72297 v2 DEVOLUTION WITHOUT DISRUPTION Pathways to a successful new Kenya DEVOLUTION WITHOUT DISRUPTION Pathways to a successful new Kenya November 2012 TABLE OF CONTENTS Acronyms and abbreviations i Foreword iii Acknowledgments iv Executive summary v Part I: Seeing the future from the perspective of the past Chapter 1: Kenya’s devolution in context 4 A turning point in a long historical process 4 Deep-rooted political economy dynamics 6 Chapter 2: The scale, scope and complexity of Kenya’s devolution 12 How the public sector functions today, and where it has failed 12 How the public sector will function after devolution 13 Implications of the transformation for public administration and service delivery 14 Implications for governance and the role of Parliament 17 Chapter 3: Demographic and geographic diversity of Kenya’s counties 20 Midgets and giants: Wide variations in population, density and urbanization 20 Poverty disparities, service inequities and social outcomes 21 Dealing with patterns of marginalization 24 Part II: Equity between levels of Government Chapter 4: Why it all starts with function assignment 30 Designing a framework for function assignment across levels of government 32 Clarifying function assignments 36 Guiding the function assignment process successfully 40 Refining function assignment going forward 44 Chapter 5: Determining how much counties need 46 Determining county needs 46 Simulating aggregate county expenditure needs 47 Converting the simulation into a real estimate of aggregate county needs 53 Finding the fiscal space to move beyond historical spending 55 Implications for design of intergovernmental financing arrangements 56 Chapter 6: Solving the intergovernmental division of revenue jigsaw 58 How will counties be resourced in practice? 58 Estimating the 15 percent? 63 Part III: Equity across Kenya’s diverse counties Chapter 7: Funding from within: County own revenue powers and capacity 68 Legal framework for own revenue powers of county governments 68 Estimating capacity from current local authority collections 73 Getting the legal framework right for own revenues 76 Addressing challenges in existing local own revenues 78 County fees and charges 80 The case for additional county revenue sources 81 Design of subnational tax regime 83 County tax collection 85 Chapter 8: Allocating the share capital equitably but carefully 90 A powerful constitutional mandate: Equalization across Kenya’s forty-seven counties 91 The CRA proposal: An important step toward redistribution 92 Managing the transition: Equalizing over time to avoid fiscal stress 94 Chapter 9: Strings attached? The scope for conditional transfers 106 The rationale for conditional transfers 106 Supporting national priorities at the local level 109 Filling gaps in the transition 113 Stimulating performance at the local level 114 Chapter 10: Making the most of the Equalization Fund 120 How much does the Fund amount to? 120 What objectives should guide its allocation? 121 What will the Fund finance and where? 122 Though small, the Equalization Fund can still be effective 123 Part IV: Translating the vision, with a smooth transition Chapter 11: Promoting intergovernmental coordination 128 Why intergovernmental relations is important in Kenya 129 Structure of intergovernmental relations in Kenya 131 Models for sector coordination 133 Challenges of parallel systems 135 Chapter 12: Managing money for county development 140 Kenya’s new budget process 140 Relationship between national and county governments 147 Budgeting and spending at the county level 151 Transitioning to the new systems 156 Helping counties manage finances accountably 157 Chapter 13: Social accountability: Transparency, accountability and participation 162 The basic elements of a social accountability system 163 Learning from past experiences with devolved funds 166 Supporting institutions for a functioning social accountability system 168 Chapter 14: Financing and management of urban areas 172 Urban development is good for economic growth 172 Urban management under devolution 173 Comparing Kenya with international experience 175 Gaps in the Urban Areas and Cities Act 176 Chapter 15: Decentralizing public services 186 Existing administrative arrangements 186 Legal framework for managing county public services 187 Balancing central control and local autonomy 190 A robust framework for managing decentralized staff 195 Chapter 16: Managing a complex transition 200 Putting Kenya’s devolution in context 200 Institutional arrangements for managing transition 202 Implementing transfer of functions 205 Transfer of public servants 208 Establishing Public Financial Management