Terms of Reference TSC Teachers Service Commission UAAU Urban Authorities Association of Uganda UAPP Urban Authorities Partnership Programme UCG Unconditional Grant UDD Urban Development Department (of MLG) UGSH Uganda Shilling ULAA Uganda Local Authorities Association UNCDF United Nations Capital Development Fund UNCDF United Nations Community Development Fund UNDP United Nations Development Programme UNHS Uganda National Household Survey (1999/2000 ­ Report Jan 2001) UPPAP Uganda Participatory Poverty Assessment Project USAID United States Agency for International Development USRP Urban Sector Rehabilitation Programme VEO Village Executive Officer WB The World Bank WEO Ward Executive Officer ZRT Zonal Reform Team vii Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda ACKNOWLEDGEMENTS This synthesis report is based on the main findings and experiences and lessons learned from decentralisation as portrayed in the three country studies ­ "A Comparative Analysis of Decentralisation in Kenya, Tanzania and Uganda". As authors of this report we would like to express our deepest gratitude to the people who have supported the Study Team. We are indebted to a number of institutions and individuals for support received during the planning and implementation of the Study. First, our gratitude goes to the governments of Kenya, Uganda and Tanzania for the strong support through all crucial phases of the exercise. The authors would also like to extend thanks to key ministries and institutions (such as KLGRP, PROLoGs, LGFC, PCU/MoLG, LGRP) in relation to local government in the three countries and other key players. We are, furthermore, grateful for the support given by the Association of Local Authorities in Kenya (ALGAK), Tanzania (ALAT) and Uganda (ULGA) and the sample local governments for their invaluable contribution and time set aside for support to the Study. We would also like to thank the World Bank Team, especially Messrs Matthew Glasser and Lance Morrell, who provided useful backstopping, commented on earlier notes and papers, and contributed to the development of the terms of reference for the Study and to Ms. Ruth Mwikali, NCG-Intermedia, for editorial support. Many people, too numerous to mention by name, helped considerably in data collection, analysis and other contributions to this synthesis report. Great thanks to all. This report is based on information collected up to April 2004. It contains the views of the Study Team, which do not necessarily correspond to the views of the World Bank, Danida1 or the governments in Kenya, Tanzania and Uganda. Points of view and errors can be attributed only to the authors. Jesper Steffensen, Per Tidemand, Harriet Naitore (Kenya only), Emmanuel Ssewankambo (Uganda only), Eke Mwaimpopo (Tanzania only) August 2004 1The Study has been financed by a Danish Trust Fund in the World Bank. viii Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda 1. INTRODUCTION 1.1 BACKGROUND TO THE STUDY Background and Objectives This report provides a synthesis and comparative analysis of the main findings from the case studies of decentralisation in Kenya, Uganda and Tanzania. The detailed findings are available from the country studies2, whereas this synthesis report draws out the main findings and lessons learned and compares the experiences across the three countries. In recent years most of the Sub-Saharan African countries have undertaken comprehensive reforms of their public administrative systems, with decentralisation as one of the major tools to improve efficiency in public service delivery and to strengthen the involvement of citizens in the local decision-making processes3. Kenya, Uganda and Tanzania are amongst the countries that, for different historical, political and economic reasons, have embarked on the decentralisation reform process. Given the strong parallels between their political and constitutional histories, and the fact that some forms of innovative decentralisation programmes have been introduced in all three countries over the past decade, their relative experience of decentralisation is likely to be mutually instructive. Taken as a group, the experiences of these countries can illustrate a variety of necessary steps towards the establishment and refinement of strong local government systems ­ and new and innovative ways to organise and implement service delivery and good governance. They also enable a documentation of the difficulties, constraints and the future challenges. The World Bank has become deeply involved in supporting decentralisation in these countries. This support has taken a number of different forms, including assistance to the development of local government systems and decentralisation processes, technical assistance to the governments at policy and strategic levels, direct funding of inter-governmental fiscal transfers to local governments (LGs)4 for investment in local services and infrastructure, and capacity building for enhanced local management. 2Steffensen, Tidenmand, Naitore: "Country Study Kenya, A Comparative Analysis of Decentralisation in Kenya, Tanzania and Uganda" August 2004", Steffensen, Tidemand, Mwaimpopo: Country Study Tanzania, August 2004 and Steffensen, Tidemand, Ssewankambo: Country Study Uganda, August 2004. 3Please refer to Steffensen, Jesper and Trollegaard, Svend: Fiscal Decentralisation and Sub-National Government Finance in Relation to Infrastructure and Service Provision ­ Synthesis Report of 6 Sub- Saharan African Country Studies, NALAD, May 2000, World Bank/Danida/USAID for a detailed treatment of the experiences in Uganda, Ghana, Senegal, Swaziland, Zimbabwe and Zambia. 4In the country outlines the abbreviations "LAs" (Kenya) /"LGs" (Uganda) /"LGAs" (Tanzania) are applied for "local authorities /local governments". 1 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda In particular, the Bank's decision-making with respect to its expanding portfolio of decentralisation activities in Kenya, Tanzania and Uganda should benefit from a systematic comparative assessment of the dynamics and lessons of the ongoing process of decentralisation in the three countries With the preparation of the Sector Work on Local Service Delivery in Kenya, the Local Government Support Programme (LGSP) in Tanzania, and the Second Local Government Development Programme (LGDP II) in Uganda, the Bank's involvement in this area is set to widen over the foreseeable future. This regional study of the experiences of decentralization addresses the need to foster cross-national learning of best practices, identifies gaps and challenges, provides valuable inputs to a more informed future strategy on donor support and, hopefully, generates ideas for each involved country on the ways and means to strengthen the process. 1.2 SPECIFIC OBJECTIVES AND PURPOSE The basic objective of the work is to review and assess the key lessons of the decentralisation reforms in Kenya, Tanzania and Uganda. The purpose is twofold: i) to foster cross-national learning on experiences with decentralisation; and ii) to support and inform the World Bank's strategic approach within its expanding portfolio of activities in the sphere of decentralization in the three countries. It should be noted, however, that the work does not constitute part of a specific project/programme preparation in any of the three countries. 1.3 SCOPE OF THE STUDY AND DIMENSIONS TO BE REVIEWED Scope and Concepts The country studies have reviewed: (i) The basic nature of the decentralization process; (ii) The chief outcomes5 and experiences of this process to date; (iii)The degree to which incentives promoting effective and efficient local service delivery have emerged; (iv)The core constraints to the emergence of these incentives. "Decentralisation" is an ambiguous and broadly used concept, and the definition varies across countries and between actors within each country. In this Study 5 An impact study of decentralisation with the immense problems concerning the question of attribution, lack of defined service delivery standards, etc, is out of the scope of this Study. Please refer to the DAC Evaluation Series: Lessons Learned on Donor Support to Decentralisation and Local Governance, Schou, Arild and Steffensen, Jesper, OECD, DAC Network on Development Evaluation, 2004, for a review of the official evaluation reports from the DAC member countries on this subject, and to the discussions below. 2 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda decentralisation is defined as "devolution" of power and competence to independent governments below the central government level, which are given responsibilities (typically within certain levels and ceilings) for determining the level and quality of services to be provided, the manner in which those services will be provided, and the source and the size of funds to finance the delivery of those services.6 Decentralisation is seen as a gradual process. It takes different forms, including political decentralisation (transfer of decision-making power to lower-level, politically elected bodies) and fiscal decentralisation (assignment of functions and transfers of power within financing). It might take place within a specific sector7 or be integrated (multi-sectoral) as transfers of power to multi-purpose authorities. Decentralisation by devolution is the declared policy of decentralisation in Uganda and Tanzania, and it is also emphasized in the draft Constitution in Kenya. From a historical perspective, decentralisation has been introduced for various reasons: to answer the problems experienced with centralised/deconcentrated systems of service provision, to get political support, to achieve improved efficiency in resource allocation, to bring decisions closer the citizens, to improve governance and accountability, to improve equity and rural development, to improve the development and strengthen poverty reduction. But the design of decentralised systems, and many contextual factors, impact on the possibilities of achieving these objectives. The basic subjects for the analysis in this study are local governments (LGs), defined as the levels of government below the central government, which are accountable to local populations through some kind of an electoral process.8 In Kenya, these are called "local authorities" (LAs); in Tanzania: "local government authorities" (LGAs), and in Uganda: "local governments" (LGs). The outcomes are not understood as outcomes in terms of the "end product", i.e. poverty reduction and development9, but rather as intermediate results in the form of improved LG processes and institutional arrangements that are assumed as essential means for achieving such development outcomes. These processes and institutional arrangements include clarity in the division of roles and responsibilities among different levels of government, the interaction between politicians and staff, the situation within LG finance, expenditure and revenue assignments and priorities, 6 This is in line with the definition made by Richard Bird, Robert D. Ebel and Christine L Wallice (ed) in "Decentralisation of the Socialist State, Intergovernmental Finance in Transition Economies, IBRD, 1995, p.11-13. Contrary to this, "deconcentration" implies transfer of responsibility from central ministries to field officers at the local and regional level, thereby becoming closer to the citizens while remaining part of the central government system. 7 Some countries like Malawi and Nepal have started their decentralisation process within a few key sectors. 8 Other features of local governments (especially in the devolved type of LGs) are to a varying degree: democratically representative and autonomous bodies, with clear mandates to provide services, having body corporate status, with certain control over their staff, having access to funds and control over their own budgets and ability to enforce by-laws, cf. a definition in UNCDF publication: "Taking Risks", p. 168-170, 1999. 9 Please refer to R Crook and A. Sverrison : "Decentralisation and poverty alleviation in developing countries ­ a comparative analysis or Is West Bengal Unique? IDS Working Paper No. 129, May 2001" for a review of the links between decentralisation and poverty reduction ­ and to the DAC Evaluation Series (A. Schou/J. Steffensen): ", 2004, op cit, which contains a review of the existing literature, and concludes that further research is needed to explore this link. 3 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda financial management performance, the capacity of the staff and politicians to perform, and the interaction with the citizens. The Study takes its point of departure in the countries´ own declared decentralisation objectives, as outlined under the first dimension of the review (see below). It is important to emphasise that there is no uniform standard model of decentralisation applicable to all countries across the world or within a region. Various approaches to decentralisation have been initiated with different speeds, forms, content and procedures. And decentralisation cannot be seen as the sole answer to all challenges in public reform. Under the right circumstances it may lead to more efficient service provision, bringing the decisions closer to those affected, improving competitiveness and innovation, enhancing participation, accountability and ownership. However, its success will depend on many factors, such as the country context, its history, the political will, and the design of the administrative and financial systems. Nevertheless, some of the major pre-conditions must be in place for decentralisation of government power actually to happen. The analysis of similarities and differences in country-specific decentralisation experiences should help in identifying practical solutions and in making existing systems of central and local government relations more efficient, effective and accountable to the citizens. A particular focus of this analysis is the extent to which individual and institutional incentives have been provided for the LGs to perform effectively. These incentives, individual as well as institutional incentives, may be influenced by economic as well as non-economic factors and may come from "above" (e.g. central government influence) as well as from "below" (e.g. downwards accountability and pressure from the citizens) ­ or a combination of both. Special attention is paid to systems and procedures that have had an impact on these incentives, and to the other factors that have promoted and constrained the development of strong incentives to improve the LGs' ability to deliver the expected results. Dimensions Explored The decentralisation processes in each country have been examined through the "lenses" of four main dimensions that, together, constitute some of the key pillars in any system of decentralisation: 1. Legal, constitutional and overall policy framework; 2. Local administrative and political structures; 3. LG finance and inter-governmental fiscal framework; 4. LGs relationship with and access to adequate human resources. 4 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Figure 1.1: Key Pillars for Decentralisation Decentralised system of service delivery assignment Legalframew acco Struc Fina untab ture nc of ility sfor ingof Humanresourc func tions ork­ pol itical LGs esinLGs Institutional arrangements for coordination of reforms A fifth dimension ­ the institutional arrangements surrounding decentralisation ­ has been examined, although in less detail, in order to provide ideas related to the changes in roles of various stakeholders, and to areas in need of further coordination and collaboration. This is typically a factor that is overlooked.10 However, proper coordination of the activities of the actors involved in reform is crucial for making effective linkages between the components of the decentralisation process itself and with the wider public sector reforms ­ especially sector and civil service reforms. It is important to emphasize that that these dimensions are strongly inter-related and mutually interdependent. A very unbalanced decentralisation process may lead to problems ­ for instance, a comprehensive transfer of financial resources and fiscal autonomy to LGs without reforms of the administrative and political structures, and without LG control over the human resources, may lead to management problems. Also, problems can occur when there is a transfer of strong decision-making power and functional responsibilities, without enabling LGs to manage/adjust part of the funding. A strategic planning, phasing and sequencing of the reforms is therefore critical. 10This is often neglected but may be one of the greatest deficiencies in most decentralisation efforts. Please refer to Paul Smoke in: "Decentralisation and Local Governance in Africa", Public Administration and Development, Volume 23, Number 1, February 2003, for a discussion of this issue. 5 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda The focus on these main dimensions will enable us not only to review elements of decentralisation already in place and the lessons from these, but also to identify areas that have not yet been fully addressed ­ and ways and means to move forward. The intention is to identify elements that are crucial in the development of a decentralised system of public administration. Within these overall five dimensions, the following main issues have been examined: Table 1.1: Dimensions and Main Issues Dimension Main issues 1) Legal, constitutional and policy · Historical development and context dimensions · Governments´s decentralisation objectives · Basic legal framework ­ enabling · Overall division of tasks and functions across layers of government 2) Administrative and political dimension · Overall structure of the system of LGs (number, layers, etc.) · Political structure of LGs · Oversight institutions and functions · Role of the associations of local authorities · Role of the statutory bodies 3) Fiscal dimension · Assignment of expenditures (profile and trends) · Assignment of revenues (composition, type, etc.) · Match between expenditure and revenue assignments · Administrative issues within revenue mobilisation · Fiscal autonomy · Financial management and accountability issues · Institutional issues and coordination in the field of LG finance 4) Human resource dimension · Local government HR capacity · Civil Service conditions and incentives · LG autonomy in HR management and accountability issues · Efforts of restructuring LGs · LG capacity building efforts 5) Institutional arrangements and · Coordination of the decentralisation coordination process · Donor-government coordination · Inter-donor coordination · Role and functions of various stakeholders · LG reform in the wider context of public sector reforms 6 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda 1.4 THE STUDY APPROACH AND THE TEAM Regional approach This exercise is a regional study of selected Sub-Saharan East African countries ­ Kenya, Tanzania11 and Uganda. The three countries are gradually expanding their co- operation in many areas, through the strengthening of the East African links, most recently with the adoption of the East African Customs Union Protocol. Given their various historical and country-specific experiences, their geographical closeness, and their strong cultural, economic and political linkages, a "benchmarking" of their experiences is of particular potential for cross-country learning and identification of good practices. In addition, a comparison of those three countries should assist in identifying factors that are facilitating and constraining the achievement of the development objectives within decentralisation, as well as common problems and challenges to be addressed by the governments and by supporting agencies. In order to ensure a productive comparison of the experiences and lessons learned, a common structure for the study approach ­ questionnaire, data collection, selected dimensions, structure of the country reports ­ has been applied. A similar format has been applied in this synthesis report, in order to ease references to the more detailed country studies and to enable comparisons across the three countries.12 Constraints and data collection As the scope of the Study has been rather limited in terms of input (three people each with two weeks of field work13 and a few days for reporting per country) it has been important that it focuses on the main trends in the development and on the key factors that have promoted and restrained the process ­ rather than on attempting exhaustively to cover all issues within the area of decentralisation. This is one of the reasons why the Study has been restricted to the five above-mentioned dimensions. It has left other important areas less well covered, such as the links between public and private sector, the links to civil society with respect to downwards accountability, and various donor interventions. It has also identified a number of areas in need of further more detailed research. It is a qualitative study, and quantitative data is mostly applied concerning the system of LG finance to enable comparisons of the main features of the system. To a great extent, it is based on existing documentation ("desk findings" from reports and studies). It has involved a focused dialogue with key stakeholders from central and local governments, in order to understand the factors that have facilitated and constrained the realisation of the decentralisation objectives. In each country, the findings have been "tested" in the field, through visits to a few selected LGs. During the field-work, interviews and discussions have been conducted with senior central 11The initial overview of Tanzania includes Zanzibar, whereas the later discussions of Tanzania focuses on Tanzania Mainland. Please refer to Annex 2.1 of the Country Study for further details. 12Please refer to the country studies for further details: i) Country Study Kenya, August 2004, ii) Country Study Tanzania, August 2004 and iii) Country Study Uganda, August 2004. 13The field-work in Tanzania was carried out during November 2003 and February 2004, in Uganda during March 2004 and in Kenya during March 2004. 7 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda government and LG officials (politicians and staff), donor agencies and other stakeholders. Data-availability has been a real challenge in all three countries, especially within the field of LG finance and human resource capacity. Generally, all quantitative data needs to be treated with due caution. However, specific sources, constraints and challenges in the data are indicated in each section/table. The Study Team The Study has been undertaken by small group of consultants: Jesper Steffensen (Team Leader) from the Nordic Consulting Group, Denmark, Per Tidemand (independent consultant), Harriet Naitore (PriceWaterhouseCoopers) for the study of Kenya , Emmanual Ssewankambo (Mentor Consult) for the study of Uganda, and Eke Mwaimpopo (Nordic Consulting Group Tz) for the study of Tanzania. The consultants have referred to the project management team composed of Lance Morrell, TTL of the AFTU1´s ESW in Kenya (EW-PO78597-ESW) and TTL of the LGDP II in Uganda (PE-PO77477 LEN) and Matthew Glasser, TTL of the LGSP in Tanzania (PE PO77477-LEN). 1.5 THE SYNTHESIS REPORT This Final Synthesis Report, which has incorporated comments received for the Draft June Report, will be presented at a regional workshop in September 2004. The report is divided into seven chapters ­ this Introductory Chapter and then each of the following six chapters focusing specifically on a particular decentralisation dimension. However, linkages between the various dimensions are referred to whenever they enrich the review. Each chapter contains: i) a comparison of the systems (benchmarking), ii) a review of the common and individual lessons learned and iii) some emerging issues and future challenges. The report only draws upon the main findings of the country studies (please refer to these for further details). Chapter Two contains a comparison of the vision and objectives behind decentralisation ­ and the overall nature of the decentralisation process. It also deals with the legal framework, especially the Constitution, the Local Government Act and the links to other sector-relevant laws and regulations. Chapter Three compares the LG administrative and political structures, the various institutions and their mutual links and relationships. It briefly deals with the links to the civil society, and ongoing and expected reforms within administrative and political structures. Chapter Four provides an overview of the systems of LG, the inter-governmental fiscal framework in the three countries (including links with, and coordination between, the central government (CG) and the local governments). 8 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Chapter Five deals with the LG human resources and the administrative capacity, LG autonomy in HR management, general trends in LG and public sector employment, LG capacity building and restructuring processes. Chapter Six provides a brief overview of the institutional framework for decentralisation, including a description of the coordination of the process, the links between the government, the donors and other stakeholders. Finally, Chapter Seven contains a summarised analysis of the factors that have promoted and restrained the decentralisation process in the three countries. There is a special focus on the incentives for the LGs to perform efficiently towards a realisation of the development goals. It also gives a brief outline of some of the main considerations that might be addressed in the future reforms by the Governments and the development partners. The chapter also contains some concluding remarks based on the findings from the previous chapters. Important questions to be discussed are: (i) To what degree have incentives emerged, promoting effective and efficient local service delivery? (ii) Which factors have promoted efficient development of proper local authorities´ incentives to perform, in terms of their legal and service providing obligations, and other obligations? (iii) Which factors have constrained the development of an efficiently functioning system of decentralisation? Annexes and the Regional Conference The report has three annexes: 1) A list of stakeholders consulted, 2) an overview of the structure of LGs in the three countries and 3) a summary of the proceedings from the working groups under a high policy level "Conference on Decentralisation and Local Government in Eastern and Southern Africa", arranged in Arusha, 22-23 September 2004 where the report and the country studies were presented by the consultancy team and the findings and the recommendations discussed. The findings and recommendations from the working groups with high-level officials and politicians from ten Eastern and Southern African countries generally endorsed the results from the Study. 9 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda 1.6 COUNTRY PROFILES Country Profile ­ Fact Sheet Issue Kenya Tanzania Uganda Population 28.7 Million (1999) 33.6 Million (2002) 24.7 Million (2002) Size of the territory 580,400 Km2 945,100 Km2 241,000 Km2 (surface Km2) GDP (2002) 12.3 Billion USD 9.4 Billion USD 5.8 Billion USD GDP Growth 1998: 1.6 % 1998: 3.7 % 1998: 4.9 % 2001: 1.1 % 2001: 6.1 % 2001: 5.1 % 2002:1.0 % 2002: 6.3 % 2002: 6.7 % GNI pr capita (2002) 360 USD 290 USD 240 USD Tax revenue of GDP 22.2 % (2002/03) 12.3 % (2003) 11.7 (2002/03) Present decentralisation 1995 1998 1992 reform start 2004 Draft Constitution Service/development Life expectancy 45.5 43.1 43.1 Literacy rate (>15 years) 84.5 77.1 68.9 Infant mortality (1000) 78.0 104.0 83.0 Under 5 mortality (1000) 112.0 165.0 141.0 Governance system Multi-party (strongly Multi-party (strongly Non-party political electoral dominated by two parties) dominated by one system at all levels of party) government Layers of government 2 layers 3 layers 3 layers In addition, a number of Central government and CG, HLG and LLGs administrative units LAs CG, HLG (urban and Parallel system of districts) and LLGs (sub- provincial and district countries/town councils) administrations Number of local 175 10,168 1,034 governments with legislative power Average size of the upper 163,923 293,525 358,681 layer of LG CG: Central Government, LG Local Government, LLG: Lower levels of Local Government, HLG: Higher levels of local government Source: Data on GDP and economy: World Bank Development Indicator Database, 2004. PER, Uganda (2003), PER Kenya (2003), and PER Tanzania FY 03. Other data is from the country studies on Kenya, Tanzania and Uganda. 10 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda 2. LEGAL AND POLICY ISSUES 2.1 INTRODUCTION In the discussion of legal and policy issues in the three countries we will initially briefly describe four key issues before proceeding to a discussion of main lessons and challenges: 1. History of LGs in the three countries: commonalities and differences. 2. The current LG legislation with an emphasis on its clarity in assignment of responsibilities to LGs and the extent to which the LG system is entrenched in the Constitution. 3. The relationship between sector legislation and LG legislation: the extent to which a clear division of roles and responsibilities is in place. 4. Current initiatives for LG reforms. 2.2 HISTORICAL BACKGROUND The three East African countries share a great deal in the historical development of their LG system. Foremost, a similar colonial history with most of the colonial period characterised by various forms of undemocratic and indirect rule by the British, followed by a gradual introduction of elected LGs in the period approaching independence ­ to a great extent as a measure for the colonial administration to control African aspirations for rapid transfer of powers from the colonial government. At independence, all three countries inherited rather strong LGs with substantial responsibilities for services. In Uganda, this even took the form of a functional federal system. In Kenya, a proposed regional system was an entrenched part of the new Constitution the country started off with. In all three countries the decentralised systems were, with different speeds, abolished after independence. The nation building exercises in all three countries required such "departicipation"14 ­ which also included the abolition of multi-party politics. While a full abolition of elected LGs only occurred for brief periods, the elected LGs have in all three countries and in most of the immediate post-colonial period, played only marginal roles in development administration. Major changes occurred when there was a reintroduction of LGs in the 1980s in Uganda and Tanzania ­ followed by a process of devolution and strengthening of LGs through the 1990s and onwards. However, these reforms occurred for very different 14Concept from Hyden, 1983 p.47 following Kasfir, 1976. 11 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda reasons. The NRM came to power in Uganda in 1986 after a protracted guerrilla war. It initiated a radical reform, of which the reform of LGs became a core of both the political and administrative transformation. In Tanzania, the LG reforms grew more cautiously out of the wider political and economic liberalisation that took place from the mid 1980s. These changes, initially, only occurred on the Mainland, and Zanzibar took longer to consider a reform of its public administration. Zanzibar is part of the United Republic of Tanzania, but it maintains its own LG system that is quite separate from the Mainland ­ in the two sections below we very briefly describe some of the characteristics15. In Kenya, the government has, until recently, mainly pursued a policy of deconcentration, with only a marginal role for LGs. However, in the recent (March 2004) draft Constitution, proposals are made for a substantial devolution of powers to elected LGs. In all three countries, the general move towards political liberalisation and experienced inability of the centralised system to provide efficient local services were the major general reasons for the recent LG reforms. The history of LGs can thus be characterised as a pendulum movement, whereby institutional arrangements for local-level service delivery over the last 40 years have alternated between systems based on devolution and centralised planning. Table 2.1: Key LG Policy Changes and Related Legal Benchmarks Policy Change Kenya Tanzania Uganda Mainland Zanzibar Introduction of 1950 African Local 1944 1949 Local democratic local District Government Act introduction Government governments Councils of 1953 of LGs Ordinance during late Ordinance colonialism Devolved systems 1963 1962 Local Weak system 1962 Constitution at independence Constitution Government devolved powers outlined in legislation to Kingdoms and detail a system extending LGs with strong modern LGs regions and LGs nationally Abolishment of "Transfer of Rural LGs 1964 1967 Local devolved systems Functions Act" abolished in Revolution Government Act. ­ LG decline and in 1969 reduced 1971 and urban merged party Decree no 2 of deconcentration the powers of LGs in 1972 and state 1971 LGs substantially New 2004 Draft 1998 LG 1995 LG 1993 decentralisation Constitution Reform Policy, legislation Decentralisation policies 1999 and 2000 and Policy, amendments to 2003 Good 1995 LG Act Governance Constitution, Strategy 1997 LG Act 15Annex 2.1 to the Tanzania Country Outline gives an elaborate description of the LG system in Zanzibar. 