systems 211 Transitioning from Local Authorities to County Governments 212 Engaging where it matters 213 Main conclusions and recommendation 216 ANNEX Annex 1: Overview of FDKP 228 FIGURES Figure 1: Kenya’s devolution presents massive challenges for political and administrative restructuring vi Figure 2: Flows of revenues from different sources for county governments viii Figure 3: Vertical imbalance: Applying the CRA formula, only two counties would receive ix less than half of their resources from transfers Figure 4: Per head equitable share allocations to counties (assuming 15 percent equitable share) xiii Figure 5: A massive reallocation of funds across counties under a fiscally prudent scenario xiii Figure 6: Counties where Local Authorities ran budget surpluses in 2008/09 xvi Figure 7: Elements of social accountability systems xvii Figure 8: Despite rapid urbanization, most counties are still predominantly rural xx Figure 1-1: Number of bills considered by parliament per year 4 Figure 1-2: Various government systems at the local level in Kenya 8 Figure 2-1: Kenya’s devolution presents massive challenges for political and administrative restructuring 14 Figure 2-2: How Kakamega will integrate former districts and Local Authorities in a new 15 county administration Figure 2-3: In the heath sector, 9 out of 10 personnel will be moved to county governments, 16 even though they are likely to remain in the same pay station Figure 3-1: There are significant inequities within counties 22 Figure 3-2: Poverty is as dispersed between counties as population density 22 Figure 3-3: There is a strong correlation between poverty and social outcomes, particularly education 22 Figure 3-5: Kenya’s spatially unbalanced growth is due to its geographic diversity and inequity 24 Figure 3-6: The uneven pattern of Kenya’s infrastructure development 26 Figure 4-1: Unbundling ensures there is clarity about who funds what 36 Figure 5-1: Except for a spike in 2009/10, development spending has seen relatively constant 54 growth from 2005/6 to 2010/11 Figure 6-1: Flows of revenues from different sources for county governments 60 Figure 6-2: Four basic options for financing counties 60 Figure 6-3: The chosen rate will apply to a lagged base 63 Figure 7-1: Water and sewerage fees, once major Local Authority revenues, have fallen dramatically 72 Figure 7-2: Kenya will become increasingly reliant on intergovernmental transfers to finance 74 subnational spending under devolution Figure 7-3: Subnational resources as a percentage of total general government spending 74 - international comparison Figure 7-4: Subnational resources as a percentage of GDP - international comparison 74 Figure 7-5: 28 of the forty-seven counties relied on transfers for over half of their resources in 2009/10 75 Figure 7-6: Vertical imbalance - applying the CRA formula to 2009/10 data, only Nairobi 75 would receive less than half of its resources from transfers Figure 7-7: How long does it take businesses to get things done? 79 Figure 8-1: Total equitable share allocations to counties 95 Figure 8-2: Per head equitable share allocations to counties 96 Figure 8-3: Total county revenues including own sources 97 Figure 8-4: A massive reallocation of funds across counties under a fiscally prudent scenario 99 Figure 8-5: Almost all counties “winners� under the CRA proposal but a major fiscal 100 burden on the center Figure 8-6: Winners and losers of redistribution in the health sector 101 Figure 8-7: Some counties will receive major new funding 102 Figure 9-1: Instruments to support local service delivery 106 Figure 9-2: The main dimensions of performance-based grant design 115 Figure 10-1: How much does the Equalization Fund amount to (KES billion)? 120 Figure 10-2: Counties covered by the Ministry of Northern Kenya 124 Figure 12-1: Increase in parliamentary independence and power 141 Figure 12-2: National budget process under the PFM Act 144 Figure 12-3: Selected Local Authority budget surpluses by county as a % of total revenue, 2009/10 155 Figure 13-1: Elements of social accountability systems 163 Figure 13-2: Accountability framework for service delivery 165 Figure 13-3: Perceptions of Local Authorities in terms of participation and transparency 166 Figure 13-4: An example of a toolbox for county governments 168 Figure 14-1: Kenya’s demographic transformation 172 Figure 14-2: Richer countries are more urbanized 173 Figure 14-3: Despite rapid urbanization, most counties are still predominantly rural 181 TABLES Table 1: A phased approach to transferring functions to counties viii Table 2: Components’ source, weights and objectives in CRA’s formula xii Table 3: Twenty-one urban centres with more than 80,000 residents will not have municipal boards xviii Table 3-1: Illustrating potential differences in policy objectives and strategies among the counties 23 Table 4-1: Broad county, national and unassigned functions 32 Table 4-2: Potential function assignment criteria for the Transition Authority to consider 37 Table 4-3: A phased approach to transferring functions to counties 43 Table 5-1: Devolving district administration – three scenarios 48 Table 5-2: Funding for devolved functions under the three scenarios using simulation year of 2010/11 50 Table 7-1: Legal basis for Local Authorities’ revenue collection 69 Table 7-2: Main revenue sources for Local Authorities, 2009/10 70 Table 7-3: Comparative under collection of property taxes in Kenya, 2009/10 71 Table 7-4: Composition of ‘other’ revenues, 2006/07 71 Table 7-5: Off budget estimates of user fees collected at public health institutions (KES billions) 73 Table 7-6: Classification of subnational taxes by degree of central vs. local control 84 Table 8-1: Components’ source, weights and objectives in CRA’s formula 95 Table 8-2: How the Constitution assigns health functions 99 Table 8-3: Risk mitigation strategies for "cash poor" and "cash rich" counties during the transition 101 Table 8-4: Counties that will receive the biggest increases have least capacity to manage funds 102 Table 10-1: How much will actually be allocated to the Equalization Fund? 121 Table 10-2: Northern Kenya’s main policy challenges and counties affected 123 Table 11-1: Selection of concurrent areas of service delivery responsibility 130 Table 11-2: Departmental heads represented in a typical district 136 Table 12-1: Principal PFM actors at national and county level 141 Table 12-2: Steps in the county budget process 152 Table 14-1: Comparison of cities, municipalities and towns under the Urban Areas and Cities Act 176 Table 14-2: Average size of African local governments 177 Table 14-3: Twenty-one urban centres with more than 80,000 residents will not have municipal boards 179 Table 14-4: Shares of local revenues collected from rural and urban areas, 2008/09 182 Table 16-1: Authorized positions identified as attached to devolved functions, 2012/13 budget 208 BOXES Box 1: What Kenyans hope devolved government will do for them xxvii Box 4-1: Recentralization: The case of Uganda 31 Box 4-2: What the Constitution says about function assignment 32 Box 4-3: Even within concurrent subject areas, not all functions have been assigned 33 Box 4-4: Framework for function assignment: Proposition by the Task Force on Devolved Government 39 Box 4-5: Possible components of ministry implementation plans 43 Box 5-2: How the big ticket items were allocated differently in the three scenarios 49 Box 5-3: Reference notes for table 5-2 52 Box 6-1: Why the equitable share will be unconditional 59 Box 6-2: Existing earmarked programs that would become conditional grants if continued 61 Box 6-3: Reasons why conditional grants are desirable in certain circumstances 62 Box 6-4: Calculating the basis for equitable sharing: 15 percent of what? 64 Box 7-1: Recommendations of the ask Force on Devolved Government on property rates reform 78 Box 7-2: Regulatory environment for businesses―has the Single Business Permit worked? 80 Box 7-3: Tax systems and state building 82 Box 7-4: Indonesia’s experience with “open� and “closed� lists of subnational revenues 85 Box 7-5: Regional experience in outsourcing local revenue collection 87 Box 8-1: A note on defining “equilization� 91 Box 8-2: The fiscal gap model 93 Box 8-3: The many complexities of estimating fiscal gaps 93 Box 8-4: Example of a needs based approach 94 Box 8-5: Is the 15% a minimum even during transition? 103 Box 8-6: How Papua New Guinea gradually introduced equalization 104 Box 9-1: Typology of conditional grants and their impact 107 Box 9-2: A strategy for phasing out conditionality 108 Box 9-3: Existing funding arrangements that would be conditional if continued 109 Box 9-4: A simple model for allocating capital grants in developing countries 111 Box 9-5: Five design issues for earmarked urban grants 112 Box 9-6: Other types of conditionality 113 Box 9-8: Pre-requisites for PBGS and recommendations for implementation 117 Box 10-1: Who are the marginalized communities? 