12 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda 2.3 CLARITY OF CURRENT LEGISLATION Current LG legislation in the three countries differs significantly in terms of clarity and the extent it is constitutionally entrenched. It can be observed that Uganda has by far the most clearly outlined LG legislation, which furthermore is embedded in great detail in the Constitution. In Kenya, the LG Act has remained relatively unchanged for a long period. It gives LGs very limited mandates as deconcentrated administrations, and sector ministries provide most of the local services. In Tanzania Mainland, a reform has been embarked upon, but not yet fully reflected in revised legislation. The Tanzanian reforms do not include Zanzibar, where LGs play a rather marginal role and operate in parallel to strong regional and district administrations. The key characteristics of the LG legislation in the three countries can be summarised as follows: Table: 2.2: Key Characteristics of LG Legislation Legislation Kenya Tanzania Tanzania Uganda Mainland Zanzibar Key LG The Local The Local Act to establish 1997 Local legislation Government Act Government the Zanzibar Government (Cap 265) of Act 1982 and MC 1995; Act, with minor 1963 with various Act to establish amendments numerous associated Acts; the district and 2002 amendments up amended 1999 town councils to 1998 and 2000 1995 LG system is No, Yes, Yes, Yes. entrenched in but 2004 draft but only very but Constitution Very detailed Constitution Constitution does briefly, is it of Zanzibar only description. include a mentioned that mentions briefly Constitution significant there shall be that there shall (1995), also chapter LGs be LGs prevents amendments without endorsements by the LGs LG legislation Yes, No. No. Yes. compiled in but several sector Described in six Three separate Very detailed one compre- laws that pieces of pieces of and succinct Act hensive LG Act overrule LG Act principal LG legislation in regulate LG legislation addition to functions legislation for regional administration etc. LGs have Burial of Functions LGs' mandate Yes. LGs have clearly defined destitute persons described in very vague and key response- functions is the only rather broad overlapping bilities for mandatory and vague with deconcen- primary 13 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Legislation Kenya Tanzania Tanzania Uganda Mainland Zanzibar function. terms ­ few trated district education, Multiple other mandatory and regional agriculture, functions are functions administrations water and permissive sanitation, primary health care, roads, among others Each level of Not applicable Very unclear Not applicable Generally, yes. LGs has clear below district Some suggested responsibilities level areas for devolution at the discretion of districts; however not fully clear and concise on the LLG functions. LLGs= Lower levels of local governments 2.4 DIVISION OF TASKS ACROSS LEVELS OF GOVERNMENT Major service provision responsibilities are devolved to LGs in Tanzania and Uganda, whereas LGs in Kenya have very limited service delivery mandates. The same limited role for LGs also applies to Zanzibar. While LGs have a rather clear service delivery mandate by law in Uganda, they are in practice facing central government control through such measures as the conditional grant system (see chapter 4) and parallel central government institutions such as the Technical Support Units in the roads and water sectors. The situation within each of the major local service delivery sectors is summarised in the table below. Table 2.3: Extent of Devolution of Key Sector Responsibilities to LGs Sector Kenya Tanzania Uganda Mainland Education Minor role. Primary education in Primary education fully Seven of the major principle devolved ­ devolved to LGs; LGs are designated however teachers secondary education still as "education recruited by TSC. Yet no with central Government authorities"; the specific role in secondary remaining LGs play education no major role in provision of educational services Health No major role by LGs responsible for LGs responsible for LGs ­ mainly primary health care. primary health care and undertaken by MoH Hospitals managed by district hospitals health boards 14 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Sector Kenya Tanzania Uganda Mainland Water Largely centralised Rural LGs mainly Rural water largely with with Ministry of responsible for rural LGs. Urban water capital Environment and water. Urban water investments managed by Natural Resources managed by autonomous centre and O&M by (MENR) authorities. Capital water boards. National Water investments largely Newly established Conservation and managed centrally Technical Support Units Pipeline Corporation by central Government (NWCPC). However tend to operate in parallel some LGs operate to LGs water boards Agriculture No major role for LGs are main responsible LGs are main LGs responsible, but current efforts are made for privatising services Roads No major role for District and feeder roads District and feeder roads LGs ­ centralized and all municipal roads and all municipal roads with creation of Road the responsibility of LGs the responsibility of LGs Board. Only a few LGs have recently been appointed as road sub-agents 2.5 CURRENT LG REFORM EFFORTS The LG reform efforts vary substantially among the three countries. Uganda is characterised by a comprehensive decentralisation reform based on devolution of authority to elected LGs and an elaborate multi-tier LG system. The reform was officially launched in 1992, and it has for several years been guided by Presidential and other high level announcements, and technically spearheaded by a committed secretariat. Substantial legal reforms were undertaken and are today prominently reflected in the Constitution, as well as in a clear LG Act. The reforms have been characterised as radical in their focus on devolution to LGs and have over the years been very high on the political agenda of the NRM Government, as the reforms have been politically as well as technically motivated. Since the reforms matured in 1997, and since the finances of LG budgets increased mainly by transfers (see chapter 4), some centralising tendencies have emerged as line ministries have sought control over the implementation of transfer of funds to LGs. The reforms are today at a crossroads, where the Government wishes to take stock of experiences as it is considering a review of fundamental elements of the system in connection with the ongoing Constitutional Review. This may include the introduction of a regional tier of government above the districts and a centralisation of the appointments and management of the Chief Administrative Officers in the districts. The Government of Uganda has today no updated long-term strategy for its decentralisation and LG reforms, but it is planned to develop such a strategy within the coming months. The Government is in the first phases of implementing a comprehensive Fiscal Decentralisation Strategy (FDS), which aims to address the challenges within the existing inter-governmental fiscal transfer system and revenue mobilisation, and has 15 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda launched a new LG capacity building framework. Please refer to chapters 4 and 5 for further details. Tanzania is in the middle of a substantial reform programme. The reform in Tanzania grew out of an ongoing Civil Service Programme in the 1990s, and it is part of a wider governance reform that also included the introduction of multi-party politics. A policy for LG reforms in Tanzania was endorsed in 1998 and an operational programme for reform subsequently launched. The reform programme has undertaken substantial training and introduced a process of restructuring LGs as well as undertaken some legal reviews. LG Acts have been amended, just as a process of harmonising sector legislation is ongoing. However, this has yet to be translated into effective devolution of finances and personnel to LGs. Legal harmonisation has encountered several problems ­ recently manifested in the passing of a Public Service Act (2002/03) that overruled efforts for devolution of personnel to LGs. Zanzibar is not part of the mainland LGRP, and it has only very recently taken steps for initiating reforms. Kenya has since the mid-1990s initiated an incremental reform of LGs that, foremost, has focused on improving the fiscal aspects of LGs without, to date, substantial legal reforms. However, as part of the ongoing Constitutional Review, much wider and very radical proposals related to decentralisation reform has been debated. The draft Constitution was endorsed by the National Institutional Conference on 15 March 2004, but, at the time of writing, it awaits the final measures for adoption. It proposes a radical devolution of powers to a LG system based on three tiers: regional, district and location. If the Constitution is adopted, it will signal the start of a comprehensive decentralisation programme based on devolution. In anticipation, the Ministry of Local Government is currently in the process of drafting various LG laws and amendments. Key aspects of the current LG reforms are summarised in Table 2.4. Table 2.4 Current LG Reform Initiatives Kenya Tanzania Tanzania Uganda Mainland Zanzibar Key guiding 2004 1998 LG LG reforms No overall policy Constitution ­ Reform Policy considered as strategy, but a documents not yet approved part of Zanzibar Fiscal Internal draft Good Decentralisation concept papers Governance Strategy, 2002 have been Strategy 2002 (FDS) is being elaborated implemented Reform Main objectives Main emphasis Mainly Primary emphasis objectives (Section 206 of on improved emphasis on on enhanced Constitution) of service enhanced local local democracy the reform to delivery. governance and and participation. foster democratic However democratisation. FDS emphasis on local governance governance Secondary rationalisation objectives emphasis on and efficiency stated in efficiency gains gains and Constitution from abolition improved of parallel accountability structures 16 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Kenya Tanzania Tanzania Uganda Mainland Zanzibar Operational KLGRP and LGRP with a Zanzibar Good LGDP II funded programmes Prologs current Governance by GoU, WB, for LG supported by approved work Strategy 2002 is Danida, reforms DFID programme to partially funded Netherlands, June 2005. by RGoZ, Austria and Funded by UNDP and Ireland Aid. GoT, and a DfID Parallel and large group of coordinated bilateral donors support by DFID, plus EU and USAID and UNCDF others FDS (2002-) 2.6 KEY LESSONS All three countries abolished, during the initial years of independence, several aspects of the LG systems ­ if not entirely scrapping the LGs. However, experience proved that local level service delivery can be best undertaken with a clear role assigned to democratically elected LGs. All three countries also now see democratically elected LGs, with substantial autonomy and powers, as key policy objectives to be pursued. In Uganda and Kenya this is entrenched in the Constitution16 and in Tanzania it is clearly stated in the political manifesto and reform policies. In pursuit of legal and policy development in support of devolution, the following main lessons have emerged: · The importance of clarity of LG legislation; in particular, the assignment of responsibilities. The current Kenyan and Tanzania LG legislation only describe LG mandates in rather vague formulations, with preciously few mandatory functions for LGs. The Uganda LG Act on the other hand has a clear schedule of LG responsibilities, which furthermore is entrenched in the Constitution. In Uganda, some challenges still remain regarding the precise division of roles and responsibilities between the many layers of the LG system. · In the absence of clear LG mandates, the responsibilities for local service delivery become muddled, and adequate re-sourcing of LGs also becomes difficult to establish. · The importance of harmonisation of sector legislation in line with LG legislation and stated policies on devolution. This, in turn, has proved impossible to achieve without a clear LG Act and strong political leadership and guidance. In Kenya, for instance, the LG Act is thus inferior to other Acts in the sense that discrepancies between what is outlined in the LG Act and what is outlined in sector legislation has always been interpreted in favour of the latter. 16In the case of Kenya, the Draft 2004 Constitution yet to be adopted. 17 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda · A clear and detailed description of the LG system in the Constitution. This has proved an effective tool for coordination of sector legislation as indicated in the case of Uganda. Such entrenchments in the Constitution may also protect the LG system against sudden ad hoc changes. On the other hand, too detailed a description of the system may also hamper necessary legal amendments. It may, for instance, not be appropriate to have the specific formula for grant allocations stipulated in the Constitution, as practical experiences should allow for adjustments. · The importance of a clear, coherent decentralisation strategy and implementation plan, based on broad dialogue and involvement of all key stakeholders, currently reviewed and up-dated to guide the overall reform process, especially the links to the sectors. 2.7 CHALLENGES AND EMERGING ISSUES The three countries currently face very different legal and policy challenges in the pursuit of reform: 2.7.1 Uganda Uganda represents a maturing devolved system of governance, where the government is now suggesting a rethink of policy and LG legislation. Two issues are of particular relevance: 1. Possible introduction of a regional tier and forms of federal governance; 2. Possible centralisation of some elements of personnel management, specifically, the appointment of the CAO. These two issues have been raised partly as a result of substantive political and cultural aspirations for greater autonomy of Buganda, in particular, and partly as a result over concerns regarding the current levels of autonomy of LGs. The two issues are thus partially contradictory: claiming more autonomy for greater units on the on hand, while suggesting a higher degree of central government control over LG affairs on the other. The proposals present some grave risks for the future of devolution in Uganda, as neither are based on substantial technical grounds: the regions are mainly pushed for political reasons without substantial consideration for the assignment of responsibilities and resources for this tier. The possible centralisation of some elements of LG personnel administration constitutes a more fundamental threat to decentralisation by devolution for two basic reasons: first, a central government appointment of the CAO will cut the most crucial accountability link between the elected local councils and the LG staff; second, such an approach to solving administrative problems in LGs (through centralisation) may result in a snowball effect, whereby local accountability mechanisms become totally undermined. 18 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda The planned development of a decentralisation strategy, based on a wide dialogue between CG and LGs and various stakeholders, may assist in guiding the future process. 2.7.2 Tanzania Tanzania has embarked on a comprehensive LG reform, which since 1998 has been guided by the Local Government Reform Policy. An overriding challenge in Tanzania has been how to translate the rather clear and radical policy into law and reformed procedures and practices. The policy is very detailed in terms of the scope of transfers of powers and resources to LGs, yet sector legislation and, in particular, legislation for the public service, contradicts or is unsupportive of the policy. Attempts are being made for a harmonisation of sector legislation, and some progress can be noted; yet fundamental issues remain, especially pertaining to the Public Service Act. Other legal and policy issues in Tanzania relate to the division of responsibilities between different layers of LGs. The wards and villages in Tanzania remain with rather weak and undefined mandates17, just as the role of the central government institutions for oversight of LGs (the Regional Secretariats) is still widely considered as unsatisfactory. 2.7.3 Kenya In Kenya, the current legal framework of LGs is widely recognised as in need of reform. The existing parallel system for local-level service delivery gives no clear mandate to LGs for the provision of key services. Instead, the various sectors and the deconcentrated administrations are given the main responsibility. Studies, confirmed by the field visits, have indicated that this not only leads to duplication and poor local governance, as citizens cannot hold government accountable at the local level for local level service delivery, but it also entails a waste of public resources and an ineffective provision of services. The draft 2004 Constitution proposes a radical reform in support of devolution to LGs. The proposed Constitution has been criticised for being too complex, as it suggests the introduction of a three tier LG system, where only the second tier (the districts) have well defined mandates for service delivery. The implementation of the proposed transition towards a devolved system will certainly be very challenging ­ and possibly painful. Many lessons can be drawn from Uganda and Tanzania in this respect. Previous structures for deconcentrated service delivery will have to be abolished; staff employed by sector ministries will have to be transferred to local governments, if they are to undertake their new functions. Modalities for sufficient (discretionary) funding need to be developed. The transfer scheme, LATF, has generated substantial experiences that can form the basis for further reforms (cf. Chapter 4 for further details). The most critical aspects will probably be: 17It should also be noted that the villages in Tanzania are rather small, with on average only 3,300 inhabitants. 19 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda · Clear assignment of specific responsibilities to the different new LGs, in line with the broad guidance of the draft Constitution and considering factors like linkages between the size of the government units and the type of services to be provided, links between the services, economies of scale, possibilities for citizens' participation and accountability, political representation, closeness, etc; · Reform of key sectors along the lines stipulated in the Constitution, with the active and constructive involvement of sector ministries; · Development of a system for balancing the new transferred functions, with properly assigned revenues to LGs, · Development of a decentralised system for management of personnel that guarantees meaningful local accountability of staff, while at the same time safeguarding the interests of personnel in terms of career development prospects and job security. 20 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda 3. LOCAL ADMINISTRATIVE AND POLITCAL STRUCTURES 3.1 OVERVIEW OF THE LG UNITS The table below provides a brief description of some of the key features of local governments (LGs) in the three countries. As it appears, there is considerable variation. Please refer to the Annex 3.1 for a detailed overview. Table 3.1: Local Government Structures Issue Kenya (1) Tanzania (2) Uganda (3) Tiers of LGs One layer Two layers (with Two layers legislative power in Rural: district ­ sub-county rural LGs, and one Urban: municipality-division layer in urban LGs) Number of all LGs 175 10,168 1,034 with legislative power 67 counties, one city 114 HLGs + 55 districts, one city council, council (CC), 62 town 10,054 LLGs 857 sub-counties, 13 councils (TCs), 45 municipal councils, five city municipal councils divisions and 34 municipal (MCs) divisions Of which no. of 108 22 122 urban LGs Of which no. of 67 92 912 rural LGs Number of the 175 114 69 upper layer of LG (only one layer) 92 districts, two CC, 12 55 districts, one CC, 12 MCs MC, eight TC Average size of the 163,923 293,525 358,681 upper layer of LGs Number of LLGs Not applicable 10,054 965 LLGs 10,045 rural villages + 857 sub-counties, 69 TCs, nine townships. five City Divisions, 34 (The 2,757 urban mitaa municipal divisions are not real LGs) Sub-ordination There is variation in the The upper level Not formally, but informally status of the LGs but coordinates the a strong support/mentoring generally no functions of the lower and coordination from the subordination levels upper layer of LG Central government Eight provinces (one is 21 Regional No administration units. deconcentrated Nairobi) Secretariats units Districts (66) + sub- District Commissioners There is a district resident districts and locations Division level DSs representative (coordination (deconcentrated units) functions) Trend Proliferation and Gradual increase in the Gradual increase in the gradual increase in the number of districts number of districts 80s. The coming reform is supposed to reduce the number of LGs 21 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda (1) Total population in Kenya: 28,686,607. (2) Tanzania: In addition there are 2,040 ward development committees, which have coordination and planning functions. Census from 2002: 33,461,849 inhabitants (3) Uganda: Census from 2002: 24,748,977 inhabitants. There are five layers of local administration in the rural areas: district, county, sub-county, parish and village, but only the district and the sub-counties are genuine LGs as corporate bodies. Layers of Local Governments Both Uganda and Tanzania have several layers of LGs, whereas Kenya has only one. In addition, Uganda and Tanzania has a number of lower levels of administrative units with political representation (although these are not corporate bodies and not defined as LGs) involved in decision-making, planning and project identification. The draft Constitution in Kenya proposes several layers of LGs as well and there is a discussion of the functions of LGs and other administrative units. All three countries have separate systems for urban and rural authorities, with different statuses, functions and funding systems. The upper level of LGs (districts/municipalities/cities/counties) are rather big in international comparison and are potentially able to capture economies of scale and efficient service provision, although the LGs in Kenya are smaller and the future reforms anticipate a reduction of the number towards larger units. Proliferation of LGs All three countries have experienced a proliferation of LGs over the last decades, which have put a further pressure on the stretched LG administrations, as key staff have been divided across additional administrations. There is a real risk of further fragmentation of limited resources in the name of "decentralisation", undermining the efficiency and the sustainability of the LG system. 3.2 POLITICAL STRUCTURES The political LG structures share a number of common features in all three countries: · The highest political decision making body is the full LG council (district council, county council, town council or municipal council, etc). · This council is composed mainly of directly elected councillors from specific constituencies (wards) ­ and, for example, not through proportional elections. In Kenya, the political parties have nominated additional members (maximum one third of the council). In Uganda, special electoral colleges elect representatives for women, youth and disabled persons. Similar arrangements are in place in Tanzania for women members. · The more regular operation of the LG Councils is conducted through a structure of standing committees. These vary slightly among the three countries, but generally involve a rather active Finance and Planning Committee, plus several other standing committees. These commonalities all derive from the common colonial history of the three countries, when LGs were based on the British system. 22 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda However, the political LG structures in the three countries also differ in important aspects. Key differences are summarized in the table below. Table 3.2: Key Differences in LG Political Structures Political Structure Kenya Tanzania Uganda Election of Mayor Elected by the Elected by the Directly elected and Chairpersons council council Party politics Yes Yes No Councillors have to Councillors have to Councillors are not stand for a party stand for a party allowed to campaign on a party basis MPs Not members, Members of the LG No role in LGs although they play council a significant role Full time None None Yes councillors The Executive Committee As evident from table 3.2, the executive political LG structure in Uganda is stronger than in Tanzania and Kenya. The Uganda LGs are led by directly elected Mayors and Chairpersons. These in turn select an Executive Committee (endorsed by the councils), which runs the day-to-day business of the LG. The benefits from such a system are, supposedly, a more substantial direct political control of LG affairs. The costs and benefits of the changes made in 1997 are, as far as the Study Team could ascertain, not documented. However, the system poses some risks of monopolized political democracy within a small and stronger group. The LG councillors in Tanzania and Kenya are theoretically playing a much less direct role in LG administration. From field interviews, staff suggested that in practice, LG councillors "interfere" substantially in administrative affairs ­ however, without the formal mandate and procedures as found in Uganda. Training of politicians in their roles and functions in Uganda and Tanzania has generally improved on the situation Members of parliament have traditionally played a substantial role in LG affairs ­ something that was changed in Uganda with the 1997 LG Act, but still prevails in Kenya and Tanzania. The inclusion of MPs in the LGs may provide a useful opportunity for sharing of information. However, experience suggests that MPs tend to dominate discussions and interfere in LG politics in a manner that does not provide local accountability, but mainly serves the MPs interests for providing local patronage. The interference of MPs in local service provision issues (rather than national policy issues) has in Kenya been further aggravated by the introduction of very substantial funds controlled by MPs for local capital development and road maintenance. This has impacted on the possibilities to make coherent local planning and in some places created conflicts between the LGs and the constituents impacting negatively on the emerging local accountability. 23 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda LG elections in Kenya and Tanzania take place within a multi-party political framework. In Kenya, this happens within a highly contested political climate, and various areas are typically dominated entirely by one or other single party. The sitting governments have, according to all accounts, subsequently penalized local councils dominated by opposition parties. In Tanzania, the elections have been very dominated by the ruling party (CCM), to the extent that the electorates in some areas could have felt they had a very limited choice in their LG elections. In Uganda, candidates for LG elections are barred from contesting on a party basis. However, in recent elections it was fairly obvious, which candidates were supporters or opponents of the "Movement". Kampala City has generally voted against the "Movement candidates" ­ an indication of the possibilities for political competition within the non-party system. 3.3 HIERARCHY OF LGS (LOWER LEVELS OF LGS) Tanzania and Uganda have a substantial number of LLGs with their own budgets and revenues (from revenue sharing with the HLGs). A distinction is made between real lower level LGs (corporate bodies, with powers to raise their own revenues, pass by- laws, approve budgets and plans) and administrative units. Administrative units have no similar powers, but they are supportive structures to the LGs. Table 3.3 Lower Level LGs Type of Unit Kenya Tanzania Uganda Full lower None Village local Sub-county local local governments (10,045 governments (rural). governments + nine townships) Divisions (cities/municipalities). Town councils (Total: 965 LLGs) Administrative None Ward Development Counties (151). units Committees (2040). Parishes/wards (5,225), Mtaa (2,757) and Villages (44,402) Vitongoji (57,137) Tanzania only has fully fledged LLGs in the rural areas: the village local governments with rather low average populations (around 3,300 people). The villages are corporate bodies, which may pass by-laws, raise revenue and pass their own budgets, etc. Ironically, the village governments were introduced in Tanzania during the period in the 1970s when LGs were abolished. So they came about as community level structures within a deconcentrated local administration and with an internal organisation based upon the ruling one-party structure. When LGs were reintroduced in 1982, the village governments became part of the LG structure, and they were potentially strong organisations for community level participation in LG affairs. In reality, their sheer size yet weak revenue base has made it difficult to build up the capacity of these units and assign them real service delivery functions. The village governments have hardly any financial resources and only one staff member: a Village Executive Officer who, from 2004, is paid from central government. Other 24 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda units, such as the ward development committees that have a higher population base and some technical staff, have no legal basis for functioning as LGs. In Uganda, the larger size of the sub-counties/town councils (approx. 24,000 people on average) has made them able to take over some functions within primary education, basic health and agriculture extension. The sub-counties have in recent years been up-graded substantially in terms of staffing, and they now maintain a core staff of several graduates. The sub-counties also retain the majority of locally generated revenue, as well as 65% of the discretionary capital grant. Thus, they can play a substantial role in local level service delivery. The principle in the linkages between the layers of LGs has been "integration with sub-ordination", meaning that the activities of the lower layer are coordinated and harmonised (e.g. planning and budgeting) with the higher LGs, without direct hierarchical control by the higher levels. In addition, the HLGs support and mentor the LLGs in their functions. In Kenya, discussions are now ongoing regarding the future structure of a devolved governance system. This includes discussion regarding the establishment of LGs at locational and possibly village level. Concerns have been raised regarding the rationale for establishing so many structures, and the current draft Constitution only presents proposals regarding LGs at the locational level. In Uganda and Tanzania, there has been a push towards further decentralisation of tasks to the LLGs, along the principle of "subsidarity"18. However, as the size of the LLGs units are larger, and the possibilities effectively to perform the increased responsibilities are better in Uganda, it has gone further in this process. 3.4 LG STATUTORY BODIES In Uganda and Tanzania a range of statutory boards have been established under the LGs. This is to ensure some level of relative autonomy of these institutions, but also to ensure substantial linkages with the LGs. They include: · District Service Commissions (in Uganda) and Recruitment Boards (in Tanzania); · Tender Boards (Uganda and Tanzania); · The Local Government Public Accounts Committee (Uganda); · Land Boards (Uganda). Similar structures have not been established in Kenya, where, for example, LG procurement is undertaken directly by a tender committee composed of the top level LG staff19. The gradual growth of such statutory bodies in Uganda is a reflection of 18According to this internationally recognised principle for assignment of tasks and functions, these should be assigned to the lowest possible level of government that can handle these responsibilities in an efficient and professional way dependant on their capacity, professional skills, size of the unit, etc. 19According to the procurement rules from 2003, the tender committee is composed of staff from the LG, without representation of the councillors. However the field missions revealed that this has not yet been fully acknowledged by the LGs. 25 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda the increasing transfer of functions to the LG level, combined with a concern not to involve the political administration too much in the daily administration. 3.5 CENTRAL GOVERNMENT OVERSIGHT FUNCTIONS There are substantial differences in the links between the LGs and the central government in the three countries. In Tanzania and Kenya, the CG has rather large offices and staff at all levels of government ­ and representatives of CG down to the lowest levels. In Kenya especially, there are parallel mechanisms for service delivery from the deconcentrated and decentralisation systems. In Uganda, this was previously the case, but the present system has only the appointed Resident District Commissioner, who is specifically responsible for coordinative functions and for overseeing those functions that are not decentralised (such as security, defence, law and order). Table 3.4: Summary of Key LG Oversight Institutions Structures Kenya Tanzania Uganda Deconcentrated Provincial Regional Resident District oversight Administrations and Commissioners Commissioners with structures District supported by limited secretariat ­ Administrations, Regional mainly responsible for with substantial Administrations of security issues. technical staff some 35 delivering services professionals and Some sectors introduce in parallel to LGs ­ District Technical Support Units rather than Commissioners ­ on a "regional" basis functioning as responsible for oversight institutions oversight and support. Some sectors have officers at the regional /local level (water/education) Central oversight MLG ­inspects LGs PO-RALG inspects MoLG inspects LG structures LGs and provides adherence to law and There are examples through RS support regulations ­ and of MLG interferes and regular provides mentoring. directly in several supervision. Auditor General. LG issues and MLG MoFPED, Treasury has powers to Auditor General Inspection. approve or reject LG budgets. Ministry of Finance IGG Auditor General Sector ministries' supervisory and Ministry of Finance monitoring departments. In some sectors: Technical Support units. 