122 Box 11-1: Kenya’s intergovernmental relations architecture 132 Box 11-2: Intergovernmental relations in Australia 134 Box 12-1: International solutions to the challenge of strong parliamentary powers to amend 145 the budget: Sweden and South Africa Box 12-2: What the Constitution says about oversight of county Public Financial Management 148 Box 12-3: Major points of difference resolved through the “harmonised� PFM Act 150 Box 12-4: Predictability of intergovernmental transfers: International experience 157 Box 12-5: Success factors for subnational performance monitoring systems 159 Box 12-6: Decentralisation and local capacity: Ugandan experience 160 Box 13-1: Tracking expenditures in Uganda: Promoting efficiency 164 Box 14-1: Constitutional provisions on urban government 174 Box 14-2: Urban functions assigned to county governments under Kenya’s Constitution 175 Box 14-3: Urban Areas and Cities Act provisions on powers and functions of city and municipal boards 180 Box 15-1: New institutions and new roles 188 Box 15-2: Indonesia: Worsening corruption in Human Resource Management after decentralization 190 Box 15-3: Key gaps in the regulatory framework for devolved civil service management 192 Box 15-4: Potential problems with concentration of power in the hands of county Public Service Boards 193 Box 15-5: Regulating casual or non-civil service employment 195 Box 15-6: Five reasons for retaining some national involvement in local civil service management 196 Box 15-7: National wage policy can compromise subnational governments 197 Box 16-1: Transitioning to Indonesia’s ‘big bang’ decentralization 201 Box 16-2: Two risks of poorly managed transition 203 Box 16-3: Institutional arrangements for transition 204 Box 16-4: Lessons from international experience of managing transition 205 Box 16-5: Four ways to ensure plans get implemented 205 Box 16-6: Legal architecture for asymmetric transfer of functions 206 Box 16-7: South Africa’s local government transition guidelines 215 Box 16-8: Checklist for establishment of county systems 216 ACRONYMS AND ABBREVIATIONS AG Attorney General FFC Financial and Fiscal Commission AIA Appropriations in Aid (South Africa) AIE Authority to Incur Expenditure FIAS Foreign Investment Advisory ALGAK Association of Local Government Services (now Facility for Investment Authorities of Kenya Climate Advisory Services) ANC Africa National Congress FY Financial Year ASALs Arid and Semi-Arid Lands GDP Gross Domestic Product BPS Budget Policy Statement GIZ Deutsche Gesellschaft für CARB County Allocation of Revenue Bill Internationale Zusammenarbeit, CBEF County Budget and Economic Forum GmbH (German: German Society for CBROP County Budget Review and International Cooperation, Ltd.) Outlook Paper GJLOS Governance, Justice, Law and Order CDD Community Driven Development Sector CDF Constituency Development Fund GoK Government of Kenya CEMF County Executive Member for GST Goods and Services Tax (Australia) Finance HIV/AIDS Human Immunodeficiency Virus/ CFSP County Fiscal Strategy Paper Acquired Immune Deficiency CGA County Governments Act Syndrome CGFM County Government Financial HR Human Resources Management Bill HRM Human Resource Management CIC Commission for the Implementation HSSF Health Systems Support Fund of the Constitution IBEC Intergovernmental Budget and CILOR Contributions in Lieu of Rates Economic Council CKRC Constitution of Kenya Review IEBC Independent Electoral and Commission Boundaries Commission CoA Chart of Accounts IFMIS Integrated Financial Management COAG Council for Australian Governments Information System (Australia) IGFT Intergovernmental Fiscal Transfer CRA Commission on Revenue Allocation IGFR Intergovernmental Fiscal Relations CSF Cabinet Secretary for Finance IMF International Monetary Fund CSO Civil Society Organization IRA Intergovernmental Relations Act DC District Commissioner JICA Japan International Cooperation DDO District Development Office(r) Agency DORA Division of Revenue Act KANU Kenya African National Union DP Development Partner KEMRI Kenya Medical Research Institute DRR Disaster Risk Reduction KEMSA Kenya Medical Supplies Agency EAC East African Community KENAO Kenya National Audit Office ESW Economic Sector Work KEPI Kenya Extended Programme on FACT Function Assignment and Immunization Competency Team KERRA Kenya Rural Roads Authority FDKP Fiscal Decentralization Knowledge KES Kenya Shillings Program (of the World Bank, funded KEU Kenya Economic Update by Australian Aid) i ACRONYMS AND ABBREVIATIONS KIHBS Kenya Integrated Household Budget PBO Parliamentary Budget Office Survey PDL Fund Petroleum Development Levy Fund KIPPRA Kenya Institute for Public PDMO Public Debt Management Office Policy Research and Analysis PEFA Public Expenditure and Financial KLGRP Kenya Local Government Reform Accountability Programme PFM Public Financial Management KMTC Kenya Medical Training Centre PFMA Public Finance Management Act KNBS Kenya National Bureau of Statistics PIT Personal Income Tax KRA Kenya Revenue Authority