26 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda In Uganda and Tanzania, the development of appropriate central government institutions for oversight and support of LGs under a decentralized set-up has proved to be a substantial challenge. In Uganda, earlier zonal and provincial institutions have been abolished, but a regional tier of LG is now considered ­ partly to ensure "coordination". Some sector ministries, such as Water and Roads, have established regionally-based technical support units to ensure central government oversight of, and technical support to, LGs. Numerous inspection and monitoring teams have also arisen, and improvements in the coordination and streamlining of these oversight institutions have being requested by LGs. In 1997 in Tanzania the strong regional administrations were almost abolished and transformed from directly implementing organization to organizations with the primary mandate of technical support and oversight of LGs. The practical transformation of the RSs has proved very difficult to undertake in practice, and it is widely recognized that the 21 regional secretariats in Tanzania currently performing these functions are quite unsatisfactory for a number of capacity and institutional reasons. MLG in Kenya inspects and monitors the matters of LGs. In addition there is an audit of the LGs´ financial management, however the capacity is generally weak in this area. There are numerous examples of direct central government interventions in specific matters of LGs, and the CG has most of the direct service provision responsibilities, rather than LG supportive functions. 3.6 LOCAL GOVERNMENT ASSOCIATIONS As documented in various evaluation literature, associations of local authorities can play important roles in enhancing knowledge and information about and across LGs, mediating conflicts of interests between CG and LGs, and sharing experiences, for example, in replicating pilot schemes and capacity building of LG staff and politicians.20 Below is a comparison of some key features of the associations of LGs in the three countries: Issue Kenya Tanzania Uganda Association of Association of Local Association of Local Uganda Local Authorities LGs Government Authorities in Tanzania Association (ULAA) and Authorities in Kenya (ALAT) Urban Authorities (ALGAK) Association of Uganda (UAAU) Number of staff 10 5 22 in the secretariat Eight professional and Vacant financial mgt, A joint secretariat between two support staff legal, HR, advocacy UAAU and ULAA has been specialist positions as established with 17 staff. In per April 2004. addition five staff in UAAU, covering all main professional positions Budget and Approx. 200,000 NA Approx. 700,000 USD funding (FA 2002) Membership fee and (FA 2002/03) 20DAC/OECD: A. Schou and J. Steffensen, op cit, p.54 27 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Issue Kenya Tanzania Uganda Subscription fee cover donor contribution the operational budget and donors the specific activities Own share 50 % NA 13 % funding by sub- Remaining from donors Subscription and donor Remaining part from scribtion, sale etc. support cooperation partners Main functions Protect the interests of Protect the interests of Protect the interests of LGs, LGs. LGs. advocacy, negotiating. Advocate and negotiate Forum for discussion Membership in all fora on LG on behalf of LGs vis-a- and exchange of reforms (LGFC, LGBC, vis CG. knowledge. LRECC etc.). Support CB of LGs. Liaise, consult with CG Information to members and Dissemination of and deliver advice sharing of experiences. information to/on LGs. Development of working Development of tools for LGs, e.g. code of working tools for LGs, conduct. e.g. handbook for Capacity building of LGs. councillors and training International projects and materials exchange Impact on LG Weak, but increasing, Weak, but increasing. Strong and increasing. reforms especially increasing Not sufficiently Participates in all major fora recognition by CG of involved in many of the on LG reforms, especially on ALGAK, reflected in reforms. LG financial matters. the Constitutional Major decisions often For instance, the ULAA has Review and in the taken without involve- recently negotiated with all LATF process ment of ALAT, e.g. on LMs on the sector policies taxes and allocation formulas Major challenges Inadequate human and Many vacant positions Stagnant in membership financial basis. in the secretariat. contribution due to Not sufficiently Inadequate human and deteriorated LG revenue recognised by the CG financial basis. mobilisation. in many issues Not recognised by the Future sustainability of the CG in many issues large amount of activities, with a decline in fees, is a challenge. Merging of the two associations will enhance the efficiency, but is challenging As it appears, the Associations of LGs in Uganda with their highly facilitated and strong secretariats have also had most impact on the policy decisions, and on the operational level (dissemination of information, submission of newsletters, library functions, capacity building of politicians, etc.). The associations in Uganda have received a strong support from some of the main cooperation partners (Danida, DFID, the World Bank, and EU), which have supported in building up their capacity. The Study Team finds that the impact of the associations in all three countries within policy reform and advocacy is increasing, and the functions are increasingly recognised by the major stakeholders, also at the CG level. All the associations are facing severe financial challenges, as the membership fees are either stagnant or declining due to financial problems in many membership LGs. 28 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda 3.7 KEY LESSONS AND CHALLENGES Derived from the comparison of the experiences, the following lessons and challenges appear to be of critical importance: · There has been a major proliferation in the number of LGs in parallel with decentralisation of functions, staff and funding to LGs. This has put additional pressure on the LGs' capacity, and in some place led to fragmentation of resources for building up a professional cadre of staff and funds for development investments21. Reviews of the "optimal" size of the LGs have seldom been done, considering factors such as the type of functions they are supposed to carry out, economies of scale, possibilities for bundling of services, distance from citizens to service and politicians, possibilities for participation and political representation. This is an area of future importance for the efficiency of the LG system. In Kenya, these considerations are urgently required as part of the implementation of the coming Constitution. It is also important as a basis for the discussions on the introduction of a "regional" tier in Uganda. In Tanzania, there is a wish to decentralise more functions to the lowest levels of LGs; hence, detailed analytical work on the advantages and disadvantages of this process for service delivery, participation, good governance, etc. should be considered. · Establishment of several layers of government in medium size countries like the three studied has been proved to have certain advantages in terms of prospects for improving participatory planning and budgeting; involving citizens in decision-making; bringing the service provision closer to the citizens. However, experience has also shown the importance of having clearly delineated tasks and functions for each layer; having a certain size of direct grant funding (or reserved funds for each layer in terms of indicative planning figures) as revenue sharing is shown to be problematic (cf. chapter 4); and having clear rules and procedures for interaction, communication and support/mentoring. · As LGs are provided with more substantial service provision responsibilities, it has been proved essential to transform the internal structures of LGs. Thus in Uganda, the LG councillors' capacities for political oversight have been strengthened by enhancing the executive elements of the councils. The power of MPs has correspondingly been reduced. · The role of the MPs in the LGs needs clarification. Direct MP responsibilities for local service provision may conflict with the LGs possibilities to make efficient local priorities and reduce the local councillors´ possibilities to build up strong accountability links to their constituents. 21It should be noted that in many other parts of the world the trend is to amalgamate LGs in order to increase the economies of scale, to ensure professional administrations and to enable these to address the future challenges with IT, communication requirements, competition, and, gradually, more complex service delivery requirements, especially within the water and health services. 29 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda · The architecture of statutory LG bodies has developed in tandem with the transfer of powers to LGs. Kenya, Tanzania and Uganda, in this respect, thus represent stages in such a development. To ensure the right balance between local political accountability of all relevant decisions, without over politicising local administrations, continues to be a major challenge. · The abolition of the strong CG regional and local control in Uganda and Tanzania has been welcomed by most stakeholders, especially the LGs. And has been reported to have led to more self-reliant and responsible LGs. However, the role of the RSs in Tanzania is still to be further defined, narrowed, and the remaining functions sufficiently capacitated, if these are to have a proactive and supportive role. The links between the CG and LGs in Kenya will have to be completely rearranged as part of the reform of the entire public administrative system. One of the lessons learnt from the three countries is the importance of clearly defined (limited) functions of CG regional authorities. These may be instrumental in ensuring certain cooperation of the service delivery outside of the decentralised areas, a platform for communication across authorities, and they may have some delineated legal supervisory functions. However, they should not interfere with the details in the LG priorities and current day-to-day operations. · The studies have documented the importance of having a strong coordination of the oversight functions, especially the relationship between the Ministry of Local Government, Ministry of Finance, the Audit and the institutions responsible for the follow-up on irregularities in order to reduce the transactions costs for LG (avoid overlaps) and strengthen legal compliance. · Associations of local authorities/local governments have a great potential for playing an important facilitating role in the decentralisation reform process, especially in areas of protecting and advocating for decentralisation reforms of benefit for the LG development, dissemination of information and best practices, and the development of specific working tools for the politicians and staff to improve performance (e.g. handbooks and guidelines). This has been the case ­ to various degrees ­ in all three countries. The LG associations in Uganda have, in particular, had a strong impact on many of the recent reforms of fiscal decentralisation, and they have played a major role in capacity building of the politicians and in information dissemination. The associations in the three countries have many of the same functions, but the experience has shown that there is a great challenge in mobilising subscription fees from the members with a low LG revenue base and with many other urgent funding needs. In this case, is not sufficient to show results at the policy level in the metropolitan areas/central government "corridors", as this may not be fully recognised by all LGs when the tough local priorities should be taken into account ­ subscription payment to the association or drugs for the health units. A long lasting support from the cooperation partners is probably needed, with a firm and realistic strategic plan to ensure a gradually increasing sustainability and own financing. 30 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda 4. LOCAL GOVERNMENT FINANCE This chapter reviews the main features of the systems of LG finance in the three countries and the experiences from recent LG fiscal reforms. The design and implementation of systems of LG finance are of critical importance for developing incentives for LGs to improve on efficiency in service provision, financial management and good governance.22 4.1 COMPARISON OF THE SYSTEMS OF LG FINANCE 4.1.1 Objectives, Strategy and the Overall System of LG Finance In all three countries studied, the development of an efficient system of LG finance is currently on the top of the agenda, with discussions of the fiscal decentralisation objectives and strategies, major tax reforms ­ as well as reforms of the inter- governmental fiscal transfer and financial management systems. However, the systemic fiscal reforms have typically been initiated some time after the legal and structural reforms, described in chapters 2 and 3, and they have been delayed in all three countries. The reforms have typically been initiated as answers (reactions) to specific problems, rather than conducted strategically in parallel/in tandem with the overall decentralisation of tasks and functions. The present activity level is generally high. Below is a comparison of a few indicators across the three countries: Table 4.1: Comparison of the Overall Policy and Legal Issues on LG Finance Issue Kenya Tanzania Uganda Fiscal No No Yes decentralisation There are some However, the first important However, this is mainly strategy declared policy steps have been taken, as focusing on the reform of objectives within the Policy Paper on LG the transfer system specific areas Reform, 1998, contains the overall principles The fiscal decentralisa- tion strategy (FDS) was A draft Sector Policy for the adopted by the Cabinet devolution of the for testing in June 2002 development budget has to answer numerous been elaborated and is problems with the pending final decision existing grant system (lack of autonomy, PO-RALG/LGRP has multiple and non- planned to elaborate a coordinated modalities comprehensive fiscal for reporting, accounting decentralisation strategy, etc.) and to introduce although no draft has yet better performance been produced. The incentives and preparatory work on the accountability measures LGSP and the recurrent 22Please refer to the study: Devolution in Pakistan - An Assessment and Recommendations for Action, 28 May 2004, Asian Development Bank, DFID and the World Bank for a description of the theoretical framework behind incentive systems in the field of LG finance. 31 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Issue Kenya Tanzania Uganda grant reform work provides a good platform for such a policy. Legal issues on The LG Act is rather The LG Act is rather Very detailed description LG finance detailed on the fiscal detailed on the fiscal issues. of the LG finance system issues. ­ both in the Constitution Significant degree of CG and the LG Act. High degree of CG control of the revenues, control of the revenues. although there is some room Some room for local for local adjustments. autonomy, although the Detailed budget LG revenue sources guidelines from CG. Detailed budget guidelines assigned have been from CG. criticised for being of a The draft Constitution low-yield and non- contains overall buoyant nature. principles on LG finance, incl. equitable In practice, most services share of revenues are financed by across levels of govt. conditional grants. Many of the recent reform initiatives are regulated by (budget) guidelines. Major recent Introduction of a Block July 2003: Abolishment of a Implementation of the reforms and Grant system (LATF) range of LG tax FDS (2003- 2006) changes in 1999/2000, with assignments encompassing additional strong incentives to flexibility in the improve performance Changes in the conditional conditional grant system, within areas such as recurrent grant system reduction in the number debt recovery, financial towards formula-based, of grants, improved management reporting transparent and predictable allocation criteria, new and participatory transfers with more LG and simplified modalities planning and autonomy within utilisation for budgeting, accounting budgeting. from FY 2004/05. and reporting. Reform of the Business Reform of the GoT financed Draft bill for reform of Licence into a Single development transfer the property tax is LG Business Permit system towards a formula- pending. system. based more equitable distribution system. Introduction of inter- Emerging reforms of governmental fiscal the IT /accounting Design of a system for coordination bodies system, with piloting of performance-based system (budget, accounts and a computer- based for capital grants and revenue enhancement). financial management capacity building, based on system clear allocation formulas Initiatives to boost the and in-built incentives for LG revenue mobilisation Budget reform: CG LG performance efforts ­ sharing of best impose ceilings on LG practices, training, new budgeting of revenues incentives in the grant to enforce more systems, etc. realistic budgets and movement towards a Introduction of the situation with "hard "second" generation of budget constraints" the LGDP, with strong linkages between transfers of development funds, LG incentives to 32 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Issue Kenya Tanzania Uganda improve administrative performance and capacity building support Use of pilot Yes. Yes. Yes. schemes Concerning the IT Under the LGRP, 38 LGs FDS was initially piloted introduction and were selected as the first within only 15 LGs. phasing in of the Single batch of reform councils Business Permit. and special support has been The devolution of the rendered to these, e.g. development budget has within financial also been phased in with management. a gradual roll-out to more and more LGs. Will also be applied in the introduction of the LGSP A computerised IFMS development transfer system is also being scheme, and may be used in piloted in a few LGs the decisions concerning the level of autonomy under the recurrent grant system CG= Central Government, LG= Local Government Strategy on fiscal decentralisation There has been an increasing awareness of the importance of development of fiscal decentralisation strategies. Uganda, based on the experience of the problems with the conditional grant system by the end of the 1990s, elaborated a comprehensive and detailed strategy (the FDS, which was adopted in 2002). In Tanzania, a strategy is planned for, but the basic pillars have already been agreed upon related to the reform of the recurrent transfer scheme and the coming devolution of part of the development budget. However, there is a need for a more holistic, all encompassing, future fiscal decentralisation strategy. In Kenya, a strategy has been awaiting the overall reform under the constitutional review process. In the meantime, the initiatives have focused on improving specific components of the fiscal system, such as development of the basic block grant system, streamlining of some of the taxes, budget reform to introduce measures to improve on the realism of the LG budgeting, improved financial management and debt recovery. A strategy is envisaged when the final decisions on the future Constitution have been taken. Although the extent to which the strategies have been formulated explicitly and comprehensively varies greatly, there are many similarities across the three countries when it comes to concrete policy objectives under the reform programmes. For instance, all three countries have formulated objectives under their reform programmes to increase LG revenue mobilisation, to improve the legal framework on LG taxation, to improve on the transfer system (improve on the predictability, timeliness, allocation criteria), to link transfers to certain performance incentives and the LG capacity to handle funds, and to improve on the LG financial management and accountability. However, the system and means vary significantly, (cf. below). 33 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda The significance of the LG finance The extent to which fiscal decentralisation has taken off varies greatly across the countries. Some of the indicators, which cannot be seen in isolation23, are shown below: Table 4.2: Key Figures on LG Finance (Budget 2002/03) Issue Kenya Tanzania Uganda GNI per capita (2002)* 360 USD 290 USD 240 USD Total public expenditure as share of GDP (1) 27.0 % 23.3 % 23.7 % LG share of the GDP (2) (B 2002/03) 1.5 % 4.7 % 5.7 % LG share of the total public exp. (3) (4) 5.1 % 19.2 % 27.0 % (B2002/03) Total LG expenditure per capita (USD) (3) 6.7 USD 12.1 USD 17.4 USD (B 2002/03) * World Bank Development Indicators for GNI (2002 figures). Data drawn from the country studies on Kenya, Tanzania and Uganda, August 2004. A number of assumptions have been used within the calculation of figures, and all figures should be treated with due caution as data availability is generally limited, cf. sections 4.1 ­ 4.2 of the three country studies. "B" = budgets. (1) Source: Republic of Kenya: Public Expenditure Review, p.1 and country studies. (2) Estimated by use of figures from the Annexes to the country studies. (3) The figure has included all donor projects. (4) Re. Uganda: the figure for the LG share of the total public expenditure will increase to 36.3 % if the donor projects are excluded, cf. Annex 4.1 of the Country Study for Uganda. The figure has included estimates of the expenditure related to own source revenues (104.8 B UGSH, equal to the 2001/02 level) as revenue data for FY 2002/03 is not fully available. If own source revenues are excluded, the expenditure per capita for Uganda is 15. 1 USD. Exchange rates estimates: Tanzania: 1 USD = 1000 Tshs. Kenya: 1 USD = 77 Ksh (2002), and Uganda: 1 USD = 1800 UGSH (2003). A review of the relative significance of the LG sector of the total public economy cannot be seen in isolation from the assignment of expenditures to LGs. However, other factors play a role as well, especially the way in which funds are routed from the centre to the final point of service delivery. The issue of expenditure assignment (division of tasks across levels of government) was discussed in Section 2. But it is worth mentioning the fact that LGs (except a few of the larger municipalities) in Kenya are not responsible for the key sector areas ­ health and education ­ and that many of the functions within roads, water, as carried out by agencies other than the LGs (line ministries, water boards, etc.). LGs in Tanzania and Uganda have more similar functions. However, there are two main factors, which should be kept in mind in evaluating the figures: i) the devolution of the development budget (capital investment and capital grants) was started 3-4 years earlier in Uganda, and ii) many of the funds flowing to the LGs and related authorities at the local level in Tanzania are "off-budget" and not registered as LG expenditures. This goes for many of the district support programmes24 and sector programmes, which in Uganda have been gradually mainstreamed and channelled through the GoU budget and account procedures. The sector capital grants in Uganda (health, education, roads, water and agriculture) constitute alone 2.3 USD per capita, and the non-sectoral development grant (LGDP-I) around 1.0 USD in B 2002/0325. These two 23 Indicators on LG fiscal autonomy, including share of own source revenues, etc, are equally important, but will be treated in the subsequent sub-sections. 24 Named Area Based Programmes in Tanzania. 25 2 USD in the urban authorities. 34 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda factors seem to explain most of the variations in the LG expenditure per capita between the two countries. It should be noted that in there are programmes with links to LGs in all three countries, which are not included in the LG budgets and accounts (off-budget), e.g. the NUSAF in Uganda, TASAF in Tanzania. 4.1.2 LG Expenditures The analysis of LG expenditures has been constrained by lack of consolidated, up- dated and reliable databases in the three countries (definitely an area in need of further strengthening). However, the available data shows the following main characteristics: Table 4.3: Comparison of the LG Expenditure Profiles Issue Kenya (1) Tanzania (2) Uganda (3) Recurrent share of 83% + 70-80 % 81% total LG exp. (6 % in debt payment) Major Sectors -Education Not broken down in 59% 42% -Health sectors, but 50% was 16% 15% -Water/roads spent on personnel and 5% 14% -Administration civic exp. in FY2001/02 19% 24% -Others and only 32% on LG - 4% operations +maintenance Share of the NA, but the costs on *19% 24 % administrative costs personnel is approx. 50 % Main priorities from LATF (2001/02): DRDP (2002) LGDP-I (1998-2001) non-sector specific Vehicles: 22% Education: 32% Roads /drainage: 43% grants (discretionary Roads: 20% Finance/planning: 20% Education: 23% grants) Water and San: 14% Works: 15 % Water/san.: 14% Council premises: 9% Agric./natural res. 13% Health: 13% Markets 9 % Water 8 % Production 5% Health 7 % Adm.: 1% Others: 5 % The table is based on the three country studies (various data banks, sample data and field missions). The figures are rough estimates and should be treated with due caution. (1) Country Study for Kenya, FA 2001/02. (2) Cf. Annex 4.2 in the Country Study. A) Recurrent share: Part of the funds classified under development for PEDP (education) and the Health Basket Fund is recurrent, i.e. the real recurrent share of expenditure may be > 80 %. Sector composition: The composition of the conditional grants is applied as a proxy. *Based on a sample of 12 LGs. The group "administration" includes some other functions. (Cf. the Country Study, Annex 4.3). The discretionary funds for own decision-making are very limited until the LGSP has taken off and the district support programmes are brought on-budget and on accounts. (2002 figures). DRDP: As there are no significant discretionary grants in Tanzania the DRDP (Dutch funded programme) is used a proxy. It provides resources for discretionary development funds and their utilisation may give an indication of the local priorities, although it cannot be compared directly with the CG grants. (3) Based on the Country Study, Annex 4.2 (data from the Ministry of Finance, Macroeconomic Dept.) Budget FY 2002/03. Administration also encompasses the cost of the political boards and functions. In all three countries, the share of the recurrent expenditure of total expenditure is very high (>70-80%). Uganda initiated a decentralisation of the devolution budget some years ago, and the results are emerging. It is expected that the capital 35 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda expenditure as a share of the total expenditure will increase in Kenya and Tanzania as well with the increase in the LATF monies in Kenya and the start of the LGSP (2004/05) in Uganda, which will encompass significant discretionary funds for capital investments. Along with the increase in capital investments, a pressure is envisaged on the related operational and maintenance costs, a challenge which is already great in Uganda under the LGDP. The table shows, that when LG are given autonomy on utilisation in discretionary development funding, combined with strong supervision and incentives, they tend to invest largely in the national priorities areas. 4.1.3 LG Expenditure Autonomy and the Budget Approval Process Control over the LG budgets and priorities takes various forms in the three countries. The table below shows some of the main "check and balances": Table 4.4: LG Autonomy on the Budget Preparation Process Issue Kenya (1) Tanzania (2) Uganda (3) Conditional Limited, >95% of the grants 88% of the grants sector -as most transfers are specific LATF block grants grants Budget Yes. No formal approval. No. approval Two stages, at the In reality - yes. The LGFC reviews provincial and central The advice from the RS and submit comments ministerial levels is typically adhered to to the LGs on the BFPs. The sectors approve the grant work-plans before releases. Budget Yes. Yes. Yes. guidelines However, there is Detailed budget There are general, as significant room for LG guidelines well as detailed, priorities on the sector grant expenditures guidelines BFP: Budget framework papers. Most of the control of LG expenditures to ensure adherence to national targets and priorities in Tanzania and Uganda is done through the conditional grant system. In Tanzania, there is a dialogue at the regional level between the Regional Secretariat of PO-RALG and the LGs about the priorities and adherence to national guidelines, but the budgets are adopted by the LG councils. However, evidence shows that the advice from the RS and, subsequently, the PO-RALG is taken seriously and has an impact. In Uganda, most of the influence on priorities is carried out through the detailed conditional grant system, with numerous budget lines and conditionalities. Kenya has chosen another route, with a system of non-sectoral grants (block grants), coupled with tight controls and approval procedures of the general budgeting process (budgeting of revenues, debt recovery, etc.). In all three countries, there has been a noticeable improvement in the prior announcement of transfer figures from CG and improvement in the budget guidelines and timing. This is reported to have had a positive impact on the LG budgeting 36 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda process. However, there are still a number of challenges within this area, as identified in section 4.2. 4.1.4 LG Revenues Composition of LG revenues and own source revenues The size, composition and autonomy of the LG revenues vary greatly across the three countries. The table below shows some of the main indicators: Table 4.5 Overview of the LG Revenues Issue Kenya (1) Tanzania (2) Uganda (3) CG domestic tax rev. 99.0 USD 31.9 USD 31.8 USD pr. capita (B 2002/03) (B 2002/03) (FA 2002/03) Size of the total LG 4.9 USD 12.1 USD 16 USD revenues per capita (FA 2001/02) (B 2002/03) (FA 2001/02) Size of the LG own 3.5 USD 1.6 USD 2.4 USD source revenues per (FA 2001/02) (FA 2002) (FA 2001/02) capita Composition of LG revenues a) Transfers 26% 79% 85% b) Own source rev. 73% 16% 15% c) Shared rev. 1% 5% 0% (FA 2001/02) (B 2002/03) (FA 2001/02) Type of LG own Water sewage fees. Development levy, Graduated tax. revenues (major Property tax. 2002/0326. User fees and charges. sources) Single Business Per. Service charges. Property tax. Market fees. Business licences. Licences House rent Produce cess. Property tax Autonomy on own Limited. Limited. Limited. source revenues Strong control of the Strong approval process The rates of the major rates and levels, through the by-laws. taxes are fixed by CG. approval process. LG involvement and LG involvement and LG involvement and influence on tax adm. influence on tax adm. influence on tax adm. Development in LG Stable. Stable/modest increase. Decrease. revenues over most But the changes in tax But recent initiatives recent 3-4 years assignments will properly seem to have reversed lead to a significant this trend. Share of own source decrease. revenues of total Share of total Share of total revenues has Share of total revenues revenues revenues has declined declined has declined Utilisation of tax Great un-tapped Great potential, Great potential, potential potential, especially especially from property tax especially from in property taxation in the urban centres property tax Revenue sharing Not applicable Yes. Yes. between HLGs and Guideline min. 20% 6 % of collected local LLGs (2003/04)* revenues retained at the LLGs FA: Final Accounts/Out-turns, CG = Central Government. B= Budget, HLG: Higher levels of local government, LLG: Lower levels of local government. 26Abolished by July 1, 2003. 37 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Source: Country Studies for Kenya, Tanzania and Uganda, sections 4.6-4.8, op cit. and PER for the three countries (2003) (2) Data: Country Study for Tanzania. The composition of revenues will change significantly from FY 2003/04, with the recent abolition of the development levy and certain other taxes. *Tanzania: From FY 2004/05, 50% of the grants in lieu of the abolished local taxes will be allocated for the villages/mtaa on the basis of population as indicative planning figures and from 2005/06 directly to the LLGs. (Draft Budget Guidelines, 2004-04). Exchange rate: TZ: 1 USD = 1000 Tsh., 1 USD = 1800 UGSH (2002), and 1 USD = 77 Ksh (2002). The table above shows that Uganda has by far the highest total per capita LG revenues. However, LGs in Kenya generate significant more own source revenues than the LGs in the other two countries, approx. 50% more per capita than in Uganda and > 100% more than LGs in Tanzania. However, the water fees and sewage fees account for a large share of these revenues, and without this type of revenue, which is mainly of benefit for few LGs responsible for urban water supply, the per capita revenue is around the same level of the Ugandan LGs. Compared to the size of the expenditures and the assignment of revenues, LGs in Kenya are much more self- reliant. The revenues in all three countries are heavily controlled by the CG, reflected in the legal framework in all countries, and in the use of approval procedures in areas like by-laws (particularly in Kenya and Tanzania) and borrowing (all three countries). The countries are in various stages of reforming the LG revenue assignments. Kenya has moved fast with the introduction of the single business permit and streamlining of the LG taxes, but a reform of the property tax (which is presently a land tax on non- improved property) is still to come. Tanzania has recently abolished a number of taxes (development levy, part of the produce cess and other taxes) in order to streamline and reduce the administrative burden, a move which will release LGs from administratively heavy and cumbersome taxes, but which will also increase the dependency of LGs on the transfers from CG. It has also created some immediate funding challenges, especially for the LLGs, which were entirely dependant on the revenues from the abolished taxes27. Uganda is in a process of reforming the property tax system, and considers the future of the other taxes, especially the Graduated Tax. It should be noted that the LGs in all three countries are getting increasingly dependent on transfers from the CG, a trend, which is expected to continue with the recent abolishment of a number of tax assignments in Tanzania and the discussions on abolishment of the G-Tax in Uganda. Revenue sharing across levels of LGs In Uganda and Tanzania, which have two layers of LGs, a revenue sharing system is in place whereby locally collected revenues are shared between the HLGs and the LLGs. In Uganda, 65% of the locally collected revenues are retained by the LLGs (part of this is shared with the villages, counties and parishes), whereas the minimum required share for LLGs in Tanzania is 20% (2003). The revenues sharing systems have been put in place to ensure local participation in planning, priority setting and decision-making close the citizens and to finance small-scale investments in services and infrastructure. 27The abolishment has also created severe problems for the LLGs, which relied heavily on the revenues from these taxes. The question of compensation for this abolishment is at the top of the agenda, (cf. the Country Study for Tanzania, Chapter 4). 38 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda In practice, the revenue sharing varies significantly across levels of LGs in the two countries, with some HLGs controlling a higher or lower share of the locally collected funds ­ and it has not been unproblematic. In Tanzania, the abolishment of a number of local taxes from July 2003 (development levy, produce cess, etc.), impacted seriously on the revenue sharing, as most of the shared taxes were abolished, leaving a very limited revenue basis for the LLGs. The compensation for the first year FY 2003/04 was allocated to the HLGs, but for the coming FYs the grants in lieu of the abolished taxes will be shared between HLGs and LLGs (50/50).28 Revenue enhancement initiatives All three countries have recently put a strong emphasis on capacity building of the LG ability to collect taxes and administer the revenues. Below is listed some of the main initiatives launched to improve on the LG revenue efforts: Table 4.6: Overview on the Initiatives within LG Own Revenue Mobilisation Issue Kenya (1) Tanzania (2) Uganda (3) Legal reforms Adoption of the Single Abolishment of taxes The graduated tax bands Business Permit ­ has (development levy, part have been changed streamlined and improved of the cess revenues, etc.) A property Tax Bill is taxation within this area to reduce the pending. administrative costs and the distortions on local The legal framework will economic development. be reviewed during the However, the reforms coming years under the have reduced the LGDP-II potential LG own source funding and increased dependency on transfers. Capacity Guidance to LGs on Training materials on Collection of best building development of revenue revenue administration practices and enhancement plans and elaborated, and training dissemination through strong focus on this area in the pipeline workshops and under the KLGRP/LATF conferences. scheme New budget guidelines put strong emphasis on this issue. Training materials elaborated on ways and means to improve LG revenues Incentives to No criteria in the grant No criteria on tax efforts Problematic, as transfers revenue raising system. However, in the in the grant system. have increased performance measures significantly. under the LATF, there are Individual incentives a number of in-built have been provided by New measures have been incentives to focus more LGs to tax collectors in introduced under the on LG revenue raising many places. LGDP-II grant system to 28According to the Draft Budget Guidelines, FY 2004/05. 39 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Issue Kenya (1) Tanzania (2) Uganda (3) (revenue enhancement improve on the LG plans, debt recovery, etc.). A share of the LG incentives to raise own Budget reform to improve revenue collected is source revenues. on the realism in LG retained at the LLG budgeting of LG rev. disposal Retention of a share of the LG revenue collected The coming LGSP at the LLG disposal capital development (65%) grant scheme has in-built incentives. All three countries have either introduced revenue mobilisation measures in some of the grant schemes or are planning to do so (Tanzania). However, in Uganda, these measures only concern one of the grant schemes- the LGDP development grant (also the case in the coming LGSP grant scheme in Tanzania); and in Kenya, these measures are applied more "indirectly" through LG incentives to development strategies for revenue enhancement and debt recovery. Simultaneously with the initiatives to improve on the LG tax administration and revenue mobilisation in all three countries, there are initiatives in Tanzania and Uganda, which will increase the dependency on transfers from CG. The recent abolishment of a number of LG taxes in Tanzania (from July 2003), and the discussions about abolishing some of future LG tax assignments in Uganda and Tanzania, will if not counterbalanced by other tax reforms, create an increasing LG dependency on transfers from the centre. 4.1.5 The System of Inter-governmental Fiscal Transfers As the transfers from CG are some of the most significant revenue sources for many LGs within the three countries, these have been subject to a detailed analysis in the country studies. The table 4.7 below summarises some of the main findings: Table 4.7: Comparison of the Systems of Inter-governmental Fiscal Transfers Issue Kenya (1) Tanzania (2) Uganda (3) Per capita size of the 1.9 USD 12.2 USD 16.8 USD transfers (B FY 2003/04) (B FY 2003/04) (B FY 2003/04) Relative size of the 31% 82% 88% transfers of the total LG revenues* (B FY 2003/04) (B FY 2003/04) (B FY 2003/04) Composition of the total transfers 1) Recurrent 0% 71% 63.0% (specific) condition- nal grants ** 2) Sector conditional 19% 28% 16.8% development grants 3) Discretionary 81% 1% 20.7% (block) grants 4) Equalisation 0% 0% 0.5% grants 40 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Issue Kenya (1) Tanzania (2) Uganda (3) Formulas for LATF is formula Will be reformed from the The present formulas distribution based, transparent and present non-transparent will be reformed in FY simple system, which is based on 2004/05 or 2005/06 to existing LG infrastructure to reflect the objectives formula based, transparent, of each transfer objective criteria (as per FY scheme 2004/05) Matching funds No No. Generally, no. The coming LGSP will The LGDP-II require co-funding like development grant some of the present district scheme has a 10% co- based support programmes funding element Link to LG tax effort No. No. Generally, No. However, there are The coming LGSP has in- The LGDP devt. grant incentives in the built incentives to improve encompasses LATF to focus on on LG revenue mobilisation conditions on the ways and means to development in LG improve LG own own source revenues, source revenues, and good performance requirements to within this area is elaborate plans, etc. rewarded in grant scheme Predictability High on the LATF High. High. (largest scheme), low In most cases, the LGs In most cases, the LGs on the smaller receive their transfer on receive their transfers transfers from RMLF time or within few days of on time, or with few (improved over the delay (improved over the days of delay most recent years) most recent years) (improved over the most recent years) Grant allocation to Not applicable Not yet, but planned for as Only for few the LLGs indicative planning figures development grants under the LGSP and for the (LGDP, PMA-non- CG grants in lieu of the sectoral grants), all compensation for the other grants go to the abolishment of taxes HLGs B=Budget, FY= Finance Year. All figures should be treated with due caution, as they are based on estimates and certain assumptions. * Estimated LG own source revenues, as the final accounts are not yet available, using figures for previously collected revenues. ** Conditional, in the sense of the requirement to use the funds within specific sectors, or areas within a sector, e.g. on health/drugs. (1) Kenya Country Study. Transfers are defined as the LATF and the road maintenance levy fund (classified as sector development grants), LG own source revenue assumed to be at the 2002/03 budget level, cf. country outline Annex 4.3. (2) Tanzania, Country Study, Annex 4.2. LG revenues in FY 2003/04 is estimated as being equal to the budgeted development prior to the reform, i.e. the real share of LG own source will be reduced significantly. The health basket fund and PEDP development funds are classified under sector devt. grants, based on estimates received from the Ministry of Finance. (3) Uganda Country Study. LG own source revenues in FY 2003/04 is assumed to be equal with the revenues actually collected in 2001/02, cf. Annexes 4.1 ­ 4.4 in the country study. The analysis shows that the design of the transfer schemes varies greatly across the three countries. The systems in Uganda and Tanzania have a number of similar features, as the conditional recurrent grants finance most of the decentralised service provision within education, health, water and roads. Although the number and specific modalities of grants differ, the basic design, and the wish to ensure that funds are 41 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda spent within a given area/activity are common. Both countries are in a process of improving the allocation formulas towards a more transparent, equitable, fair, objective and predictable system. Uganda has moved far in the direction of making a rather advanced but also complicated system of transfers encompassing 37 different grants: unconditional and conditional grants, equalisation and development grants. The lessons from this system, and the challenges experienced, could be instrumental for the other countries, especially the ongoing initiatives to address the previous design problems. The reform under the FDS encompasses a significant reduction in the number of grants, an increase in the LG autonomy, a closer linkage of the performance and capacity of the LGs with the level of grants, and flexibility in the utilisation, harmonisation and simplification of the transfer modalities (reporting, accounting, banking, etc.), and improvements in the linkage between the grants and the LG own source revenue mobilisation and accountability.29 Tanzania is also in the middle of a comprehensive reform encompassing: i) improvements in the allocation criteria for the conditional recurrent grant; ii) further autonomy in the utilisation of grants; and iii) a gradual introduction of genuine discretionary development grants, based on transparent objective criteria. Most grants are transferred to the HLGs. Only Uganda30 has started transferring funds directly to the budgets of LLGs (sub-counties) for development investments and to the administrative units (wards/parishes) in form of indicative planning figures. All the conditional grants are still transferred to the HLGs for disbursements. Under the coming LGSP in Tanzania, funds will be allocated to the LLGs as indicative planning figures to ensure stronger local involvement in planning and priority setting, whereby a local area is ensured a fixed planning envelope for realistic constraint based planning. However, the funds will be channelled to the HLG, and financial management will be kept at that level where there is larger capacity to handle the funds. The experiences from Uganda with the use of "indicative planning figures", whereby lower units have the right to chose how resources assigned to them are used, have been promising. 4.1.6 Performance Incentives All three countries have experimented with an elaboration of LG budget allocation performance incentives, especially addressing the generic elements of the administrative performance (cf. also section 4.1.4). Uganda has the most comprehensive experiences from the DDP and LGDP programmes, whereby the size of the development grants have been measured and adjusted against the LG performance within a number of generic areas like financial management, transparency, good governance, etc. The systems have provided remarkable and encouraging results. The results from these programmes have been followed up under the FDS implementation framework, where the flexibility in the utilisation of the recurrent conditional grants (autonomy) will be directly linked to the performance of the LGs within specific areas of planning, budgeting and accounting. 29Please refer to the Country Study for Uganda, Chapter 4, for a detailed review of these reforms. 30In Tanzania, there are a number of programmes, which transfer funds further down to the local level, but these are typically supported by technical units and arrangements outside of the LG system (e.g. TASAF). 42 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Kenya has introduced performance measures under the LATF, where the transfers are subject to the fulfilment of certain minimum conditions and adjusted against a number of performance measures. Tanzania has tested performance measures in specific districts under the UNCDF- supported programme in Mwanza Region, and it has recently designed a comprehensive Local Government Support Programme, (LGSP), which in Component One has in-built performance incentive measures concerning areas of generic administration performance and good governance. It is also considering a linkage of the flexibility in the utilisation of recurrent grants (level of autonomy) to the capacity and performance of LGs. Table 4.8: Comparison of Performance Incentive Systems Issue Kenya (1) Tanzania (2) Uganda (3) System LATF. Coming LGSP (will LGDP-II Discretionary block initially cover 40-50 LGs). Development grants: grants for recurrent Capital grants: minimum minimum conditions + costs, development conditions + adjustments adjustments against and debt recovery. against the LG performance the LG performance Minimum conditions within generic areas within generic areas + adjustments against the performance FDS ­ recurrent grant within generic areas system Typical Financial statements, Planning, Planning, performance areas Debt recovery Accounting, Accounting, performance, Fiscal capacity and co- Procurement, Involvement of funding performance, Co-funding, citizens in planning, Procurement, Revenue mobilisation Revenue Project implementation performance, enhancement capacity. Good governance planning Revenue mobilisation (transparency, performance. information, Good governance involvement of (transparency, information, citizens etc.) involvement of citizens etc.) Adjustments of 60% if MCs are Basic allocation if the MCs Basic allocation if the grants fulfilled and 40% are fulfilled and adjustment MCs are fulfilled and extra if performance of grants +/- 20% depending adjustment of grants measures are on the performance by +/- 20% depending complied with on the performance FDS: further autonomy if performance is high Results Positive, To be reviewed Positive, documented documented improvement in LG improvements in financial management planning, financial (planning, accounting management and good and reporting) governance (transparency, interaction with citizens, etc.) 43 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Issue Kenya (1) Tanzania (2) Uganda (3) Future Considerations on Up-scaling of the Further development expanding the programme nation-wide of the performance performance measures for the measures Considerations on recurrent grant utilisation of the principles schemes for other grants as well Considerations on use of sector-specific performance measures Although there are country specific nuances, the LGDP-II (Uganda) and the LGSP (Tanzania) have many similarities in the basic design. MC= Minimum Conditions. PM = Performance Measures. 4.1.7 LG Borrowing and Debt LG borrowing is an area for which, generally, there is limited information available. However, evidence from all three countries shows that many LGs are facing severe problems with the size of the debt, especially the outstanding overdrafts and arrears. The table below provides a general overview of the situation in each country: Table 4.9: Local Government Borrowing and Debt Issue Kenya Tanzania Uganda (1) Autonomy to Loans need the approval Need approval from the Need approval by the borrow of the Ministry of LG. Minister of Finance and Minister of Finance, PO-RALG (previously the Minister of LG) Size of the debt The aggregate size of the The aggregate size of There is not an aggregate total LG debt is 2-3 times the total LG debt is figure for LG debt, but the size of one year´s around 65% of one information from sample total LG own source year´s total LG own LGs shows that arrears and revenues source revenues outstanding payments are significant (20-40% of one year's own source revenues) Which LGs Strongly concentrated in Strongly concentrated Arrears and outstanding the few larger cities, in certain (5-10%) of payments are a general especially Nairobi and the LGs problem within most LGs, Mombasa. Long term urban as well as rural, debt from NCC however with great constitutes >50% of all variations. debt Major problems Significant debt build up Lack of a strong system Lack of strong commitment to statutory bodies and of commitment control control and cash flow housing schemes, pension and cash flow budgeting systems. funds, etc. budgeting. Downward trends in own Severe structural Lack of payments to source revenues. problems within the statutory bodies. Structural problems with a larger cities. Lack of incentives to excessive number of staff Weak financial improve on the debt in some of the LGs, and the management capacity and recovery and weak design of the transfer incentives to improve financial management system. Development Strong focus on debt Possibilities to Strong focus on ways and and future recovery and restructure/improve the means to improve on LG government improvements under the Local Government revenues through improved strategies to LATF scheme, including Loans Board are incentives (LGDP), 44 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Issue Kenya Tanzania Uganda (1) address the debt recovery strategies, explored improved technical capacity problems agreements between LGs (sharing of best practices). and the creditors, increase TA support to weaker in funds available for LGs within the field of Plans to restructure the debt recovery and financial management organisation of LGs and to performance incentives to ensure CG funding of the improve on the debt basic minimum LG staffing resolution structures (1) According to the new Public Finance and Accountability Act approval by the Ministry of Finance is required. According to the LG Act, borrowing needs approval by the Minster of Local Government if the amount exceeds 10% of the amount the LC council is eligible to borrow (which is 25% of the locally generated revenues, provided that the LG demonstrates its ability to meet its statutory requirements). Kenya is the country with the most severe problems of LG indebtedness. It has also been in the forefront of designing innovative mitigating strategies and measures to address the problems. However, in all three countries, some of the LGs have debt/arrears on a scale that necessitates specific structural measures to bring in a better balance between the revenues and expenditure, reviewing the entire structures and procedures of the LGs, combined with the introduction of stronger incentives to improve on the future financial management performance. 4.1.8 Financial Management Issues The capacity and performance of LG within the field of financial management have been matters of serious concern in all the three countries, and a number of initiatives have been launched to address fiduciary risks, mismanagement and inefficiency in handling of the revenues and expenditures. Poor status in financial management has been an argument for, especially, the line ministries to reduce the speed of the decentralisation process ­ and it is perceived by most stakeholders to be a real constraining factor. Below is listed a few of the critical dimensions on financial management performance, using "regional standards".31 Despite the worrying situation in many LGs, there has been a marked improvement within a number of financial management areas over the most recent years. This is especially the case within basic accounting, where the training and capacity building efforts have started to have an impact. Similarly, the incentives to improve on financial management performance in the development grant systems (Uganda/Kenya) have played a significant role in the improvement of, particularly, reporting standards and financial recording. Common weaker areas are: i) overoptimistic revenue budgeting; ii) lack of linkages between planning and budgeting; iii) weak internal audit and control functions due to confused roles and lack of facilitation of the control and internal audit functions; iv) lack of transparency in procurement procedures; v) very weak revenue collection procedures and practices, including inefficiencies and cases of corruption; vi) lack of sanctions whenever financial management problems are discovered in audit reports, inspection visits, etc. 31Based on CFAAs in various countries and experiences from the Team members. 45 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Table 4.10: Financial Management Issues Issue Kenya Tanzania Uganda Budgeting Weak. Medium. Medium. Unrealistic budgets (30- Still great variances Overoptimistic budgeting 40%). between budgets and of revenues (+/- 30-40 %) De-link between budget and outturns (+/- 20 %) plans Procurement Weak/medium. Weak. Medium. The basic tools (tender Low awareness about TBs are generally boards, etc) are established; rules and procedures. functioning and however the composition of New composition of TBs improvements have been the boards in some places is just initiated achieved within not adhering to the transparency and requirements information Up-dated Weak. Weak/Medium. Fairly good. accounts Many LGs have not However, the quality of Most LGs are up-to date produced accounts in many the accounts is often Great challenges within the of the years between 1979- weak LLGs 1999. Recent improvements have been noticed Use of IT Weak. Rather advanced as the Weak. Piloting of a IFMS has just IMFS was introduced in a Piloting of IMFS just started in a few (8) LGs large sample (38) some started in few LGs years ago, and as lessons have been learnt from this testing Revenue Weak. Weak. Weak. collection Budgets are not realised Budgets not realised Budgets not realised performance >30% underperformance >30% underperformance >30% underperformance Internal Weak. Weak. Medium. control and Internal audit units have only Internal audit units have Internal audit units are audit recently been established and been introduced recently, operational in most LGs, are still grumbling with their but are still very weak. although still rather weak roles. Absence of local public LG public accounts Absence of local public accounts committees committees are operational, accounts committees although still rather weak Audit status ­ Weak. Medium. Medium/advanced. Audit is done There are very few audit Delays approx. 6 months Delays in auditing of the reports with an official HLGs by av. 6 months. opinion on the status/quality Auditing of the LLGs is still carried out on a project basis, and a strategy needs to be developed Auditors Weak Medium Medium* opinion No opinion as all accounts 17 % clean 5% clean needs opening balance and 58 % qualified Approx. 95 % qualified completion of accounts from 25 % adverse Few adverse previous years. (2002) (2002/03) Follow-up on Weak. Medium/advanced. Fairly advanced. audit Nearly absent, as there are no Audit reports are Although there is still 46 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Issue Kenya Tanzania Uganda "real" audit reports followed-up by a great room for Committee in the improvement Parliament, but sanctions MoLG/Parliament are seldom applied. Committee and IGG Financial Weak. Medium. Relatively high reporting It has recently improved with The quality is constrained performance ­ in the the LATF requirements to by the use of different reporting to CG, but still produce financial statements accounting systems and weaker concerning procedures downwards to citizens Accountability Weak. Weak. Medium. (downwards) Has improved with the Limited publication of Improvements in publishing LATF requirements to information and of transfer data, budgets publicly transfer data and to involvement of citizens in and accounts, and involve citizens in the the control involvement of citizens in planning process planning and budgeting Please refer to the country studies for further details. * There is no consolidation of the audit reports. The information is based on interviews with the Audit and best judgement. 4.1.9 Institutional Framework ­ LG Finance The institutional framework on LG finance varies greatly across the three countries. Uganda has by far the most elaborated system of CG-LG budget clearance and negotiations, spearheaded by a Local Government Finance Commission and a number of newly established bodies to oversee and coordinate the various reforms on LG finance. These bodies all have representatives from CG and LGs. Tanzania has as an interim measure established a system of coordination committees (block grant implementation teams) in connection with the recent and ongoing reform of the recurrent grant system. A more permanent solution is being discussed to ensure proper data, information, analysis, inter-governmental coordination within the area of LG finance and to strengthen the capacity within the MoF and P0-RALG. In Kenya an advisory committee is established under the LATF transfer scheme, but there is no committee or body to ensure research and coordination of the overall issues within the field of LG finance. This may be established as part of the constitutional review process. In all three countries, the ministries of finance and the ministries in charge of local governments share functions on coordination and monitoring of LG finance. Table 4.11: Institutional Coordination within LG Finance Issue Kenya Tanzania Uganda Ministries in Ministry of Local PO-RALG Ministry of Local charge of LG Government Ministry of Finance Government finance Ministry of Finance Ministry of Finance, Planning and Economic Development Coordinative No No Local Government Finance bodies with CG Is discussed as part of the It is discussed to establish Commission and LG constitutional review. a more permanent forum (studies, advices, fiscal representatives for deliberations on LG reforms, monitoring) (overall) finance (analysis, studies, negotiations etc.) 47 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Issue Kenya Tanzania Uganda Specific bodies Yes In connection with the Yes, all with for certain The LATF scheme has an reforms of the recurrent representatives from CG functions advisory body with grant system a number and LGs representatives from the grant implementation LBCC (budgets, grants, major stakeholders (public teams were established etc.) and private sector) with representatives from LGROC (accounts, IT etc) CG (P0-RALG, MoF, LRECC (revenue Line ministries and LGs) enhancement) 4.2 COMMON LESSONS AND CHALLENGES The comprehensive case studies of the fiscal decentralisation experiences in the three countries have generated a number of common (and country-specific) lessons learned on the design and implementation of systems of LG finance. Some of the main ones are briefly highlighted below: The importance of a fiscal decentralisation strategy · The experience from Uganda, but also from Tanzania and Kenya, shows the importance of having an elaborated and well developed fiscal decentralisation strategy in place, and the problems with the absence of such a strategy. In Uganda, the incremental development of the conditional grant system, lack of coordination across the sectors, without a strategy linked to the overall decentralisation objectives, led to a system with high LG transaction costs, less autonomy and downwards accountability. With the recently adopted FDS there is a window of opportunity to improve on these issues, although it is too early to judge the results. In Tanzania, the systemic fiscal reforms were delayed, but have now taken off. And there are plans to elaborate an overall fiscal strategy to link the various reforms. In Kenya, the elaboration of a strategy will be an urgent activity when the final decisions are taken on the overall structure of the system of LGs, based on the outcome of the constitutional review. Budget approval procedures · The three countries have various systems in place to ensure that LGs are adhering to the overall national service delivery objectives and standards. The experiences from the three case studies have shown that a rigid ex ante control of individual LG budgets leads to delays in the budget preparation and inefficiency in the budget execution process. The capacity of the CG to deal with all these budgets individually has also been over stretched, leading to insufficient support and bottlenecks ­ and budgets where the legal status is unclear. Well-developed budget guidelines, combined with capacity building support and ex-post audit and follow-up, has proved to be more feasible, and has generally improved the quality of the plans and budgets. Other common challenges within planning and budgeting are problems with the linkages between the plans and the budgets in a resource constrained manner, and with numerous CG planning and budget guidelines, support from donors to elaborate various planning approaches, which have often not been sufficiently consolidated and harmonised. 