PP Public Participation KRB Kenya Roads Board PREM Poverty Reduction and Economic KURA Kenya Urban Roads Authority Management Unit (of the World LA Local Authority Bank) LAIFOMS Local Authorities Integrated PSASB Public Sector Accounting Standards Financial Operations Management Board System PSC Public Service Commission LASC Local Authority Service Charge PSR Public Sector Reform LASDAP Local Authority Service Delivery RMLF Roads Maintenance Levy Fund Action Plan SBP Single Business Permit LATF Local Authority Transfer Fund SNG Subnational Government LGFC Local Government Finance SPP Specific Performance Payment Commission (Uganda) (Australia) M&E Monitoring and Evaluation SRC Salaries and Remuneration MDG Millennium Development Goals Commission MoF Ministry of Finance TA Transition Authority MoLG Ministry of Local Government TB Tuberculosis MoMS Ministry of Medical Services TFDG Task Force on Devolved Government MoPHS Ministry of Public Health and TSA Treasury Single Account Sanitation TTDG Act Transition to Devolved Government MP Member of Parliament Act MTEF Medium Term Expenditure TWG Technical Working Group Framework UACA Urban Areas and Cities Act NACC National AIDS Control Council UCLGA United Cities and Local NARC National Alliance of Rainbow Governments Association Coalition UK United Kingdom NASCOP National AIDS and STI UNCDF United Nations Capital Development Control Programme Fund OCPD Officer Commanding Police Division USAID United States Agency for OECD Organisation for Economic International Development Co-operation and Development VAT Value Added Tax PBGS Performance-Based Grant System WDR World Development Report ii FOREWORD W hen Kenyans voted decisively in favor of a new Constitution on August 4 2010, we―Kenya’s friends and partners―recognized that the country had reached a historical milestone. Kenya’s new Constitution envisages far-reaching changes to the way that government operates and relates to its citizens, with a view to making it more fair, efficient, transparent, and accountable. Devolution is a core dimension of this ambition, and one that will profoundly affect the daily lives of Kenyans from all walks of life. Following the next elections, Kenyans will increasingly be looking to their new county governments to deliver essential services. Enabling counties to deliver will imply a massive reorganization of state functions around these new units which will, over time, acquire discretion over significant budgets, staff and programs. Since devolution has generated such high hopes among Kenyans, it is imperative that it does not disappoint. Kenya’s devolution is ambitious, even by international standards. Implementing the transition to devolved government, in tandem with many other reforms to the institutions of national government, poses enormous challenges. It is inevitable that things will not always go according to plan, and that implementation will reveal problems which were not anticipated. The breadth of the transformation makes it imperative to consider clearly what these changes will involve, and how best to prepare for them. This is precisely the focus of this report, “Devolution without Disruption: Pathways to a Successful New Kenya�, which we are delighted to present. The report covers the critical issues that Kenya’s policy-makers will need to address, as the country seeks to fulfill the constitutional promise of a more devolved government that is closer and more responsive to the people. It addresses the critical importance of reconciling, in the design of intergovernmental financial arrangements, the twin goals of maintaining fiscal stability and efficiency, while also pursuing equalization across Kenya’s regions. It provides suggestions for updating the framework for Public Financial Management―at both the national and the county levels, while supporting meaningful social accountability from the start. Finally, it discusses critical transition issues, including how urban areas will be managed and what reorganization of the civil service is required to ensure that service delivery not only continues without interruption, but is also improved over time. Implementing devolution will be a Kenyan-led process. By walking this new road, Kenyans will find their own solutions to the development problems that they face. This report does not advocate for or against devolution, and it does not pretend to provide definitive solutions to the problems that exist. Instead, it aims to identify for policy-makers the key issues that devolution brings forward, together with suggestions and options for addressing those issues. It highlights the trade-offs that are inevitably involved in