48 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Accountability and transparency in financial management · Although the involvement of citizens in the planning and budgeting has improved in all three countries, it is still a great challenge to ensure active involvement of citizens (not just attendance in meetings) concerning the local priorities. Early announcement of the resource envelope (transfers) from the CG capacitated LG planning units and CG promoted fiscal incentives to "push" LG further in the direction of transparency and downwards accountability have all been enabling factors. The mechanisms are still in the early stages of implementation, but means like participatory planning tools, open budget conferences, involvement of citizens in project implementation and monitoring, and demands for LG transparency, communication means like publications of financial information, notice boards with budget and accounts figures, LG radio announcements, public hearings etc. have shown some encouraging signs. Expenditure priorities · The lessons learned from the development programmes, where LGs have been given the autonomy to make cross-sectoral priorities concerning the development investments (discretionary grants), have been very encouraging, especially when diligent performance incentives (measures) have been elaborated and sanctioned/followed-up in a transparent, objective and firm manner. A short "negative" list may be required, like the one elaborated under the LGDP-I and II in Uganda (e.g. vehicles and machines cannot be funded from the local development grants) and the coming LGSP in Tanzania. The utilisation of funds in Kenya, where these items are not excluded, should be followed closely. However, in Kenya there are other incentive-means (rules on minimum amount for capital investments, maximum for personnel, and minimum for debt recovery), which can influence the local priorities without relying on CG micro-control. These indirect measures, which leave sufficient autonomy to adjust to the local service needs, have shown interesting results in the three countries. They also add nuances and new perspective to the current debate on the advantages/disadvantages of conditional versus unconditional grants. It is remarkable to notice that under the LGDP-I (2000-2003) more than 90% of the discretionary development funds have been utilised within the five national priority sector areas (health, education, water/sanitation, agriculture and roads). Revenue assignments ­ the overall composition/profile of revenues · In strongly decentralised countries in terms of the assignment of expenditure, such as Uganda and Tanzania, a significant amount of grants from the CG would typically be required to fill up the gap between the expenditure and feasible revenue assignments32, to balance/equalise the development across the country and to ensure fulfilment of certain national development targets. However, the studies ­ especially the experiences from Uganda ­ have shown 32It is well recognized that some vertical imbalances in most cases will be needed, but that excessive vertical imbalances between assignment of expenditure and revenues with limited LG autonomy at "the margin" in order to adjust local revenues to local needs, may lead to a decrease in accountability and inefficiency in the utilisation of public funds, (cf. e.g. Steffensen /Trollegaard, 2000 op. cit and the World Bank, 2004, op. cit. 49 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda that there are drawbacks in a situation with too heavy dependence on the grants from the CG, especially in terms of: i) reduced downwards (increased focus on the up-wards) accountability, ii) problems with strong earmarking of the grants (as "the one who pay for the music wants to decide the songs") and iii) emerging problems with the LG incentives to generate own source revenues. All this undermines the entire sustainability of the system of LG finance, ownership of the investments and the relationship between the local politicians and their constituencies. Evidence from the field and various studies show that strong disconnection between the expenditure and taxation may encourage "free-riding" in tax payments and LG tax collection efforts, and a lack of pressure on efficient utilisation of the revenues ­ and reduced accountability, as the costs are not directly obvious for the LGs and their citizens.33 This is especially a problem when it is coupled with problematic revenue assignments, non-conducive legal frameworks, and weak administrative capacity to ensure utilisation of the tax potential. · Hence, the design of a system of high yielding LG revenue assignments and support actually to utilise the LG tax potential in a fair, efficient and equitable way is a tremendous challenge in each of the three countries. However, the transfer of new revenue assignments to LGs without improving the existing sources poses a risk for undermining the overall national tax yield. Hence, various measures have to be initiated and implemented simultaneously and in a balanced manner. Types of tax and the abolishment of taxes · The present LG tax assignments in the three countries have been criticised for being of low yield, non-buoyant, non-predictable, inequitable, unpopular in nature, and expensive to collect compared to the CG tax assignments. However, experience has shown the importance of having well prepared LG tax reforms linked to the overall national tax reforms. There are no easy ­ "over the night" ­ solutions to these challenging problems. In the meantime, much can be achieved by improving the tax administration and collection procedures, not only in purely administrative efficiency terms (registers, valuation, collection procedures), but also in terms of the links and interactions with the citizens, transparency and dialogue, establishment of better links between tax payments and service delivery. Sharing of best practices and capacity building should be important ingredients in any reforms within this area. · Abolishment of LG tax assignments without identifying new sources of sustainable LG revenues, creates problems for the entire viability of the LG fiscal system. Revenue sharing between HLG and LLGs · The experiences from the revenue sharing across levels of LGs in Tanzania and Uganda have been mixed. On the one hand, the share of the collected revenues retained at the LLG level for small scale services and local council 33These findings resemble recent findings of i) a review in Pakistan, Asian Development Bank, DFID, World Bank, 2004 op. cit, p.6; ii) from Rémy Prud´Homme in Fiscal Decentralisation in Africa: A Framework for Considering Reform, Public Administration Development and Development, 23, 2003 and iii) Steffensen and Trollegaard, May 2000, op cit. 50 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda expenses have ensured certain funds for political activities, in some areas motivated LLGs to raise funds (especially in Tanzania where certain shares have been given to the tax collectors) and contributed to a degree of local ownership. On the other hand, the complex revenue sharing arrangements have led to a fragmentation of limited resource, reduced the possibilities for investments in services, and increased the costs of political meetings/ interactions. The available amount of funds for service provision for a typical village in Tanzania has been less than 500 USD, leaving very few options for investments in service and infrastructure. The sub-counties in Uganda are typically much larger, and they have retained a significantly larger share of the resources; hence, have typically significantly more resources. However, the villages/parishes (administrative levels below sub-counties) in Uganda have had very few sources available for real service provision. The lessons have also shown that this is an area that requires firm CG oversight and supervision, if the rules are to be complied with, as many LGs have deviated from the legal framework.34 Please refer to the sections below for the sharing of the grants from CG. Incentives to improve LG revenue mobilisation · The establishment of the right incentives to collect LG taxes has been a great challenge, especially in the cases of Uganda and Tanzania, with large increases in transfers from the CG35. Transfer systems, with tax effort criteria and performance measures for tax collection, have provided initial results in Uganda. Also, the focus on the revenue enhancement strategy and debt recovery in Kenya seems to generate some positive lessons. However, these incentives should be combined with cautious measures, not just to mobilise revenue for the sake of revenue enhancement, but also to do it in a poverty- sensitive, development-oriented fashion, focusing preferably on the major tax sources, combined with improvements in the efficient utilisation of the collected funds. This requires an enabling legal framework, as well as capacity building of politicians and staff (especially tax collectors). Privatisation of tax collection may provide a higher yield in certain areas, but the mixed experiences from Tanzania and Uganda have shown that this require a very cautious approach, and strong control and supervision by the contract giver (the LG). Tax administration · LG tax administration has been one of the most problematic areas of LG finance in all three countries. Some of the findings of this study, and supported by findings from other studies36, are that the issue is complex and includes problems such as: i) complex and restricted legislative framework on LG 34Cf. e.g. Uganda: LGFC: Revenue Sharing Study Phase II, Final Report 2002, and Tanzania: "Local & Lower Local Government Revenue Sharing and Management of Finances at the Village, Ward and Mitaa Levels", UNCDF/DFID, Final Main Report, 2003. 35Please refer to Rémy Prud´Homme: "Fiscal Decentralisation in Africa : A Framework for Considering Reform in Public Administration and Development", The International Journal of Management Research and Practise, Volume 23, No. 1 , 2003 p.24-26, UNCDF, for a more theoretical explanation of the issue. 36E.g. studies by Odd-Helge Fjeldstad (CMI, NBIR, REPOA) in: "Formative Process Research of the Local Government Reform in Tanzania", analysis of data from six councils 2000-2003, Final Draft, 19 April 2004. 51 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda taxes, with limited autonomy to adjust the rates; ii) multiple and complex tax bands/levels, iii) poor administrative capacity to access the revenue base, collect the taxes and enforce the taxes, iv) intentional resistance from the tax payers to pay taxes, v) corruption and embezzlement of revenues; vi) political pressures on the LG tax administration to relax on revenue collection. There is a strong need for combined efforts to simplify and improve the tax legislation, to support the LG capacity to administer and collect the taxes, and to review and strengthen the incentives to improve the revenue mobilisation regarding the major taxes. Property taxes · It is interesting to note the high untapped revenue potential from LG property taxation reported in many studies on LG finance in the three countries - evidence which has been confirmed during the short field missions. This is an area in need of future support in a multi-faceted manner, covering the legal framework, which is highly restrictive at the moment, support to smoothening of the valuation procedures, registration, collection methods, etc. Transfer systems · The study has identified a number of achievements within the functioning of the inter-governmental fiscal transfer system. The transfers from CG are being increasingly reliable and predictable, providing LGs with a firmer basis for planning and budgeting (well-known resource envelope) and better possibilities to achieve the service delivery objectives. In most cases, the LGs are receiving only slightly less from transfers than what was budgeted, the balance typically attributed to lack of LG reporting for previous instalments and/or delays in the reporting/transfers creating problems of absorption in the final weeks of the FY. Although this is the general picture, there is still room for improvement within this area, e.g. on the CILOR and transfers from the RMLF in Kenya. · A mixture of conditional and unconditional grants will typically be necessary to provide a balance between autonomy and guarantees for LG adherence to key national service delivery targets and priorities. However, experience has shown that too many conditionalities and limited flexibility may cause problems with efficiency, accountability and may distort the incentives for politicians and staff to perform efficiently. There is room for further development of the promising experiences from the so-called "second generation" of transfer schemes, which adjust the size and the level of autonomy against the LG capacity to absorb, capacity to reduce the fiduciary risks and improve performance within administration and good governance, [cf. below]. · Hitherto, these performance-based transfer schemes have only been utilised within the development transfers schemes (grants which are easier to adjust without implications for the current LG business). However, there is room for a cautious "roll-out" of the principles to the recurrent grant schemes, e.g. by linking the level of flexibility in the utilisation of grants against the LG capacity and performance, allowing more room/autonomy for the best performing LGs, thereby "catalysing" good performance. · Equalisation of the disparities in fiscal capacity (revenue mobilisation potential) and expenditure needs across LGs has been carried out mostly 52 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda through the conditional grant system. Uganda has introduced a very insignificant equalisation grant, which constitutes less than 0.5% of the total grants. It is available for the weaker LGs in Uganda, but its impact is very limited. None of the countries has been particularly successful in addressing the regional inequalities, which have typically increased over the last five years. The factors/criteria in the transfer systems are typically made in isolation and lack a holistic view on the various transfer schemes: how they play together and how other transfer flows (on-budget and off budget) impact on the economic development, especially in the poorest areas. New allocation criteria have been prepared in Uganda and Tanzania for the sector conditional grants, but the need for an overall equalisation system (through the criteria in a general block grant system or through a specific equalisation grant) to ensure balancing of the various grant systems may be considered. However, the allocation of funds should also keep the absorption capacity in mind, especially administrative capacity (planning, financial management, monitoring and accountability), and LG fiscal capacity to cater for operational and maintenance costs. Significant increases in transfers to weaker areas, without sufficient linkages to capacity building and improvements of the absorption capacity, may lead to a waste of resources. · Related to the conditional grants, the study has noted a great proliferation of rules, guidelines and requirements for LGs, which impact negatively on the transaction costs and draw attention away from implementation and operational and maintenance functions, efforts on monitoring of projects, etc. There is an awareness of this under the reform programmes in all three countries, and initiatives to harmonise and streamline the budget guidelines, accounting and reporting systems etc. under the various government transfer schemes and the district support programmes, will be a great future challenge. There is also a need in all countries to ensure a better harmonised, streamlined (mainstreamed) system for linkages between the planning and budgeting guidelines. The two processes of planning and budgeting often run in parallel, rather de-linked from each other. · Conditionalities may create a shift away from local accountability towards focusing entirely on compliance with the central government reporting requirements and targets (upwards accountability). But properly designed, broader incentive-based requirements for LGs to improve on the local downwards accountability ­ such as publication of transfer figures, publishing of plans, budgets, accounts and audit reports, notice boards, involvement of citizens in planning and budgeting and monitoring, transparency and dialogue, establishment of user committees/project implementation units with citizens, etc ­ may also strengthen the LG incentives to open-up for dialogue with the citizens and may, over the time, enhance citizen participation and downwards accountability. The experiences from the communication strategy of the LGDP and the performance measures on good governance, linked to the size of the transfers in Uganda and the LASDAP/LATF process in Kenya, have yielded promising results for this form of indirect support to enhanced local accountability. · Early announcements of the grant envelope and indicative planning figures (improved in all three countries) provide a hard budget constraint and a more resource constrained and realistic planning and budgeting process. 53 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Development budget · Although this is an area that has improved recently, there are still common challenges in addressing the various development funds routed through vertical, central government and/or off-budget channels to local service providers. The decentralisation of the development budget has typically started later than other LG fiscal initiatives, and there is a lack of proper overview and information of the total allocation of resources for local service provision. Furthermore, the multiple funding flows create planning and budgeting problems, overlaps and a lack of cross-sectoral coordination and priorities. Recent reforms in Uganda (under the LGDP-II and GoU/joint donor coordination/FDS) and Tanzania (as part of the preparation of the LGSP and the draft Letter of Sector Policy for Devolution of the Development Budget) have sought to bring development funding on-budget, to channel the funds through the LG system on-budget, to ensure common planning, budgeting, accounting, reporting and auditing procedures and standards are promising and deserve future attention and support. · The challenge is two-fold ­ 1) to "fold" the district/area based programmes into the governments´ transfer schemes and 2) gradually to merge the sector development grants into the non-sectoral discretionary development grants to ensure better local priorities along with increased capacity to handle the funds. · The recent experience with transfers of development funds to the lower levels of LGs (Uganda) have been positive and will be applied in the future LGSP in Tanzania as well; however, in the form of indicative planning figures at the ward level. In the design of these sharing arrangements thorough consideration is important concerning balancing of the aims to involve the lower levels in planning and participation (bringing funds closer to the citizens), and to ensure sufficient available funds for meaningful budget envelopes/and options for real improvements in services. The intention will be to avoid a great fragmentation of limited resources and to ensure sufficient capacity to handle the funds in terms of planning, financial management and project implementation. Incentives are important but often overlooked · There has been an increasing awareness in the recent reforms that systems of LG finance have intended, as well as unintended, incentives. Some of the existing conditional grant systems in Tanzania and Uganda have allocation criteria that will "punish" improvements in the service delivery, e.g. if the there is an improved access to clean water in a particular LG, the grants may be reduced for the subsequent years. LGs are also forced to use funds within e.g. health, where the related area of water/sanitation may be more in need of attention in that particular LG. Another example is the perverse incentives to hire and bring excessive staff on the LG payrolls, as these have been financed from the grants (Uganda and Tanzania). · This calls for a flexible strategy, which promotes incentives for LGs to buy into the national priorities, allowing for the local priorities across (and within) the sectors, e.g. within certain ceilings. · The emerging application of incentive means in the transfer schemes in the three countries (Tanzania still only applied on a pilot scheme, but with plans for a significant roll-out) are among the most interesting and promising recent developments, and they have documented impacts in terms of improving 54 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda financial management procedures and good governance.37 However, the incentives have been mostly directed towards the performance of the LG as an entity, and the elaboration of incentives for individual staff performance has been less in focus, although equally important, (cf. Chapter 5). Financial management and financial position (indebtedness) · Weak financial management capacity and performance have definitely been a stumbling block for a quick decentralisation of functions and funding to LGs in the three countries. It has been easy for opponents of decentralisation to point to obvious weaknesses in the LG performance, although the performance has seldom been directly compared with the performance within the central government agencies. This study has identified serious weaknesses in most of the LG financial management "cycles" from planning to auditing. However, the marked improvements in certain areas should not be neglected. The strong support to capacity building rendered from the local reform programmes, the gradual improvement of the incentives in the transfer systems to strengthen financial management performance, have started generating results ­ especially in areas such as basic accounting and financial reporting. Some of the key lessons have been: i) Incentives integrated in design of the transfer systems have an impact. ii) IT systems may enhance financial management performance, but demand tremendous efforts in a weak financial management environment. There is a strong need for integrated capacity building support and up-grading of staff, settling of the backlog of previous accounts, and current support within financial management/IT application. It has also been noted that there have been few cost-benefit analyses of the introduction of IT in LGs in the three countries. iii) Financial management capacity is more a question of a combination of targeted training, restructuring and incentives than absence of people in positions (although there are certain vacancy problems in a few of the rural areas). iv) Establishment of a number of control authorities, like internal audit, LG public accounts committees, evaluation committees for TB procedures, inspection and auditing organs, may improve the general situation, but the impact is heavily dependent on the institutional inter-linkages (the way they work together and follow-up on each others work), the clarification of their roles and responsibilities and the manner in which these are facilitated and utilised by the political level in the LGs. "Follow-up" is critical for success. Many analyses and auditing reports are shelved immediately and have little impact. Coordination of the fiscal reforms and current challenges · The experiences from all three countries show the importance of properly coordinated fiscal reforms. In Uganda, the coordination has improved with the strengthening of the Local Government Finance Commission, adoption of the fiscal decentralisation strategy framework, and later establishment of more broad based coordinative bodies ­ i) LG Budget Coordination Committee, ii) the LG Releases and Operations Committee and iii) the LG Revenue Enhancement Committee, with involvement of all major CG and LG 37These findings are supported by other analyses of the impact of fiscal incentives, e.g. Roger Shotton (ed) in Local Government Initiative: Pro-Poor Infrastructure and Service Delivery in Rural Asia ­ A Synthesis of Case Studies, UNCDF, 2003, which has documented the same experiences in Bangladesh. 55 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda stakeholders. This has led to improved budget guidelines and more coherent messages from the CG to the LGs on various issues, reduced the overlaps in activities, and pooled the limited resources for reforms. The experiences have also shown the importance of clarifying the roles of each agent, and the links between these coordination bodies and the implementing agencies (MoF, MoLG, Commissions, etc.). In Tanzania, the current working groups, established for the design of the reform of the recurrent grant system, have also been instrumental in bringing all stakeholders on board with critical reforms of the grant allocation criteria and objectives, especially the line ministries, which will often tend to be hesitant about such changes. Consideration is ongoing about forming a more permanent forum for LG finance analysis, dialogue, and coordination. In Kenya, the LATF advisory committee has involved all the key players around a particular scheme, but a need has been identified to broaden this cooperation to other areas, e.g. in the form of a LG Fiscal Commission or a current negotiation coordination body/mechanism. 4.3 EMERGENG ISSUES AND NEED FOR FURTHER SUPPORT A few key areas38 warrant further consideration in the reform work regarding the system of LG finance: 1) There is a very strong and urgent need to improve the basis for all reforms of LG finance ­ namely the fiscal databases. Reliable figures on LG aggregate revenues, expenditures, service delivery, human resources, etc. are very limited and scattered all over many institutions and sources. There is a great need to consolidate, coordinate and mainstream this important area, which is the basis for many other reforms. 2) Development of LG fiscal strategies and realistic, time-bound and costed sequencing of the implementation activities is of critical importance. This should be linked with the overall decentralisation strategy and the wider public sector reform. 3) Despite the significant support from the governments and the donors, there has been, and still is, a strong need for improving: i) the linkages between the planning and budgeting - the process is typically de-linked, overlapping and/or not timely and properly synchronized ­ and ii) the involvement of citizens in these process in form of citizen participation and influence on the planning and budgeting process through means of improved transparency, communication, and open access to various decision-making fora. 4) The LG own source revenues: Although these may not be the major share of the total LG revenues in most LGs in the three countries, they are of critical importance for the local accountability, ownership, sustainability and viability of the entire system of decentralisation. There is a need for support 38Please note that the other areas highlighted are important as well, but the authors have felt that these few areas warrant special attention in the shaping of the future reforms. 56 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda to the design and implementation of improved systems of revenue assignments and mobilisation, including legal reforms, support to improved tax administration practices, focusing on improved local revenue mobilisation, considering the overall development objectives ­ i.e. designed in a poverty-sensitive way, without distorting the local economic development. This will be an immense future challenge, which deserves concerted effort from the governments and the main development partners. 5) There have been remarkable improvements in recent years in the coordination of support for LG development investments, establishment of joint donor basket funds and common project designs ensuring a more equitable distribution of funds and procedures mainstreamed with the government procedures. However, this process will still be challenging in the coming years, and will require better linkages with the sector support programmes, in order to ensure synergies between the LG based programmes and the SWAPs and a gradually merging of the district support programmes, with components of capital investment support to LGs, with the CG emerging development transfer schemes. 6) Development of incentive based grant allocation systems has provided interesting and encouraging lessons, which deserve further consideration in the next generation of these schemes, encompassing thoughts on more output-oriented indicators, links to the sector performance, etc, keeping the system manageable and simple. Financial incentives may be considered, combined with other incentives, such as wide public announcements of LG performance, league tables, rewards for best practices, publications with key figures, etc. 7) Support to financial management needs to be viewed in a holistic manner. The challenge will be to integrate the activities on the IMFS, with improvement of other systems ­ planning, budgeting, accounting, procurement, reporting systems, etc. and comprehensive capacity building programmes. IT solutions cannot stand alone, without reform of these basic pillars. 8) Innovative approaches to deal with financially distressed LGs: Some LGs are in a situation where structural and fundamental measures have to be taken with support from outside. 9) Development of systems and procedures for improved downwards accountability is one of the means for reduced fiduciary risks and efficient service provision. Development and implementation of innovative approaches to improved downward accountability need strong support. 10) Finally, it is an overarching challenge for CGs to move away from a role of tight ex ante micro control to one of guidance - ensuring an enabling environment for the LGs, supervision and ex post sanctioning, combined with further development of good incentives for current LG improvements in service delivery and good governance. 57 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda 5. HUMAN RESOURCES The literature on Local Government personnel systems is not as developed as the literature on fiscal aspects of decentralization. However, it is well recognized that an efficient and professional cadre of LG staff is essential for LG delivery of quality services. Furthermore, it is evident that some level of local control of executive staff is essential for the local accountability of staff and thus autonomous and responsive LGs ­ which, in turn, is assumed to lead to more effective services. However, local level control over staff has some risks, including: · Bias in selection of staff according to local preferences without due consideration of professional competencies (with, rather, an emphasis on e.g. ethnicity or relationship with a specific politicians) ­ recruitment of "sons and daughters of the soil"; · Stronger local political pressure on LG staff ­ including pressure on them to undertake illegal payments to councillors or other financial mismanagement; · Limitations on career prospect for the higher level cadres, as possibilities for promotion within a specific LG are limited; · Excessive local recruitment drives and/or non-payment of staff salaries or pension contributions, which leads to accumulated debts that ultimately may have to be carried by the central government. On the other hand, the above weaknesses can arguably also be found in central government managed systems, and the weaknesses in LG administration of personnel can likewise be argued to be results of e.g. the low level of financial resources availed to LGs, insufficient capacities for HR management, inappropriate regulations or lack of incentives for proper LG management of personnel ­ all probably to be better resolved by improved support to LGs rather than centralization of responsibilities. In East Africa, different models have been applied to resolve the challenges associated with LG personnel management. These have, over time, included all three basic models for personnel management that, according to Mawhood,39 generally have been practised in Africa: · Separate, in which every LG acts as a completely autonomous employer; · Integrated, in which the LG executive is composed of central civil servants; · Unified, in which the LG staff are employed locally but organized nationwide in a single civil service parallel to the central one. This chapter seeks to establish in a comparative manner across the three countries: · The state of current legal and policy guidance to management of LG personnel; · Quantitative trends in the development of local and central government personnel; · The general human resource capacities in LGs; · State of service conditions and HRD practices in LGs; · Experiences with restructuring of LGs; · Experiences with capacity building in LGs. 39Mawhood, 1983: Local Government in the Third World ­ the Experiences of Tropical Africa. 58 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda 5.1 LEGAL AND POLICY FRAMEWORK The systems of management of personnel in LGs differ substantially among the three East African countries. Uganda has a rather clear legal framework, whereby LGs hire and fire through each of their respective District Service Commissions. However, terms and conditions of service for employees are determined nationally and similar to those for central government employees. As LGs in Uganda largely depend on central government transfers for payment of salaries their staff establishments, except for a few financially stronger LGs, are in this manner also effectively ­ although indirectly ­ controlled. In Kenya, the LGs have been delegated the responsibility for hiring and firing lower cadre staff (salary scales 10-20), whereas senior staff are hired and fired by the Public Service Commission. Terms and conditions for LG staff differ from central government employees and are negotiated between the unions and the LGs. As LGs in Kenya rely on own revenue for salaries, they can in practice exercise more autonomy in decisions regarding the size of their lower cadre staff establishments. In Tanzania, the system for management of LG personnel is under transition and currently rather confused. A new Public Service Act was passed and amended in 2003, in a manner that essentially centralized LG staff management and abolished other existing LG legislation and regulations for staff management. However, in practice the Act has not been implemented in LGs. It is furthermore expected that the Act will be revised shortly in order to bring it in line with the LG Reform Policy. The subsequent discussion is, therefore, based on the legislation and regulations as currently applied in LGs40. This system was introduced from 2001 in the reforming LGs, but it has yet to be implemented nationally, just as the entire legal framework remains to be confirmed. It entails a partial devolution of responsibilities for hiring and firing to LGs ­ mainly junior technical staff and support staff, whereas so-called "category A staff" (the Directors and Heads of Department) still are largely centrally managed through a Local Government Service Commission41. The practice of LG personnel is further complicated by some line-ministries that continue to issue directives regarding transfer of "their staff", in spite of LG legislation. 40The amended LG Act and the 2000 Staff Regulations ­ however, these are initially only implemented in the so-called first batch reforming 38 councils; for a full discussion please see the Tanzania Country Outline. 41Abolished by the Public Service Act, but a department under the new Public Service Commission will undertake similar functions. 59 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda In both Kenya and Tanzania (and to a much lesser degree in Uganda), there are some substantial differences among procedures for management of different staff categories. These apply partly to different sectors, but mainly to whether the staff are senior (the executive staff: heads of departments and senior technical staffs) or junior (junior technical staff and support staff) ­ this is reflected in a simplified manner in the below table42. Table 5.1: Overview of the Personnel Management Functions Function Kenya Tanzania Uganda Personnel provision Recruitment, senior PSC LGSC under PO- LG (DSC) staff RALG, with some hearing of LGs Recruitment, junior LGs Council establishes a LG (DSC) staff Recruitment Board Appointment of senior PSC President for City LG (DSC) staff Council Director. Minister for distr. town or municipal Director Appointment of junior LGs LGs LG (DSC) staff Employment LGs LGs LG Transfers, senior staff PSC (MLG) Minister and LGSC Not applicable except on request Transfers junior staff N.A. N.A. from LG Promotions, senior PSC LG recommends to staff appointing authority LGs (DSC) Promotions, junior LGs Director recommends staff to council Staff development Performance The Clerk by the appraisals, senior staff Provincial LG LGO/PSC; LG otherwise HODs and CAO immediate supervisor and PSC/MLG Performance Clerk and LG appraisals, junior staff immediate supervisor Training Mainly through LGs ­ supported LGs externally funded and by capacity managed building grants programmes from MoLG 42For detailed discussion see chapter 5 of each of the country reports. 60 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Function Kenya Tanzania Uganda Other staff development LGs LGs LGs Salaries and other benefits Salaries and salary PSC and the scales ALGE and unions LGSC Ministry of Public Service. Incentive packages ALGE and unions LGs LGs Staff relations and discipline Disciplinary and PSC/MLG ­ Minister for appellate authority for appellate body not Directors, LG DSC with the senior staff clear advised by Public Service Committee for Commission as the Personnel for appellate body. HODs Disciplinary and Clerk LG; appeals LGs appellate authority for to PSC junior staff Termination of services Termination of senior PSC/MLG Council submits DSC staff resolution to Minister Termination of junior LGs LGs DSC staff Centralised ALGE Association of Local Governments Employers functions DSC District Service Commission Decentralised HODs Head of Departments functions LGSC Local Government Service Commission PSC Public Service Commission The system for management of LG personnel in Uganda is thus an example of a "unified system" discussed in the introduction, whereby LG staff is employed locally but organised nationwide in a single civil service parallel to the central one. Tanzania and to some extent Kenya have indicated that they intend to move in the same direction, but in both of these countries the systems applied for LG staff management can best be characterised as examples of "integrated systems" whereby the LG senior staff is composed of central civil servants. 61 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda 5.2 TRENDS IN LG AND CENTRAL GOVERNMENT EMPLOYMENT LGs share of total public employees is increasing in all three countries. However, reasons differ. In Uganda and Tanzania, the share is very substantial (73% and 63%) and foreseen to increase further. This is the result of the ongoing reforms, whereby staff for key social service delivery sectors (in particular primary education and health) have been transferred to LGs. These transfers happened earlier, but as the PRSPs emphasis continues to be on improvements in social service delivery (e.g. universal primary education), these social sectors continue to expand their staff substantially, whereas there is an attempt to keep administrative staff at the same level. In both Uganda and Tanzania, the phases of the public service reforms that focused on "right sizing" were completed almost a decade ago, and total public employment has, during the last years, increased. Table 5.2: LG Share of Public Employment in East Africa Total public 1998 1999 2000 2001 2002 employment Uganda 168,956 177,520 178,741 196,311 211,420 LG share 65% 67% 68% 71% 73% Tanzania 275,178 266,426 274,408 271,674 280,830 LG share 57% 58% 62% 62% 63% Kenya 661,300 650,300 644,500 610,900 612,100 LG share 12% 12% 13% 13% 14% Source: Central Bureau of Statistics (Kenya), Ministry of Public Service (Uganda), CSD Payroll data base (Tanzania) In Kenya, the situation differs substantially, as the public service is still "right sizing", and as the LGs have a much more marginal role in service delivery. Thus, it is no surprise that they only employ a much smaller proportion of public employees. Furthermore, as discussed later, in Kenya most of the staff in LGs are junior or support staff. The gradual increase in LG share of public employment in Kenya is foremost a sign of how the civil service reform focuses mainly on "right sizing" the central government and parastatals, whereas LGs so far have not undertaken any substantial restructuring and "rightsizing" ­ although it is much needed (see section 5.3 and 5.5). 5.3 HUMAN RESOURCE CAPACITY AT THE LG LEVEL The level of human resource capacities across the three countries is not easy to compare, as the type of staffing obviously depends on the type of LG (higher ­ lower; urban ­ rural). The task of comparing the three countries is complicated by the absence of national level statistics and difficulties in establishing comparable categories. The table below should thus be taken with some caution, although it gives a broad picture of the different levels of human resource capacity at "typical" (rural) LGs in the three countries. The table only gives an indication of the level of professional staffing at district headquarters and excludes, for instance, professional medical staff at district 62 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda hospitals. In Uganda and Tanzania it furthermore excludes any professional staff assigned to lower levels of LGs. From the table it appears that Tanzania and Uganda are roughly comparable when it comes to professional staff levels at a typical rural district HQ. The major difference seems to be in the works department, which, in the case of Uganda in recent years, has been substantially boosted by posting of staff for manning District Water Offices. Kenyan LGs have no professional health, education or agricultural sector staff (exceptions are the larger LGs that also have been assigned responsibilities for education and health). In Kenya, professional staff posts in administration, finance and works departments are furthermore most often characterised by vacancies or staff in acting positions. The major differences arise when staffing at lower level LGs are compared. The sub- counties in Uganda have substantial technical staff, while no such technical presence is available in Tanzania or Kenya. A sub-county in Uganda will today, in most cases, have a graduate sub-county chief, a sub-accountant and several professional staff working in the productive sectors. Table 5.3: District Staff ­ Professionals in "Typical" LG 43 Department Uganda Tanzania Kenya Administration/Management 5 4 5 Finance 6 4 5 Education 7 4 0 Health (HQ only) 6 5 0 Agriculture, Marketing and Natural 13 14 0 Resources Works (including water) 14 6 3 Community Services 4 4 0 Planning 3 3 0 Internal Audit 2 1 2 Total 60 45 15 In all three countries, there are significant vacancies in key professional posts. Although data is not available in a form that enables cross-country comparisons, there are some key features that stand out: · Vacancies are, for the most part, prevalent in non-managerial senior technical positions, where staff may seek better opportunities in the private sector (engineers, doctors, etc). Particularly in Kenya, vacancies in senior staff positions are a problem, with vacancies ranging between 30 and 50% of all personnel in salary scales 1-9. · Job stability is an issue for senior managers. The senior manager (head of district administration or head of finance) has a "hot seat"; in Kenya and 43Sample data from visited rural LGs (see country outline reports and annexes for details). 63 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Tanzania these officers are very frequently transferred, and in Uganda LGs they frequently dismiss or interdict these officers. · In both Uganda and Tanzania there are very distinct geographical disparities in staff allocations. In each country, specific marginalised areas can be identified where conditions (lack of access to services or insecurity) make it very difficult to attract competent staff. In Kenya, the main disparities are between the larger urban LGs and smaller rural LGs. 5.4 SERVICE CONDITIONS AND HRD PRACTICES IN LGS 5.4.1 Remuneration Salaries of LG staff are in Tanzania and Uganda centrally determined (with rare or minor exceptions for urban LG staff in Uganda). LG staff salaries thus follow central government staff. In each country there is some debate regarding how different staff positions should be graded ­ in particular, senior management positions in LGs and central government. Pension schemes vary between central government and LG employees, and in both countries the management of LG pensions has faced various problems and is an area that LG employees often complain about. Although LG salaries in Tanzania and Uganda are determined centrally, LGs may, according to the laws and regulations, still introduce their own additional incentive systems. Only a few LGs have launched substantial additional pay packages (Kampala is a noticeable exception, as they pay their Clerk more than a CG PS), but several LGs have established other benefit schemes, including housing, loan schemes, etc. As most LGs rely so substantially on central government transfers, it is obviously only the stronger LGs with substantial own source revenue that can effectively manage pay schemes. Weaker LGs may find it even more difficult to attract qualified staff. In Kenya, salaries and other benefits of LG staff are negotiated between the unions and the Association of Local Government Employers (ALGE). The salaries between LG and CG staff in Kenya differ substantially, and they are generally higher in LGs than CG. However, CG employees receive higher housing allowances and other benefits that, to a large extent, may counter the higher LG salaries. The major complaint at the LG level for senior staff is not so much about salaries, but rather about the general working conditions (see 5.4.3). 5.4.2 Wage Bill and Pay Reform Salaries in the public sector are generally perceived as "low" and, in particular, for several mid-level technical staff categories it has been documented that the pay is substantially lower than in the private sector. Thus, particularly in Uganda and Tanzania, more explicit pay reform efforts have been attempted. While very substantial increases took place in the 1990s44, it is fair to conclude that further pay reform efforts have stalled in all three East African countries. All three face substantial pressure to limit further public expenditure, as overall economic growth, population growth, and revenue collections cannot afford more. Kenya has in general 44E.g. in Uganda teachers monthly salaries increased from 3 USD to 69 USD over six years (1991-96). 64 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda a relatively "expensive" public service and Uganda and Tanzania have, after a period of "right sizing", seen a general expansion of their public services ­ mainly through the recruitment of teachers and other staff in the social sectors as the countries pursue PRSP goals. Further pay reform has therefore to be highly selective. However, pursuit of such targeted reforms, based on highly technical criteria, has proved to be extremely difficult in the face of political interventions45. Figure 5.1 Government Wage Bill as % of GDP 10 9 8 7 6 5 4 3 2 1 Kenya 0 1998 Uganda 1999 2000 Tanzania 2001 Tanzania year 2002 Uganda 2003 Kenya Data source: WB Database and IMF as quoted in PER Kenya 2003 5.4 3 Personnel Management Practices A decentralized personnel management system is only fully established by law in Uganda, and it is only slowly taking root. The established District Service Commissions have started to work, LG Councils and their staff have started to understand the rules after a number of LG staff complaints have been settled in court (especially complaints from Chief Administrative Officers regarding unlawful dismissals). Some of these court decisions have been very costly for LG Councils and have made them much more careful in following rules and regulations. The devolution of personnel is not even completed in Uganda where, for instance, systems for decentralized payroll management are only implemented on a pilot basis. The LG personnel management systems in Kenya and Tanzania remain substantially centralized, and in Tanzania the system has in recent years been characterized by a great deal of uncertainty ­ with LGs receiving conflicting circulars from various 45Kithinji Karagu and Rwekaza Mukandala: Public Service Pay Policies, PwC, Draft 2004. 65 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda central government offices. Staff transfers continue to be undertaken very frequently, especially of senior management staff. Recent statistics suggest that Tanzanian LG Treasurers, on average, stay less than two years in the same post; while in Kenya it is not unusual to have the Clerk changed more than once in any one year. Modern systems of staff performance appraisals have been only recently introduced in all three countries ­ and the experiences have not yet been analyzed. No comprehensive studies have been undertaken in any of the countries to analyze HR management issues in depth. Some attempts have been made to benchmark practices in Tanzania, but no reports have been published. In all countries the Inspectorate Departments have made some observations on LG personnel management practices, but in a far less systematized manner than observations on financial management issues. The topic certainly warrants more systematic analysis in all three countries. 5.5 LG RESTRUCTURING In all three countries the need for restructuring of LGs is recognized, but for different reasons. In Uganda and Tanzania, the reforms and subsequent changed responsibilities of LGs have required a rethink of LG structures. In both of these countries, the current LG structures are, to a large extent, the outcome of central government decisions rather than local priorities, although some distinct local differences in employment at the lower levels have been determined by LGs. Central government transfers pay the majority of LG personnel in Tanzania and Uganda, and local decisions regarding the size of LG staff establishments can in practice not be made by LGs without central government support. The systems have, in this manner, some perverse incentives for LGs to lobby for increased personnel, as increased numbers generally lead to increases in the fiscal transfers. In Kenya, the need for restructuring mainly arises from a long history of LG employment of lower cadre staff. Most LGs in Kenya recognize that their staff establishments are "bloated": that too many lower level cadre staff have been employed and undertake functions that may be better undertaken by the private sector. The approaches to restructuring of LGs differ substantially. In Kenya LGs are largely left to their own devices to restructure their organization and retrench staff. Some incentives are provided through LATF for LGs to restructure (funds are not provided to those LGs that spend more than 60% of their budgets on personnel) ­ but no direct technical support. In Uganda, the Government currently seeks to establish some "standard staffing structures" in LGs. A consultancy under MoPS looked into what would constitute appropriate staff structures for different categories of urban and rural LGs. The categories of LGs are based on population, land area and the size of LG revenues. The 66 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda proposed new model structures will require very substantial additional central government funding for the LG wage bill ­ and the proposal is still under debate. In Tanzania, the LGRP has initiated a process of restructuring of LGs through an elaborate, facilitated process that so far has been completed for only a third of the LGs. The 17 steps process includes development of strategic plans for each of the LGs, based on local consultations and "customer surveys". Each LG then determines what functions should be prioritised, what functions could be outsourced, etc. Staff audits are undertaken, and measures are explored to retrench unqualified staff. The great benefit of the approach has been the emphasis on the particular situation of each LG. The drawbacks include: high costs for facilitation of the exercise, the slow pace (four years to complete one third of the LGs), uncertainty regarding the institutional framework for LG personnel management (who has the right to hire and fire?) and lack of commitment upfront for financing the proposals. LGRP has recently calculated the costs of the recommended retrenchments, but it has not yet fully established the costs of filling the many vacant positions. A redeployment of teaching staff has recently been considered but not effected ­ the analytical work in support of this exercise documented substantial flaws in central government management, as teachers were found to be very unequally allocated across the country, with a strong bias in favour of urban and richer LGs. A common feature of the two exercises in Uganda and Tanzania is that they recommend substantial increases in LG employment with no prior budget approvals ­ thus practical implementation of restructuring proposals in both countries may take long to implement fully. 5.6 CAPACITY BUILDING IN LGS In all three countries the importance of LG capacity building is recognized, but the approaches to it differ substantially. In Kenya, it is realized that previous systems for LG capacity building have largely collapsed after the 1980s. It used to be based upon well-funded LGs requesting training from national institutions with wide ranging LG training courses. Under the KLGRP and other donor funded programmes some training and other capacity building of LGs has been initiated but on a rather ad hoc basis. LGs continue to fund training and equipment from own source revenue, and in recent LATF guidelines each LG is requested to spend a minimum of 2% on capacity building. However, this suffices for only a few short skill development courses. No effective central mechanisms for QA of training material, training providers, or assistance to TNA, are in place. The MLG is working on a LG training strategy. In Tanzania, there is, as in Kenya, no overall system for the guidance of LG capacity building (no training policy, etc), but the level of training and other capacity building (equipments, TA, etc.) is far more substantial ­ although so far entirely provided in the form of centrally defined LG training programmes and discrete donor funded projects. LGs in Tanzania fund a very small fraction of ongoing capacity building. 67 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda However, the recently formulated LGSP will establish a LG capacity building grant that will provide discretionary funds for LG demand-driven training. The most elaborate framework for LGs capacity building is in Uganda. A capacity building unit has been established within MoLG; it has facilitated the development of some 26 standardized modules for LG skills development; it works on QA and the oversight of private and public sector training providers; and it guides the financing of LGs' capacity building in the form of a discretionary LG capacity building grant that allows LGs to establish their own specific capacity building requirements on a demand-driven basis. In addition, CG supports LGs with a few supply-driven training programmes, such as financial management, councillor induction training, support to development of HR functions, etc. Evidence to date suggest that the discretionary capacity building grant, in combination with the incentive structures of the capital grant transfers (chapter 4), provide very effective mechanisms for capacity building of LGs in a manner that corresponds both to national standards and requirements as well as local needs and priorities. However, the system has also showed the need for a CG coordination and QA of the training materials and support to the establishment of strong HR functions within LGs to enable them to utilize the capacity building grants efficiently. 5.7 KEY LESSONS LG staff structures: LGs' autonomy in determining staff structures is rather limited in all three countries. In particular, in Uganda LGs have a clear legal mandate to structure their internal organizations, but they are effectively constrained by the degree of available own source revenues. In Kenya, LGs have less legal autonomy, but they have in practice, due to their higher degree of own funding, decided to undertake a substantial recruitment of lower cadre staff in a manner that is both unsustainable as well as un-conducive for service delivery. The Kenyan case thus serves as a warning for granting indiscriminate autonomy at the LG level for staff recruitment ­ when there is low accountability to the citizens. However, the problems in Kenya are also the result of (a) only granting autonomy to lower cadre recruitment to LGs; (b) lack of clear designation of functions to LGs; (c) lack of guidance and training of LGs to undertake restructuring; (d) perverse local accountability mechanisms where councillors predominantly are judged by the favours they can grant individuals rather than by their contributions to sound LG policies. Thus, there is insufficient pressure from both above and below to improve on performance and efficiency. LGs power to hire and fire differ significantly across the three countries. In Uganda, the power to hire and fire has been significantly devolved to LGs (to their District Service Commissions). Although the experience has not been substantially analyzed to date, evidence suggests that the devolved system has functioned relatively well. The central government, nevertheless, contemplates a centralization of appointments of the Chief Administrative Officers. In Kenya, the system is centralized for senior staff and decentralized for junior staff. This has functioned very poorly; LGs have an extremely high number of senior staff vacancies ­ with low morale and bloated lower staff structures. In Tanzania, this issue has been very contentious and still remains 68 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda unresolved, in spite of clear policy statements in 1998 concerning the devolution of staff management. LGs' ability to enhance staff performance is an assumed outcome of devolution of staff. This remains to be fully documented, but in Uganda ­ and to some extent in Tanzania ­ evidence suggests that this has occurred. The annual LG assessments under LGDP II have impacted positively on LGs' enhancement of staff performance, in particular in the areas of development planning and financial management ­ not least at sub-county levels. These improvements could not have taken place if LGs could not hold their staff accountable. Technical tools, such as the introduction of open staff appraisal systems, are introduced in all three countries ­ but the impact also remains unanalysed. LGs ability to attract staff has, in particular in poorer LGs, been curtailed by lack of financial resources. More wealthy LGs ­ in particular urban LGs ­ have been able to provide extra incentives in the form of additional allowances or outright salary increments. LG salaries are in Uganda and Tanzania similar to those of central government employees; whereas in Kenya they are negotiated locally and are substantially above central government staff (but other allowances are lower). The main factors that determine the LGs' ability to attract staff are thus job satisfaction and security, possibilities for promotion and staff development, and facilities at the duty station. The problems of attracting staff for certain technical categories to remote and poor LGs appears chronic in both Uganda and Tanzania and ways and means to resolve this, e.g. by introduction of special allowances or other targeted benefits, should be considered. Capacity building can successfully be undertaken on a demand driven basis if an appropriate system of financing, incentives, oversight and quality control is developed. However, donor-funded support to LG capacity building nevertheless still overwhelmingly supports centrally-managed and supply-driven capacity building efforts. Restructuring; the determination of LG staffing structures can probably not be undertaken without substantial central guidance, in a manner where several, often parallel, processes are coordinated. The various often conflicting concerns that need to be balanced include: · Delivery capacities of specific sectors ­ generally pushing for increased staff in each sector; · Macro-economic concerns regarding tax levels and subsequent arguments for limitation on public sector employment; · Pay reform issues, where efforts are made to target salary increases in strategic areas; · Balancing needs of CG and LGs; · Pressure from various interest groups ­ in particular, labour unions. The MoPSs traditionally do coordination, but it may require more involvement of MLGs, sectors, and LGs 69 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda 5.8 CHALLENGES AND EMERGING ISSUES LG HR management remains an area with many challenges, and it is at the centre of the debate on devolution. Figure 5.2 below outlines the intended accountability mechanisms in LGs. The core assumption of devolution is that locally elected LG councillors are effectively accountable to the local citizens, and that this in turn: · Will foster administrations that plan and budget in accordance with local needs and priorities, so making resource allocations more cost effective, · Will ensure local accountability in the use of resources in a manner whereby corruption is curtailed and allocated resources are spent according to plans and budgets. Intended Local Accountability Mechanisms Figure 5.2 The Local Government Local government staff collect taxes and provide services erfi and einlpicsid dna ot rts reih, po re rdaw re different forms of participation: votes, questions and maybe recalls Councillors (and LG Boards) Monitors Informs on performance resource and gives availability incentives for and LG improvements performance CENTRAL GOVERNMENT (PO-RALG/MoLG, MOF, AG) Of the three countries, Uganda has most clearly introduced a system for managing staff that can enable the above accountability mechanism to be effective. Kenya and Tanzania are both committed at the policy level to move in a similar direction. There are some general issues that need to be addressed in the design of a decentralized staff system. These were raised in the introduction of this chapter. But, 70 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda more specifically, in all three countries some distinct challenges have recently emerged: · In Uganda, it is currently being debated whether to centralize the management of the CAOs, yet experiences from the current system are not well documented; evidence suggests that the system only gradually is being established and starting to yield results in the form of local accountability of staff. However, the discussions indicate a need to continue to support improvement in the relationship between the administration and politicians in LGs. · In Tanzania, a Public Service Act was recently passed that overruled preliminary attempts at the devolution of staff to LGs; · Support to restructuring has recently been undertaken in both Uganda and Tanzania ­ while both recommend some retrenchments, they also argue for substantial increases in actual staff deployments in LGs ­ something not yet budgeted for in the MTEF; · In Kenya, the Constitution argues for the devolution of staff to LGs ­ but precise modalities for institutional arrangements need to be put in place. This will be a major challenge, not least since many types of sector staff are foreseen as being transferred to LGs. Finally, is should be noted that HR management in all three countries remains very under researched: · Basic statistics on LG capacities across LGs (staff vacancies, staff qualifications, etc) are not available in any of the three countries in summary reports. In general, the MLGs maintain some records of staff situations in LGs, but often with data in a non-comparable format ­ and data has never been aggregated and analyzed. · Statistics on basic practices (transfers, disciplinary cases, promotions, etc.) are usually available for CG staff ­ but not for LG staff. · Most importantly, no empirical analyses have been undertaken of the assumed relationship between devolution of staff management, local staff accountability and staff performances in delivery of services ­ something that is central to the LG reforms. Uganda is the obvious case for such a study, as the system for devolution of personnel was introduced sufficiently long ago to have impacted, not only on management practices but also on staff performance. 71 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda 6. INSTITUTIONAL ARRANGEMENTS 6.1 INSTITUTIONAL ARRANGEMENTS One of the often overlooked but very important factors, contributing to the prospects for a successful decentralisation reform process, is the institutional arrangements/ coordination surrounding decentralisation46. During the country studies, these were explored, and some of the main results appear in the table below: Table 6.1: Overview of the Coordination of the Reform Process Issue Kenya (1) Tanzania (2) Uganda (3) Agents for KLGRP within the MLG. LGRP ­ part of the MoLG. coordination of Integrated part of the ministry but separated No overall reform unit after the LG reform ministry. physically. the phasing out of the process decentralisation secretariat. Small separate support Larger unit, with more A number of coordination unit ("semi autonomous" than 30 professionals agents: PCU/MoLG, status) LGFC, LBCC, LGROC, LRECC, (especially on fiscal matters) Government ­ Ad hoc Joint govt-donor Current (not frequent) donor - Consultative Forum, meetings between MoLG coordination chaired by the PS/PO- and the Decentralisation RALG. Donor Sub-Group (DDSG), Common Basket Fund based on specific needs and Steering Committee, requests IMTC, IMWG Donor ­ donor Ad hoc informal meetings Donors participating in DDSG has regular monthly coordination the Basket Fund for the meetings LGRP Linkages to the Limited Some, although there are Limited overall public examples (Public Service sector reform Act) which show the need process for further improvement Linkages to the Limited Generally limited. Limited. sector reform However, the LGRP legal This has been noticed, and process harmonisation component a decision to ensure a better tries to streamline the LG coordination between the Act and the sector laws decentralisation and the sectors will be initiated under the Joint Annual Review of Decentralisation process (JARD) Unit for No No, but it is partially Yes, recently established coordination of coordinated by LGRP under MoLG to ensure capacity to LGs more coordination and quality in the CB rendered to LGs 46Please refer to the DAC Evaluation Series: "Lessons Learned on Donor Support to Decentralisation and Local Governance", Schou and Steffensen, OECD DAC Network on Development Evaluation, 2004, for a review of general experiences of government­donor and donor­donor coordination. 72 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Issue Kenya (1) Tanzania (2) Uganda (3) Joint Not yet. Some Joint reviews in Has started recently with programming, corporation has been connection with reviews the FDS preparatory work reviews, initiated between the WB of the LGRP and design and the preparations of the evaluation and and DFID on joint of the recurrent and LGDP. So far rather project monitoring reviews of development grant specific. decentralisation systems It is expected that the JARD will take the first step in this process 6.1.1 Institutions for Coordination The way the overall coordination of the decentralisation reform process is organised varies among the three countries. In Tanzania, there is a strongly capacitated Local Government Reform Programme Unit, with task managers and support staff within each of the key areas of reform, zonal support reform teams in six regions to spearhead the reforms in the "batch one" reform councils and to support the CB of LGs. Mechanisms for the coordination of sector reforms has until recently been weak, and the establishment of the separate LGRP Unit has on occasions made coordination difficult within PO-RALG. As an example, there are the two different approaches to LG development planning ­ strategic planning as an outcome of the LGRP process and O&OD planning under PO-RALG. Institutions such as an inter-ministerial LGRP Committee have not performed to the extent it was hoped they would, as the entire government is overburdened with the coordination of many different reform efforts. Recently, more emphasis has been put on sector coordination by the recruitment of a task manager to facilitate this, and with greater attention paid to the issue in the Inter- ministerial Steering Committee. Within the field of LG finance, Tanzania has an emerging system for coordination under the established committees, developed in connection with the reform of the recurrent grant system. It is currently considering the establishment of a more permanent body. The experience has been that this is of the utmost importance if fiscal reforms are to be efficiently implemented. In Uganda, the general coordination of the decentralisation process is under the MoLG, and it was in the beginning spearheaded by a strongly facilitated Decentralisation Secretariat47, which now is in the process of being phased out. Within the field of LG finance, Uganda has a comprehensive institutional set-up, spearheaded by the Local Government Finance Commission and the newly established coordinating bodies for the FDS implementation and the transfers systems (LGBCC chaired by LGFC) and for releases and operations, accounting etc. (LGROC chaired by MoF) and revenue enhancement coordination (LRECC chaired by LGFC). All these committees have representatives from the key ministries as well as from LGs, and they have been very instrumental in spearheading and coordinating recent reforms. In Kenya a small reform unit within MLG (KLGRP), supported by DFID, is spearheading and coordinating the reforms. Except for an advisory committee for the LATF grant scheme in Kenya, there are no permanent bodies for coordination of LG fiscal reform matters (negotiations with LGs, analysis, reform policy development, etc.). 