the decision-making process, and draws on international experience, where relevant. The World Bank, the Government of Australia (which funded this work), and the Fiscal Decentralization Knowledge Program team (that prepared this report) look forward to continuing their partnership with the wide range of Kenyan stakeholders who seek to make devolution a success for a better new Kenya. Johannes Zutt H.E. Geoff Tooth World Bank Australian High Commissioner Country Director for Kenya iii ACKNOWLEDGEMENTS T his report was prepared by the World Bank’s Fiscal Decentralization Knowledge Program team. The core team includes Kathy Whimp and Aurélien Kruse, who led the overall report, as well as Christopher Finch, Geoff Handley, Wangari Muikia, Frederick Owegi and Jonathan Rose who each wrote chapters, as well as Anne Khatimba and Lucy Mhina who provided critical editorial support. The report benefited from the invaluable insights of a broader team: Wendy Ayres, Gabriel Demombynes, Jane Kiringai, George Larbi, Christina Murray, Diana Ortiz-Zuluaga, Onur Ozlu, John Randa, Garry Reid, Larry Schroeder, Roger Sullivan and Sue Viney. The team is particularly grateful for the strategic guidance it received from Wolfgang Fengler, Sue Graves, Humberto Lopez and Johnannes Zutt. Peer reviewers Jonas Frank, Stuti Khemani and Paul Smoke provided insightful and detailed feedback. Lastly, but not least, production of this report was made possible through an incredibly fruitful collaboration with Kenyan institutions, namely the Ministry of Finance, Commission on Revenue Allocation, Commission on Implementation of the Constitution, Transition Authority, Ministry of Local Government, and the Association of Local Government Authorities of Kenya, to name but a few. The support of the Australian High Commission and Australian Aid to the World Bank’s Fiscal Decentralization Knowledge Program is also gratefully acknowledged. iv EXECUTIVE SUMMARY 1. Kenya’s new Constitution marks a critical juncture in the nation’s history. It is widely perceived by Kenyans from all walks of life as a new beginning. Indeed, many feel that post-Independence Kenya has been characterized by centralization of political and economic power in the hands of a few, resulting in a spatially uneven and unfair distribution of resources and corresponding inequities in access to social services: the opposite of an inclusive state. Born of the political opportunity created by the 2008 post-election violence, the Constitution that was finally adopted after almost a decade of unsuccessful reform attempts presages far-reaching changes. Its vision encompasses a dramatic transformation of the Kenyan state through new accountable and transparent institutions, inclusive approaches to government, and a firm focus on equitable service delivery for all Kenyans through the newly established county governments. 2. Devolution is at the heart of the new Constitution and a key vehicle for addressing spatial inequities. A more decentralized government makes eminent sense given Kenya’s diversity and past experience with political use of central power. Decentralization has been increasingly seen and adopted worldwide as a guarantee against discretionary use of power by central elites, as well as a way to enhance the efficiency of social service provision, by allowing for a closer match between public policies and the desires and needs of local constituencies. Kenya’s Constitution entrenches devolved government by guaranteeing a minimum unconditional transfer to counties under the new dispensation. 3. Devolved government presents an opportunity to address the diversity of local needs, choices and constraints in Kenya. This is a very diverse country with ten major and more than thirty minor ethnic groups. Needs are very different between the arid and semi-arid North, the highlands, the rural Northern Rift, the urban centers of Mombasa, Nairobi, and Kisumu, the coast, and Western Kenya. Counties will be better placed than the national government to deliver social services, because they have specific challenges and the local knowledge to address them. For instance, in the case of health, lagging counties still need to catch-up in providing basic health services, while the leading urban counties will be faced with new types of diseases (mostly non-communicable such as diabetes and cancer). With these stark differences it makes little sense to provide the same mix of services across the country. And even if there are no dramatic improvements in service delivery, people prefer to make decisions themselves rather than following directions imposed by a central government. With a constitutional guarantee of unconditional transfers from the center, Kenya’s counties will have the means and the autonomy to begin to address local needs, and their citizens will be more able to hold them accountable for their performance. 