47Danida was the first donor to support the DS ­ support which went on to 2003. 73 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda 6.1.2 Government­Donor Coordination The Government­Donor coordination has been a challenge in all countries, partly due to lack of capacity in the key ministries, partly due to the numerous donors operating within the field of decentralisation with different modalities for donor support, not always synchronized and aligned with the government procedures. Experience has shown the value of strong coordination, especially in the design of joint programmes such as development grant and capacity building programmes. The donor­donor cooperation is most institutionalised in Uganda, where there is a permanent, all-encompassing Decentralisation Donor Sub-Group DDSG. In Tanzania, the coordination is regular amongst the countries involved in the Basket Fund of the LGRP; however, there is no permanent all-encompassing DDSG. In Kenya, it is carried out by ad-hoc dialogues between "like-minded" donors. Experience has shown that strong coordination can reduce the transaction costs both for CG and LGs, for example, by having joint review missions, by mainstreaming and harmonising of procedures (it is often easier and better to harmonise around government procedures than around one of the procedures of one of the donors). However, as for the area above, there is room for improvement. 6.1.3 Decentralisation and the Wider Reform Agenda The linkages between the decentralisation reform process and the sector reform processes have generally been weak in all three countries studied, a problem that is well-known from other places as well.48 However, there are also similar problems in the linkages with the wider issues of public sector/civil service reforms. First, decentralisation has had a minor role in the PRSP and in the PRSC dialogue. Second, the links between the Sector-Wide Approaches (SWAPs) and the decentralisation reform process have been blurred, and the processes have been carried out largely in parallel without proper coordination. The LG reform process has not been sufficiently linked with the reform plans within the civil service system, including the pay reform discussions. In Uganda, these remedies have been noticed, and the recently launched LGDP-II is expected to address these links by supporting annual reviews of the decentralisation process (stock-taking exercise), by reviewing the links to the sector activities and the contradictions in the reform progress, and, ideally, by "paving the way" for a decentralisation poverty reduction strategy credit (D-PRSC)49, a type of budget support, with related objectives, targets and on-budget funding scheme, closely linked to the sector reforms. These initiatives are urgently needed, as many of the sector policies, guidelines and concrete support activities have contradicted the overall decentralisation objectives, for example, the establishment of parallel technical support/implementation units within the water sector, legal inconsistencies, etc. The progress in this area would be of great interest for other countries that are preparing joint programming and funding of LGs. 48Cf.. e.g. UNCDF, by Paul Smoke, 2003, op cit, Schou/Steffensen, 2004 op cit, and David Watson, DFID Paper prepared for the AGF V working group, 2002. 49It has been expected that this credit will take over from the Second LGDP, but the experience of this programme should be reviewed during the midterm review in 2005/06. 74 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda 6.1.4 Coordination of Capacity Building Capacity building activities have been numerous, scattered and hardly coordinated in all three countries. They have been less pronounced in Kenya due to limited resources and capacity to support the LGs. There has been no overall coordination and clear donor entrance for support, and an absence of a framework to ensure high quality, non-overlapping, resource cautious approaches. Based on the experience of these problems, and the experience from capacity building grants on a demand-driven basis to LGs, Uganda decided to elaborate a capacity building framework, including establishment of a CB unit ­ and later a policy, strategy and common coordination. Among the first steps was the development of common training materials, which have been designed with support from most of the donors, and with the involvement of LG/CG/donors in the steering and quality assurance. There has also been support to the LG human resource functions, to enable them better to utilise the demand-driven CB grants. These activities are promising, but the momentum needs to be kept up. 6.1.5 Joint Programming, Reviews and Evaluations There has been remarkable progress in the areas of joint programming, reviews and evaluations over the most recent years. In Tanzania, the LGRP is funded by a joint donor basket fund composed of nine donors, and the programme has currently been jointly reviewed and a cooperation has taken place concerning the preparation of the coming LGSP. In Uganda a joint donor basket fund was established to finance the following activities: i) the Fiscal Decentralisation Study, which provided pertinent input to the government´s Fiscal Decentralisation Strategy (FDS); ii) later, the preparatory work of the LGDP; iii) the studies of the links between the LGDP and the other donor programmes financed by a joint donor basket fund; iv) the initial activities to support the FDS implementation (budget guidelines, grant allocation formulas); v) Training materials were co-financed by several donors, and five donors are presently co- funding the support to the LGDP-II directly ­ and others through parallel but well linked and coordinated activities. There are still examples of overlapping activities, and weak coordination, but the general picture has improved tremendously. In Uganda and Tanzania, the future challenge related to integration and mainstreaming of LG support programmes will be two-fold: i) to ensure that the district support programmes are mainstreamed and ultimately merged with the GoT development grant schemes, ii) to ensure that there is a process whereby the sector development grants, and the project specific development grants, are gradually merged with the LGSP/LGDP frameworks and respective funding modalities ­ which will be in line with the overall GoT/GoU principles for the transfer of discretionary capital development funds to LGs. In Kenya, the major challenge will be to follow up on the results from the Constitutional review process, which envisages a transfer of several key service delivery functions to the LGs. An entirely new funding system, with proper links to the sectors, needs to be elaborated. The transfer scheme ­ LATF - in an expanded and adjusted format, can probably provide the vehicle for the future donor support to development investments. Only few donors are involved in the support to decentralisation, and these have informally attempted to coordinate various 75 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda initiatives and reviews. However, the challenge will be in the joint programming of LG support after the adoption of the Constitution. 6.2 LESSSONS AND CHALLENGES Derived from the comparison of the experiences, the following lessons and challenges appear to be of major importance and may be summarized as: · The role of a strong MoLG is critical for overall coordination of decentralisation reforms. However, substantial decentralisation reforms require high-level political guidance and coordination. Decentralisation requires reforms of the entire public service, and all sectors, in a manner that is beyond the mandate of a ministry such as a typical Ministry of Local Government. The limits of "MoLG" type of coordination became evident recently in Tanzania, with the passing of the Public Service Act ­ and the experience can easily repeat itself. One way to ensure high-level political guidance to reforms is to have some form of cabinet committee established, or have the reform principles entrenched in binding documents ­ if not in the Constitution itself. · The role of a strong and decentralisation oriented Ministry of Finance, especially to ensure a well-functioning, predictable and timely intergovernmental fiscal transfer system and to ensure links to and involvement with the sector ministries is important. · Gradually merging of the donor supported district/areas based capital programmes with the overall government intergovernmental grant schemes, and harmonisation of transfer modalities (planning, budgeting, accounting, reporting and auditing) around the government system and procedures · Forums for joint government-donor coordination may facilitate the reform process, but meetings need to be regular and well prepared. · Donor coordination is critical to reduce transaction costs for all parties. · Joint reviews, evaluations, preparation of programmes, are instrumental to coordination and facilitation of programme implementation. · The weak links in the overall public administration reform process, e.g. the civil service reforms, restructuring, budgeting reforms, etc, and the sector approaches are areas in need of urgent attention ­ it will be important to ensure that other reform areas, especially the SWAPs, are not undermining support to decentralisation. · The key ministries for the operational coordination of LG reforms, the MoLG/MLG/PO-RALG/Ministry of Finance, need to be strengthened and facilitated to enable them to spearhead, coordinate, supervise and monitor the overall decentralisation reform process in their interface with LGs. 76 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda 7. CONCLUSIONS AND AREAS IN NEED OF FURTHER REVIEW AND SUPPORT This Chapter summarises the factors that have promoted and constrained the decentralisation process in the three countries, with special focus on the LG incentives to perform towards realisation of the development goals. The assumption, supported by qualitative reviews in the field, is that a good out-turn on these main dimensions (policy and legal framework, political and administrative structures, LG finance, HR and coordination of the reform process) increases the likelihood of efficient decentralisation processes impacting positively on the service delivery and overall development. It finally gives a brief outline of some of the main considerations that may be addressed in the future reforms by the governments and the development partners. The chapter also contains some concluding remarks based on the findings from the previous chapters. 7.1 INCENTIVES FOR LG SERVICE DELIVERY AND DEVELOPMENT50 7.1.1 General Overview of the Incentives The table below provides a summary of the factors (incentives), which have promoted the decentralisation process in the three countries studied: Table 7.1: Overview of the Status of the Factors Promoting LG Incentives Dimen- Positive Kenya Tanzania Uganda sion incentives Legal and Conducive Constitution No; but is under No; very weak on Yes; although in policy elaboration LGs some areas over- dimen- detailed Clear legal framework Overlapping Not very clear, Rather clear sion for assignment of responsibilities and especially for the framework, expenditures parallel provision LLGs, and scattered although weaker in several acts. for LLGs LG Act harmonised LG Act is by-passed by Harmonisation has Yes; however, with sector legislation sector acts. The sector been initiated, but problems within legislation dominates still challenges, e.g. specific areas + in the civil service planned initiatives Clear and coherence No, Yes (1998), but is To be developed. Decentralisation not up-dated The FDS (2002) Strategy and covers most of the implementation plan LG finance issues 50This section has benefited from the methodology outlined in the recent publication: Devolution in Pakistan - An Assessment and Recommendation for Action, ADB, DFID, The World Bank, May 28, 2004, p.23-25, op. cit, but uses a number of other dimensions and indicators. 77 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Dimen- Positive Kenya Tanzania Uganda sion incentives Admini- Overall structure No ­ great disparity Yes, at the HLG, but Generally Yes; strative ensures capable LG across LGs, but the not at the LLG. The however, the and units of optimal size coming reform will LLGs are rather gradual increase in strengthen the district small (3,300 inh.) no. of LGs has put political level and increase the for efficient service pressure on the structures size and introduce provision scarce capacity. LLGs The role of the administrative units need clarification. Clear roles of No Common problems A challenge, but politicians vis-a-vis the in mutual has improved over administration understanding of the the recent years roles No interference from No. No. Yes MPs in local matters The interference is seen Interference is seen (complied with) as a great problem as a problem Well functioning No, however recent Weak. Yes. statutory bodies to improvements in the It is under However, still with ensure separation of Tender Boards establishment, e.g. room for power/non-political TBs have been improvement, as interference in technical recently reformed politicians tend to matters interfere Involvement of citizens Weak, but improving Weak, but Moderate, and in planning and improving improving decision making Finance Clear Fiscal No No, but under Yes (FDS). Decentralisation development. First However, the plan Strategy with steps taken on the is very much implementation plan systems of recurrent focused on the and sequencing and development transfer system, i.e. grants less on reforms of the own revenue sources Clear assignment of No Partly. However, Reasonably clear. expenditures still unclear areas, However, less clear e.g. in relation to for the LLGs sector laws and concerning LLGs Autonomy in budget Budget approval Reasonable. Reasonable. preparation process delays However, the However, the decisions. conditional grants conditional grants CG interference in put strong limits on put strong limits on specific decisions LG priorities LG priorities Buoyancy and potency Moderate. Weak; Moderate/weak: of taxes assigned However, there is a especially after the There is a great great non-utilised abolishment of a non-tapped potential, partly due to number of taxes. potential especially weak LG capacity to Property tax in the in the urban areas collect, CG urban areas has a (property tax) interference and significant potential problems with the property tax High level of own Yes >80% No <15% No < 15% source funding 78 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Dimen- Positive Kenya Tanzania Uganda sion incentives Good incentives for Reasonable. Low. Low. revenue raising Grants are small. No tax effort criteria However, recent Incentives in the major in the present initiatives under grant scheme to system. LGDP have improve on LGR High dependency on addressed this as grants well as CB and other means. The major grants are not linked to the tax efforts and contribute to 80% of total revenues Transfers predictable, Yes (for the major Yes. Yes. transparent and timely grant: LATF). In a process of being In a process of There are however formula-based improving the problems with the formulas. CILOR and RMLF Some minor delays generate problems Formula based Transfer systems are LATF has taken Not yet. (emerging) Partly. performance-based to important first steps in However, the The LGDP and the promote strong that direction coming LGSP recently initiated incentives for efficient capital grant system FDS have a service provision will have a number number of in-built of in-built incentives incentives. However, not yet expanded to the major sector grants Strong incentives to Previously no; but has No; but will Reasonable under improve on financial improved significantly hopefully improve the LGDP scheme, management with the LATF with the LGSP, supported by a which includes these number of incentives oversight functions Strong oversight No. No. Moderate. supervision, well E.g. the audit has great However, the audit However, great coordinated backlogs and audit follow-up problems with the have improved coordination. There is a weaker effort concerning the LLGs. Human Autonomy to hire and Partly. No. Although the Yes, through the resources fire Only possibilities for staff regulations individual LG the lower staff positions issued in 2000 District Service provided some legal Commissions basis, they have constantly been undermined by various CG initiatives that have limited LG in practice Autonomy to discipline Only for lower level Partially Yes cadres 79 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Dimen- Positive Kenya Tanzania Uganda sion incentives Satisfactory career path The entire LG staff Reasonable career Reasonable for LG staff system to a large extent possibilities. possibilities for discredited; thus not a However, senior career very attractive career staff express development, path concerns regarding although the most full devolution of senior staff feel the systems, as they constrained for see promotion employment within within CG as one single LG desirable rather than potential promotions through transfers Good salary incentives LG staff salaries better LG and CG salaries LG and CG salaries ­ the LG salaries are at than CG salaries. Other generally similar, similar. Some the CG level and good benefits are better in but e.g. SASE only limited scope for possibilities for CG employment extended to CG introduction of rewarding good staff. In theory, local incentives ­ performance possible for LGs to mainly benefiting introduce local staff in financially incentives, but this stronger LGs is only applied on a marginal scale Regular payment of A frequent problem in Yes. This is no Yes. This is no salaries some LGs longer a problem, longer a problem, except for newly except for newly recruited staff as recruited staff as they wait for they wait for registration on registration on central payrolls central payrolls Satisfactory pension NO, No, but under No, and retirement benefits LGs fail to pay improvement. General crisis of contributions. Attempts LGs default in pension system and to solve problem contributions. major problems in through debt recovery Central government transfer from initiatives (LATF). now deducts. LAPF central to local under restructuring. pension systems. No No. There is a Partly - some Yes. CG interference in day- substantial interference improvements, but CG interference is to-day business on a regular basis. Staff still substantial minimal and are transferred very interference in the mainly in the form frequently without strict form of frequent of guidance adherence to regulations transfers and inspections. A few contradictory cases of IGG instructions issued instructions to LGs by different parts of have, however, CG been observed Conducive interaction Great variation across Some Has improved between politicians and LGs improvements, but substantially ­ staff still substantial recent court cases problems have clarified the boundaries of rules and regulations Autonomy on the salary Strong control of the No; as most of the No; as most of the budget upper staff levels. personnel are personnel are Substantial autonomy in financed by financed by recruitment of lower conditional grants conditional grants cadres 80 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Dimen- Positive Kenya Tanzania Uganda sion incentives Institutio Committed and strongly KLGRP; but not LGRP/P0-RALG; MoLG needs nal facilitated champion for particularly strongly however, not stronger capacity to arrange- decentralisation situated in the wider particularly well coordinate the public reform context integrated in the wider reforms ments for ministry or in the decentrali wider public reform -sation process Strong government- No. Reasonable. Moderate. donor arrangements for Ad hoc However, these are Meetings are not coordination mostly concentrated regularly within one programme (LGRP) Strong donor-donor Ad hoc, informal Reasonably strong; Rather strong, with coordination however, mostly a permanent set-up (principles, programmes concerning specific and well prepared etc.) programmes (LGRP, meetings. LGSP, etc) (DDSG) Proper links between No No No. decentralisation reforms Initiatives have and sector reforms been taken and will be followed- up on this Proper links between No No No decentralisation reforms and the wider public reform (civil service etc.) Common processes of No Yes; but mostly Yes; however, preparation, design, surrounding the there are still reviews and monitoring LGRP and LGSP numerous bilateral to improve coordination donor reviews, and reduce transaction which need further costs for all parties coordination a. As observed from numerous reports and confirmed during this study. The table above provides a broad overview of areas where progress has been made, and where there are great challenges in establishing strong LG incentives. Although these indicators, listed in the second column, have played an important role in all three countries, the issues cannot be seen without incorporating the country specific context. The country specific experiences have shown that51: 7.1.2 Kenya Factors that have promoted good incentives for LGs: · The introduction of the LATF has generated funds for discretionary LG funding linked to improved LG incentives on financial management, transparency and debt recovery. The pre-determined and predictable allocations from LATF have improved the possibilities for proper planning and budgeting and improved the incentives to make real local priorities in service delivery. 51Please refer to the country studies for the three countries for further details, as this chapter only attempts to summarise the key ones. 81 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda · Discretionary funding (from LATF and own source revenues) has enhanced the political dialogue on the spending priorities and improved involvement of all stakeholders. · Recent initiatives to streamline and simplify the taxes ­ the introduction of the Single Business Permit ­ have been a positive factor for the LG revenue mobilisation and linkages to the private business environment. · Initiatives to improve on the participatory planning process, with the introduction of the local authority service delivery plan (LASDAP), and inbuilt incentives in the transfer system to inform and involve people in the planning process have shown remarkable results in a system which, historically, has been characterised as been very "closed". · The KLGRP has definitely played a very strong role in promoting the most recent reforms and in strengthening of the intergovernmental relationship. One of the strengths of the KLGRP is its close alignment with the MLG as an integrated unit within the ministry, rather than a typical project management agent/institution, and a committed unit of core staff. Factors that have particularly constrained the process: · Lack of a decentralisation strategy ­ this has created some lack of clarity in the direction and focus for CG and LGs. · Lack of a conducive legal framework with clear guidance on the range of LG autonomy, existence of many approval procedures ­ these have especially led to delays and lack of incentives for LGs to take their own initiatives. · Cumbersome budget approval processes without a clear framework for ceilings and general "rules of the game" ­ have led to lack of LG incentives to improve on the budget execution and to use the budget as an actual planning tool during the fiscal year. · Lack of incentives to comply with basic regulations, lack of oversight/ sanctions for non-performance, and lack of staff incentives due to HR resource arrangements (especially the lack of LG control over the staff) ­ have all led to deteriorated financial management performance during the 1980s and 1990s. · Structural problems in the larger cities, which are heavily indebted to an extent where incentives to improve are absent. · Limited development of the legal framework and practices around one of the potentially high yielding taxes ­ the property tax. · Strong CG control of the rates on tax and user fees have reduced incentives to improve on tax mobilisation performance. · Lack of autonomy on staff matters. The interference from CG in staff matters, transfers of staff due to political and other reasons, have led to frustrations and lack of initiatives for the staff and lack of capacity building efforts, as LGs cannot be sure to capitalise on the investments in HR development. · Weak links between the LGs and the sector activities. · Breakdown of the capacity building activities in the 1990s. 7.1.3 Tanzania Factors that have promoted good incentives for LGs: · A clear decentralisation policy (1998) from the onset of the reform process. 82 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda · A committed unit ­ the LGRP ­ has spearheaded the LG reforms; however, with significant related administrative costs. · Increasing efforts by MoF to provide intergovernmental fiscal transfers in time and to inform about the amounts. · Strong focus on capacity building from the onset of the reforms in parallel with the systemic reforms. · Legal harmonisation reforms have ironed out certain contradictorily laws and regulations ­ however, not optimally within e.g. the civil service. · Recent improvements in the transfer system, which have become more predictable, transparent and timely, and which will (2004/05) improve on the allocation criteria and increase autonomy. · Recent improvements in the allocation criteria for the government development grants scheme, and agreement on the design of a programme with gradual devolution of the development budget (LGSP), linked to LG performance incentive systems. · The foundation (analytical work and preparations) for a comprehensive restructuring of the reforming LGs, with structures based on local priorities and needs ­ however, the funding issue is still to be resolved. Factors that have constrained the process: · Lack of continued clear political guidance of the reforms; the 1998 LG Reform Policy has never been updated and the commitment of e.g. sector ministries and CSD to the policy has at best been questionable. The status of the LG Reform Policy vis-a-vis many other policies ­ e.g. sector specific policies and policies for the Public Sector Reform is unclear ­ different ministries have advocated for different concepts of decentralisation. · Delays in the systemic LG fiscal reforms, which have constrained the LG autonomy on the finances, and reduced their discretionary power. This has, at the same time, had a negative impact on the interests in planning and budgeting, as limited resources are available for local priorities. · Systemic problems within revenue assignments, with a weak revenue base and potential in the rural LGs, and weak capacity in the urban areas to utilise their wider tax base, especially the property taxation. · Lack of incentives for LGs to improve on financial management, as the transfer system is rather automatic, without performance incentives, coupled with weak pressure from below (downwards accountability). · Lack of uniform planning tools applied by all stakeholders ­ many tools tend to be rather project specific. · Hitherto fragmented, non-coordinated programmes for devolution of the LG development budget, with numerous district, sector, issue-support programmes not properly harmonised and mainstreamed. A recent focus has been put on improving this, through the LGSP initiative and the GoT´s reform of the development transfer scheme. · Lack of LG control over the human resources and unclear legal framework ­ this is perceived as one of the main remedies for improvement of the LG staff incentives. · Lack of linkage of the LG reform to the wider public reform issues, including reform of the civil service (pay schemes, incentives etc.). 83 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda · A structure with very small villages, with limited capacity and opportunities for service provision. 7.1.4 Uganda Factors, which have promoted good incentives for LGs: · Strong political will from the top level for decentralisation. The experience has also shown that the public announcements from the top political level have a strong impact on the LGs performance. · A number of champions of decentralisation (from the onset, the Decentralisation Secretariat, later the Local Government Finance Commission, the active facilitating role of the associations of local authorities, and the committees established under the fiscal decentralisation strategy implementation framework) have raised the awareness of all involved, disseminated information, ensured compromises, and facilitated the capacity building and "mentoring" of LGs. Visits to LGs have confirmed the impact of these initiatives on the LG incentives to utilise the new opportunities. · Rather detailed legal framework clarifying the expenditure and revenue assignments in a comparable clear and coherent manner and the roles and functions of each actor ­ politicians and staff. · More recently: establishment of coordinative bodies, with involvement of all key stakeholders affected by the reforms of the transfer system, revenue enhancement, budgeting and accounting issues. These fora have facilitated the understanding, the involvement of LGs, and LG ownership of the reform process. · Establishment of a number of statutory bodies to regulate areas of importance for LGs (DSC, LGPAC, DTBs etc.) where there are reasons to de-link local decisions from direct political influence. · Strong focus on the involvement of the lower levels of local governments in planning and decision-making. · Strong performance incentives inbuilt to the development grant system under the LGDP, where the eligibility and the size of the grants are conditioned by the LG performance within key performance area ­ planning, budgeting, good governance, etc. It was observed that this has had a very positive impact on the LG incentives to improve on these critical areas for efficient service delivery, particularly when coupled with available funds for demand-driven capacity building ­ the lesson learnt is clearly that financial tools can be designed to improve on LG performance. · Certain autonomy related to human resources has been achieved, keeping a distance from the directly political involvement in hiring and firing. The compromise between the two sides ­ full autonomy and full control by the politicians (with the inbuilt risks in this) ­ seems to have been established in a rather successful manner. · Increase in the cooperation amongst all partners around common initiatives and programmes, for instance within the reform of the transfer systems and revenue mobilisation initiatives. 84 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Factors that have constrained the process: · Lack of an overall decentralisation strategy after 1997 and corresponding sequenced implementation plan to keep all stakeholders (coordinating ministries, line ministries, donors and others) on the right track. Development of such a plan is intended, with support from the LGDP. · Continuous separation of LGs (districts) prior to the building of sufficient capacity to split, and without prior analysis of the feasibility in terms of man- power, sustainability (fiscal basis) and economies of scale, and other factors. · Lack of fully up-dated, comprehensive, reliable and well-integrated, all encompassing data bases on LG finance statistics and other pertinent information. This has made it difficult to make informed choice within critical aspects of the reform work.52 · An overly complicated conditional grant system, with multiple modalities (planning, budgeting, accounting and reporting), has led to reduced downwards accountability and a tendency for LGs to do "paper for money" work. The system was created to ensure confidence that funds are utilised within the national priority areas according the PRSP targets. However, the detailed design has created some unintended incentives and remedies, which have required a new strategy ­ the FDS. · Lack of, and delays in, the legal reforms pertaining to the LG revenue assignments with sufficient strong incentives in the grant system to mobilise own revenue sources, especially in a situation with a large increase in grants, which presently constitutes approx. 85% of the LG revenues. Among the problems observed in recent years have been increasing difficulties in funding the basic administrative positions in certain LGs, in funding of operational and maintenance costs derived from new investments and own co-funding contributions to ensure ownership in the development processes. There has been an increasing awareness and focus on this problem. A number of initiatives (capacity building, better incentives to improve on LG revenues, etc.) have been launched recently, and additional mitigating measures are planned for in the near future (especially concerning the legal framework and sharing of best practices in tax administration). · Lack of involvement of the key sectors in the decentralisation reform work and linkages of the reforms to the wider public administration reform process, including restructuring of the civil service system, pay reforms, etc. · Although oversight functions are important, the multiple and non-coordinated CG oversight functions (MoLG, LMs, Audit, MoFPED, donors, etc.) have created transaction costs and frustrated LG staff, taken away attention from the actually provision of services. 7.1.5 The Political Context of the Reforms The political contexts of the reforms are not a primary focus of this study, but some overall observations need nevertheless to be made. In Uganda, the decentralization reforms were introduced by the NRM Government in a situation where the new Government in 1986 had come to power through a long 52Initiatives have been taken to improve on this coordination under the MoLG/LGFC. 85 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda guerrilla war ­ but without a clear electoral mandate. The Government had to build a political base throughout the country, in competition with the older structures of the political parties that again had been moulded on long established religious and ethnic identities. The Resistance Councils, which later became the political wing of the LGs, served that essential political purpose. When the administrative elements of the reform were introduced in 1992, the Ugandan government had political reasons for transfer of substantial functions to the new LGs. The government could also proceed with the reforms in a quite radical manner, as the existing public administration was so devastated by several years of war and misrule that the bureaucracy itself could present only limited resistance to reforms. The public service was in such a poor state that the government indeed had to proceed with public reforms in a radical manner. In Tanzania, the reforms occurred against a very different political background. The LG reforms came mainly about as a technical necessity subsequent to the Civil Service Reforms in the early 1990s. They are part of a wide reform package that involves the introduction of multi-party democracy, privatization and civil service reforms. The CCM Government introduced these reforms out of a realization of the need to liberalize the economy and policy, but CCM has maintained a high level of control of the process, and it continues to have a very clear political majority in both Parliament as well as in LGs. CCM has no interest in radical changes, but pursues incremental reforms that are perceived as leading to improvements in service delivery and are non-threatening to its hold of power. In Kenya, the current debate on the Constitution is part of a wider regime transition after some 40 years of rule by KANU. While in opposition, the current government argued for wide ranging reforms; the future will show whether these will be implemented (e.g. the proposed weakening of the office of the President). It is too early to say what the final outcome on the Constitutional debate will be, but, while some reform based on devolution of powers to LGs seems unavoidable, it is far from certain that its implementation will be radical. The new government has, for instance, not curtailed MPs' involvement in service delivery, and it may find it hard to transfer substantial powers to LGs, a substantial number of which will be controlled by the national opposition. In summary, the decentralization reforms in Uganda took place in a very particular political context that is absent in the other countries ­ thus one may not expect similar radical transitions ­ only to the extent that decentralization reforms leads to significant gains in service delivery and/or political gains for incumbent governments.53 7.2 CONSIDERATIONS ON FUTURE REFORMS AND SUPPORT Although there are common problems, challenges, and significant room for cross- country learning using the experiences from the recent decade of comprehensive 53See also Therkildsen, Ole: Efficiency, accountability and implementation: public sector reforms in East and Southern Africa, UNRISD 2001, Uganda's Radical Decentralisation Reforms ­ an example for others? Working Paper, Centre for Development Research Copenhagen 2002. 86 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda decentralisation reforms, it is nevertheless imperative that programmes are country- specific. However, as outlined at the end of this section, there are several common issues that warrant further analysis in all three countries of continued cross-country relevance. 7.2.1 Kenya The shape of the future LG system, derived from the Constitutional review, will determine the need for support to the process. However, there are areas where further support will be needed under any circumstance: Legal review The present LG Act is not conducive for the LG service delivery; it contains a vast amount of tight and rigid controls measures, which have shown to impact negatively on the LG incentives and performance. A review of the LG Act is therefore an initiative in need of support54. As part of this, a more clearly delineated assignment of expenditure responsibilities should be established for each layer of government. Policy and administrative structures There is a need for strong capacity building support to improve the interaction between the political and administrative levels of LGs, to clarify the roles of each stakeholder, and to ensure more efficient management structures. There is also a great need for support to organisational restructuring in most LGs, adjusted to the individual needs of each LG (depending on factors such as the type of the LG, its size, political objectives, etc.), especially to ensure a better balance between the professional level and the lower staff categories. The role of MLG in policy making, coordination of the reform process, supervision, compliance inspection, and mentoring of LGs, are critical, and support to capacity building within the key functions of the Ministry is important. The capacity of the ministry needs to be enhanced in order to enable it to support and coordinate the capacity building of the LGs. Similarly the importance of a well-functioning and sufficiently capacitated MoF cannot be exaggerated as this is critical for the stability and predictability of the funding flow to LGs, and as an important partner in all LG fiscal reforms. LG finance There is a need to follow-up on the initial reforms of the LG tax assignments, reviewing and reforming the existing taxes, especially the property tax, and exploring structural reforms for assignment of taxes. Even within the existing tax system there seems to be a significant potential for improved revenue mobilisation, provided that LGs' incentives and capacity in taxation are strengthened (cf. below). Hence, it is recommended to support a broad range of activities within this area, in addition to the existing ongoing reforms, including reviews of the legal framework, of the LG tax potential, of the LG practices, and support to dissemination of knowledge and good practises on tax administration. Finally, genuine capacity building of the people involved in LG tax mobilisation is equally important. 54This review has already started but put on hold, until the constitutional review process is completed. However, it may be prudent to continue reviewing some of the general principles and clauses. 87 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda The success of the first years of the LATF deserves full recognition, and support for a further elaboration of the system. Analytical work is required to evaluate the optimal size of the future transfers, the optimal allocation criteria, the menu for investments, the "second" generation of the incentives to promote good governance, and the links to other financial transfers systems directed toward local service delivery, (cf. below). A further refinement of the incentive system within the existing LATF transfer scheme would potentially provide additional benefits in terms of i) LG accountability (involvement of citizens, information to citizens); ii) efficiency in administration (e.g. in budget execution, procurement and reporting); iii) sustainability (if the focus on the links between the transfers and own revenue mobilisation is ensured); and iv) ownership, for instance if various co-funding arrangements or other means are introduced. Experiences from other countries may be interesting for the present situation in Kenya, where the introduced incentive measures have proved to have a positive impact on the LG performance. LG debt recovery and structural reforms are urgently required. As mentioned, the problems vary across LGs, where some of the larger urban areas have structural problems, which cannot be solved without major innovative measures. The incentives built into the LATF system should be further supported, and supplemented with special support programmes, targeting the relatively few most-urban authorities with very serious financial problems. It is important that the introduction of a computerised financial management system ­ (LAIFOMS) ­ is combined with strong support to the basic components of financial management capacity, systems and procedures. The introduction of IT systems without properly working basic budgeting, accounting systems and financial management practices and basic skills, is a risky business. There is, therefore, a very great need for support to general capacity building on financial management through a combination of training, on-the-job assistance, awareness raising and strong institutional incentives within the transfer system(s) to improve on the performance. This should ideally be combined with LG civil service reforms, (cf. below). Human resource development The draft Constitution proposes a radical transfer of powers and functions to LGs. One of the areas that is described in the vaguest manner is the anticipated system for decentralised management of staff. Only if each LG is given full autonomy to hire and fire its staff can one expect true local accountability and the establishment of staffing structures that are tailored to local needs. However, it is clear that such a system cannot be introduced overnight and there will be the need for some central coordination and oversight. The establishment of a Local Government Service Commission has been discussed, and it may provide a required regulatory role. This may also ensure that remote and poorer LGs receive competent staff. However, most often such service commissions tend to centralise recruitment and transfers rather than guide, and great care should be taken in determining the right balance between a gradual introduction of LG autonomy and central oversight and support. 88 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda As part of a revamping of the LG system, it will be necessary to introduce a more comprehensive strategy for capacity building of LG staff and councillors. The MLG is in the process of developing such a strategy. This needs to address issues such as: i) long term funding of training and other capacity building rather than relying on ad- hoc project funding; ii) support and guidance to national training agencies to ensure that they provide relevant and cost effective solutions to LG training needs; iii) modalities for demand-driven capacity building; iv) mechanisms for quality assurance and training evaluations. Institutional framework and coordination of the reforms There is a strong need for support to the institutional framework surrounding the LG reform process. Especially, there is a need within the area of fiscal reforms to establish a forum where inter-governmental fiscal relations can be studied and discussed, where information can be exchanged, disseminated, and compromises reached. This may take the form of a permanent Local Authority Fiscal Commission, although other options are worth considering. The development of strategies and operational plans for the future reform process should be among the first steps in the reform programmes, which needs to be followed-up by annual joint reviews of the progress. This would ideally be coordinated by MLG, but with involvement of a broad range of stakeholders. Finally, a better coordination between civil service reforms (pay reform, overall wage bill, LA restructuring) and decentralisation should be pursued. Further need for support if the plans under the Constitutional Bill are adopted If and when the fundamental changes that are envisaged in new Constitution (within the LG structure, functions, institutional set-up and finance system) are launched, there will be an even a greater need to strengthen the support within all the above- mentioned areas. Some additional tasks will be critical: i) Clear delineation of the functions of each level and type of government/authority. ii) Calculation of the costs of the new functions and identification of the source of funding. iii) Review of the future assignment of revenues to LGs, in terms of taxes, fees and charges, new transfers systems etc. A part of this will be the review of the best funding mechanism for the new sector functions (e.g. conditional, unconditional grants, user fees, etc.). The best funding scheme will typically vary, depending on the type of service to be delivered. iv) Harmonisation of the staff regulations, culture, working procedures for the staff from the various districts, which are amalgamated into new LGs. v) Additional training and capacity building of politicians and staff in the new functions, both within generic areas like management and leadership, but also in managing new sectors like education and health. vi) Changes of the system of monitoring, reporting and auditing, adjusted to the new structures and functions. Common strategies and interlinked programmes supported by the "common-minded" donors should be developed, steered by the GoK/spearheaded by MLG and the 89 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda KLGRP, to ensure coverage of the reform needs in both the rural as well as in the urban authorities (which are quite different). Donor-supported programmes do not necessarily need to be joint programmes. However, they should be developed with an eye on possible mutually strengthening components and synergies. Concluding comments on Kenya The LG reform process in Kenya, after some decades of deteriorated LG performance, development of poor practices of financial management, human resource management, non-coordinated and inefficient service provision, has witnessed rather encouraging results over the last 4-5 years. Many structural problems are still to be addressed, but the initiatives, that have been taken under the KLGRP, have generally moved the process in the right direction. However, a number of initiatives are required and urgent, in any case. These are especially related to the need to get away from a "micro-control" system of centralised checking and interference in local matters, to a system of interaction through the establishment of general conducive legal procedures, based on general guidelines and incentives to adhere to core national targets such as the development of services, good governance and reduction in the fiduciary risks and monitoring and support. Although there are certain real risks that delays in the overall reforms may impact negatively on the LG reform process, the present environment for support is generally judged to be encouraging. 7.2.2 Tanzania With the present reforms within the legal, administrative, financial and human resource areas, especially with the coming LGSP, the emphasis should be on consolidation, deepening and broadening the support to the existing reform process, rather than on the elaboration of new programmes and support schemes. The public service reforms and the question of LG staff autonomy are closely linked to the reforms of the recurrent grant system, where the full benefit of these initiatives will only be achieved if the LGs are given more freedom to adjust the staff to the local needs, to hire, fire and discipline their staff, thereby enhancing accountability. The following areas may deserve increased attention and support: Political and administrative structures Exploration of the optimal size and structure of the system of LGs and the assignment of tasks to each layer of LGs. This would also include a more precise definition of the roles and responsibilities of Regional Secretariats that are supposed to monitor and support LGs ­ yet they remain rather dysfunctional. Finance The ongoing reforms of the recurrent and development grant schemes are promising and warrant full future support. There is also a need for improvements in the accounting systems, registration procedures and the practices for devolution of development investments to LGs. This is partly a purely technical reform with respect to improved systems, and partly a reform that requires better interaction with the funding agencies on streamlining and harmonising the modalities for transfers (planning, budgeting, reporting, registration, etc.) 90 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Further development of the financial incentive system within the inter-governmental fiscal transfer system is another area of critical importance ­ both within the recurrent grant system (by adjusting the autonomy and, maybe, in the medium term the size of the grants against the LG performance) and within the development grant system (by bringing gradually more and more area-based programmes and sector development grants under the coming LGSP transfer window and modalities). There is a need to review the future tax assignments and possible revenue enhancement strategies, with the aim to ensure: i) better links between the service (benefits) and the tax payment (the costs); ii) a rational and simple system easy to administer; iii) improved equity in the tax payment, with a stronger focus on a few high yielding taxes, citizens' rights and obligations and iii) better LG capacity to handle these functions. Finally, support is warrant for improvements in the LG financial statistical data and data processing, in both PO-RALG and the Ministry of Finance. This should improve the basis for informed decision-making within this area. Human resources Clarification of the legal framework for the management of LG staff is absolutely essential for progress on the decentralisation reform, and much impinges on decisions regarding possible revision of the current Public Service Act. There is a special need to support the development of a capacity building framework, including a coordinative forum, policy and strategy formulation, quality assurance unit, standard training materials, etc. Institutional arrangements It was from the onset of the LGRP in 1999 decided to have a semi-autonomous section with well-facilitated staff to drive the reform process. The LGRP has over time grown in terms of numbers of staff; and although it in principle has been decided to mainstream the LGRP into the PO-RALG's current structures, it remains a substantial task to do so as many of the current activities defined as "reform issues" will need to be undertaken by PO-RALG in a more routine manner in future. Some activities under the responsibility of the LGRP Good Governance Outcome Manager have now been mainstreamed and a consultancy will be launched later in 2004 to develop a strategy for the mainstreaming process. A need has also been identified to improve the linkages between the LG reforms and the overall public sector reforms within areas such as pay reforms, modalities for accountability of the civil servants, restructuring and retrenchments, pension reforms, etc, possibly by way of establishing some form of Presidential Oversight Committee or similar arrangement. Concluding comments on Tanzania The overall evaluation is that that the decentralisation process in Tanzania has made important achievements. However, if the overall balance between decentralisation of tasks and responsibilities, finance, personnel and capacity is not ensured, the reform process will face serious challenges. Particularly, the development of the framework for management of LG staff in a manner that ensures accountability between 91 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda politicians and local staff, will be crucial for a successful deepening of the process. Furthermore, it is important to keep up the momentum, recently obtained within the systemic fiscal reforms, and further to develop the discretionary funding available for LGs, linked to improved performance, accountability (downwards, horizontal and upwards) and incentives for the LGs to improve on their performance. 7.2.3 Uganda The present activity level under the various technical reforms is relatively high, in the view of the capacity of the stakeholders to manage and absorb the changes. Considering the great number of parallel ongoing and planned reforms related to decentralisation, it suggested that there should be a focus on supporting the ongoing initiatives rather than on attempting to elaborate radical new procedures and systems. Hence, the main focus should be on strengthening the existing emerging improvements of the coordination and institutional arrangements ­ and on the implementation of the FDS initiatives. Institutional arrangements and coordination It is important to consolidate and ensure implementation capacity in further mainstreaming of the various pilot schemes into general government procedures and operations and long lasting solutions. Support should be rendered through the existing institutional arrangements and the newly established coordinative fora for managing part of the LG reform process. A better coordination of sector reforms and decentralisation should be supported, especially through the coming joint annual review of decentralisation and follow-up on the implementation of the elaborated recommendations. This is closely related to the need to start the preparations for the support after completion of the LGDP II ­ the so-called Decentralisation PRSC55 (change from programme to a kind of budget support for decentralisation) ­ and the need to explore the best possible support tools, based on a careful review of the experiences from the ongoing activities, especially the FDS and LGDP-II. Finally, a better coordination between civil service reforms (pay reform, overall wage bill, LG restructuring) and decentralisation should be pursued. Fiscal decentralisation strategy (FDS) The FDS is at a critical stage, as it will be rolled out to all HLGs from FY 2004/05, next year also covering the about 900 LLGs. The reform anticipates great changes in the budgeting and accounting procedures, in the links between LG politicians and staff, in the links between CG and LG, etc. There is a great need to support this process, both in terms of TA support to the key CG institutions involved in the reform (MoLG, MoFPED, LGFC, LGBC, LGROC, LRECC) and to the LGs in the use of new planning, budgeting, accounting and reporting procedures. Further work on the incentive system, linking LG performance and the level of autonomy, is key for the success of these reforms and warrant strong support. 55This concept came up prior to and during the preparations of the second LGDP. The D-PRSC is supposed to provide a generic window for support to funding of LG service and infrastructure mandates, using some kind of budget support modality to be further elaborated. 92 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda Finally, it is paramount that the FDS is providing a real increase in the LG autonomy within the intergovernmental fiscal transfer system, i.e. lead to a real reduction in the number of grants and restricting budget lines56, and gradually move in a direction of increasing the discretionary non-sectoral development grants by means of merging of the sector development grants with the LGDP non-sectoral grants and mainstreaming of the remaining district support programmes, which have capital investment components. Structures and human resources It is recommended that care should be taken with radical changes, such as the introduction of a regional tier and the centralisation of CAOs prior to further analysis of the existing structures and functioning of the administrative system. The experience with devolution of staff to LGs in Uganda is unique, and it deserves a substantial review of the experiences to date. Concluding comments on Uganda Although the political and economic context of the reforms in Uganda were highly specific, then the Ugandan experience nevertheless indicates how devolution can contribute significantly to overall improvements in governance and service delivery. Most countries in Africa resist, for example, transfer of responsibilities for hiring and firing to LGs ­ yet this appears (though not fully documented) to have worked rather well in Uganda. Many specific areas of the reform nevertheless warrant substantial further analysis ­ and, probably, time to prove their full impact. The cost effectiveness of many of the measures, the quality of services under devolved management, the procedures for decentralised HR resource management, are among the factors that need further analysis ­ especially prior to any changes of these. The strategy elaborated within the area of fiscal decentralisation provides important lessons, in terms of means to address the problems with the hitherto gradual development of very complex technical systems and modalities for the transfer of funds in a situation with a sudden strong increase of intergovernmental fiscal transfers. There are lessons, too, related to in-built challenges for the achievement of the overall policy and objectives, in the interfaces between the sectors, and the decentralisation objectives ­ lessons for ways and means to elaborate and adjust strategies and tools for improvement of the reforms, based on compromises between all involved stakeholders. 7.2.4 Emerging Issues and Areas for Further Cross Country Studies Below are identified a few areas that are critical for the functioning of the systems of LGs, which are of cross-country relevance, and which need further research, documentation and considerations. First, there is a further need to search for the optimal size of the LGs, many reforms have been initiated around the world without proper analytical work on: 56A "budget line" is a requirement within a conditional grant to spend a certain amount/percentage on a specific area, e.g. drugs within the health grant. 93 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda · The fact that various types of task require different scales/sizes of the government units; · The inter-linkages between the various tasks ­ some tasks are interrelated and need to be conducted at the same layer of government; · The funding capacity and economies of scale ­ options for viable units; · The possibilities for citizens´ participation; · Governance issues such as opportunities for active political representation. · Historical and cultural links. These studies would be highly relevant for the reform process in all three countries in order to inform the design of the overall structure of LGs, and the assignment of specific expenditure assignments across each layer of government. Second, work is required on revenue assignments ­ how to improve on the LG revenue mobilisation, reduce the increasing vertical imbalances, and improve the sustainability of the system of LG, considering the wish to avoid distortion of the local economy, and keeping attention to the overall poverty alleviation objectives. The challenge is to improve LG tax yield in a poverty sensible manner. Many studies of LG taxes tend either to argue in favour of abolishment of most local taxes, centralising taxes, due to their distortive impact on poor people and local business ­ or they argue strongly for purely technical improvements of the capacity of the LGs to increase existing LG taxes (for instance by focusing on privatisation of tax collection) without considering the impact of this on the overall development objectives. However the challenge is to explore and identify systems, tax assignments, revenue sharing models and tax collection procedures that can address both objectives simultaneously. Third, related to the above, there is a need to design intelligent transfer systems, with fair, simple, transparent allocation criteria, which promote the incentives for proper performance within areas of importance for service delivery, good governance and sustainability, without leading to micro-control and too much interference from the CG in the local priorities. There is also still a need to streamline and harmonise various approaches to transfers of development funds from central to LGs. Fourth balancing between pursuit of decentralisation by devolution and other strategies for improvements of the public service, including the establishment of user groups and other more direct involvements of citizens in management of services. Fifth, further studies are needed on the functioning and possible development of LG personnel management systems that can achieve the right balance between LG autonomy over LG staff and, accountability, while avoiding undue political influence, bloated administrations and ensure satisfactory professional career paths etc. Finally, the country studies have shown a great need for further elaboration of tools, procedures and practices for strengthening of the downwards accountability. 94 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda 8. LIST OF LITERATURE · Bird, Richard Robert D. Ebel and Christine L Wallice (ed) in "Decentrali- sation of the Socialist State, Intergovernmental Finance in Transition Economies, IBRD, 1995 · Crook R, and A. Sverrison : "Decentralisation and poverty alleviation in developing countries ­ a comparative analysis or is West Bengal Unique?", IDS Working paper No. 129 May 2001 · Hyden, G. 1983: No Shortcuts to Progress - African Development Management in Perspective, London, Heinemann. · Kasfir N 1976: The Shrinking Political Arena ­ Participation and Ethnicity in African Politics, with a case study of Uganda. University of California Press · Kithinji Karagu and Rwekeza Mukandala 2004: Public Service Pay Policies, PwC. · Mawhood, Philip 1983: Local Government in the Third World ­ The Experiences of Tropical Africa. · OECD/DAC: "Lessons Learned on Donor Support to Decentralisation and Local Governance", by Schou Arild and Steffensen, Jesper, OECD, DAC Network on Development Evaluation, 2004 · Paul Smoke in: "Decentralisation and Local Governance in Africa", in Public Administration and Development, volume 23, Number 1, UNCDF, February 2003 · Steffensen, Jesper and Trollegaard, Svend: "Fiscal Decentralisation and Sub- National Government Finance in Relation to Infrastructure and Service Provision ­ Synthesis Report of 6 Sub-Saharan African Country Studies", NALAD, May 2000, World Bank/Danida/USAID · Steffensen, Tidemand, Sswankambo, Country Study Uganda ­ A Comparative Analysis of Decentralisation in Kenya, Tanzania and Uganda, August 2004 · Steffensen, Tidemand, Naitore, Country Study Kenya ­ A Comparative Analysis of Decentralisation in Kenya, Tanzania and Uganda, August 2004 · Steffensen, Tidemand, Sswankambo, Country Study Tanzania ­ A Comparative Analysis of Decentralisation in Kenya, Tanzania and Uganda, August 2004 · Therkildsen, Ole 2001: Efficiency, accountability and implementation: public sector reforms in East and Southern Africa, UNRISD. · Therkildsen, Ole 2002: Uganda's Radical Decentralisation Reforms ­ an example for others? Working Paper, Centre for Development Research Copenhagen. · The World Bank, The Asian Development Bank, DFID: "Devolution in Pakistan - An Assessment and Recommendations for Action" May 2004 · The World Bank: World Development Indicators, Database 2004 · United Nations Capital Development Fund: "Taking Risks ­ Background Papers", September 1999, UNCDF · UNCDF, by Shotton, Roger (ed) Local Government Initiative: Pro-poor Infrastructure and Service Delivery in Rural Asia- a Synthesis of Case Studies, UNCDF, 2003 95 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda 9. ANNEXES 9.1 LIST OF STAKEHOLDERS CONSULTED · ALAT/Tanzania · Association of Local Government Authorities in Kenya (ALGAK) · Bagamoyo District Council/Tanzania · Bushenyi District/Uganda · DFID/Kenya · DFID/Uganda · Dodoma District Council/Tanzania · Dodoma Zonal Reform Team · European Union/Kenya · Ilala Municipal Council/Tanzania · Ireland Aid/Uganda · Kampala City Council/Uganda · Kanungu District · Kenya Local Government Reform Programme under the Ministry of Local Government · Kirinyaga County Council/Kenya · Kisarawe District Council/Tanzania · Layers/Kenya · Local Government Finance Commission/Uganda · Ministry of Education/Tanzania · Ministry of Finance and Planning/Kenya · Ministry of Finance, Planning and Economic Development/Uganda · Ministry of Finance/Tanzania · Ministry of Health/Kenya · Ministry of Health/Uganda · Ministry of Local Government, Project Coordination Unit/Uganda · Ministry of Local Government/Kenya · Ministry of Local Government/Uganda · Ministry of Public Service/Uganda · Ministry of Water and Livestock Development/Tanzania · Ministry of Water/Uganda · Morogoro Municipal Council/Tanzania · Mukono District/Uganda · Mzumbe University/Tanzania · Nakisungu Sub-County/Uganda · Nyeri Municipal Council/Kenya · Office of the Auditor General/Uganda · PO- RALG /Tanzania · PO-PSM/Tanzania · PO-PSM3/Tanzania · PO-RALG/LGRP/Tanzania 96 Final Synthesis Report ­ A Comparative Analysis of the Experiences of Decentralisation in Kenya, Tanzania and Uganda · PROLoGs/Tanzania · Public Service Commission/Uganda · Royal Netherlands Embassy/Uganda · The Road Management Board ­ Secretariat/Kenya · The Royal Danish Embassy in Tanzania · The Royal Danish Embassy in Kenya · The Royal Danish Embassy in Uganda · The World Bank/Kenya · The World Bank/Tanzania · The World Bank/Uganda · Tororo District/Uganda · Tororo Municipality/Uganda · ULAA/UAAU/Uganda · UNCDF/Kampala 97