4. But Kenya’s devolution is very ambitious, and therefore commensurately risky. It is a massive undertaking from a logistical point of view. The day after the next general election, Kenya’s system of government and public administration will be profoundly remodeled. It is inevitable that teething problems will be encountered during the transition. There are diverging views on how far and how fast the transition should be implemented. Since Independence, Kenya’s leaders have held diverging views about devolution. From one perspective, it offers the potential to redress perceived ethnic and political bias by giving local communities far greater control over resources and decisions about service delivery. However, from another perspective, devolution could potentially undermine national unity, v Devolution without disruption – Pathways to a successful new Kenya by encouraging fragmentation of the state along partisan lines or by ‘decentralizing corruption’, leaving citizens worse off if local elites are able to capture resources to the detriment of the majority, or if newly established counties fail to put in place the systems needed for effective and transparent service delivery. 5. In the short run, managing the transition to the new system and people’s expectations will be critical. From a political viewpoint, the devolution process has generated tremendous hope in the population, and sometimes unrealistic expectations of how quickly things can and will change in the ordinary lives of Kenyans. From a logistical viewpoint, this is a highly complex undertaking, which Kenya has embarked upon in a context of political division. The challenge will therefore be to manage expectations of how much and how quickly devolution can deliver, and to make sure that the transition to devolved government causes as little disruption (and litigation) as possible to the delivery of services that are essential for the welfare of the people of Kenya, as well as for the health of its growing economy. Equal distribution of wealth across Kenya may be desirable politically, but it is impossible economically. With a few exceptions, counties will be too small to generate the economies of scale which companies require in order to be successful. Firms will only come to Kenya and expand if they can operate in the whole country and beyond. Moreover, no matter what remote counties do to attract them, most will chose to locate their operations in Kenya’s big cities to benefit from the markets around them. 6. The next two years will be critical because the foundations of the devolution architecture will be impossible to alter the foundations, at least not without knocking it down and starting again. Kenyans sense that a momentous restructuring of their country is underway. There are high expectations and much anxiety. Kenya’s devolution is not only a critical milestone in this country’s history; it is also remarkable in global terms. Many countries―both rich and poor―have transferred power and resources to lower levels of government. Few have done so to entirely new subnational units which they have had to establish from scratch. Kenya will undergo a dual transition: a transfer of power and resources from the center to the subnational level and a simultaneous reorganization of local government, with the consolidation of existing local structures into forty-seven newly-created county governments (see Figure 1). Figure 1: Kenya’s devolution presents massive challenges for political and administrative restructuring. Solid waste management, public health, parking and street lighting, markets, slaughterhouses, waste sewerage, storm water drainage, billboards, noise control, �re �ghting, game parks 20+ De-concentrated Administrations 20+ District Administrations Health, agriculture, livestock, �sheries, planning, housing, lands, transport, Liquor licensing, disaster management, rural electricity, sports and culture, plant and animal quarantine, control of drugs and pornography environment and conservation Source: World Bank (2011b) vi Executive Summary 7. The devolution train has already left the station: the challenge is to make sure it arrives at the right destination, safely and on time. The politics of devolution explain the high intensity of hopes and expectations that have been pinned to it. It also means that there are high risks if expectations are not met. There are great opportunities and enormous challenges waiting for Kenya, in a critical election year, which will determine the fate of the country, politically and economically, for years to come. This report takes a look at the critical issues facing Kenya’s policy-makers today. It does not argue for or against devolution (a decision that belongs solely to Kenyans), but presents suggestions and recommendations on how best to navigate the tough choices ahead. It’s main focus is to help Kenya to manage a delicate transition. Financing county needs: transferring the right amount with an appropriate mix of instruments 8. Decentralizing power requires transferring resources from the center to the local level: but there is no single answer to the question of how much, how fast and in what form. Unlike in many other decentralized countries, Kenya’s counties will not have substantial sources of own revenues. Consequently, they will depend on the national government for transfers. Deciding the amount of each county’s transfers involves a two-step process to ‘fairly’ divide national resources. First, resources must be divided “vertically� between the national and county governments, in such a way that each is adequately resourced to carry out its mandated functions. The vertical division must also take account of historical under-privileging of service delivery. Second, the county share must be split across the forty-seven counties in a way that recognizes their different inherited needs and also addresses historical inequalities between them. This will be particularly difficult in an overall context of fiscal stress, with limited scope to increase overall public funding. Third, national government needs to determine where it is appropriate to provide conditional grants from its own resources. 9. A golden rule of decentralization is “funding follows function�, which is why the function assignment process is so important. While Kenya’s Constitution provides high-level guidance on the respective responsibilities of the national and the county governments, much more granular work is needed in order to provide a basis for sharing resources. The Transition Authority (TA) has been set up with the mandate to carry out a detailed assignment of functions, but the legislated process for function assignment is cumbersome. It is envisaged that each county will apply to the TA for each function to be devolved to it on an ad hoc basis. This approach is likely to be overly complex to manage, and may also lack transparency (because of the resulting complexity), potentially undermining accountability at the local level. In this report, an optional framework is proposed for organizing a phased ‘bulk’ transfer of functions to counties, in a way that minimizes ad hoc arrangements and maximizes efficiency and transparency (Table 1 shows one proposed option). 10. By prescribing a minimum transfer to counties, Kenya’s Constitution has not pre-empted the usefulness of an aggregate costing of county needs. This is essentially for two reasons: first, the overall cost of functions to be devolved is likely to be significantly more than the constitutionally guaranteed share of 15 percent of audited national revenue, and; second, functions will be phased out over time during the transition period, and as a result counties may not initially receive the full amount guaranteed to them. Working out a ‘fair split’ of available resources in a context of limited fiscal space, as well as the most appropriate mix of grant instruments (i.e. between the equitable share and conditional grants) will require detailed evidence of how much counties will need, for what and by when. vii Devolution without disruption – Pathways to a successful new Kenya Table 1: A phased approach to transferring functions to counties. PHASE Timing of transfer Readiness criteria Types of functions Examples PHASE 1 Immediately after • None/automatic. Functions not • Local government functions. county formation. currently being • County assemblies and performed by any executives. national agency. • Establishment of basic systems (PFM, HR, Procurement etc.) PHASE 2 As soon as basic • Basic county HR and Basic operation of • Operation of rural and district county systems are PFM systems. service delivery health facilities, ambulance operational. • Sectoral plans. programs. services. • Provision of agriculture extension services. PHASE 3 Only when specific • Specific system More complex • Purchasing and distribution readiness criteria are or capacity programs and supply of pharmaceuticals (only after met. requirements. chain management. stock control and ordering system in place). Source: World Bank staff analysis. 11. In one ‘moderate’ scenario, using 2010/11 as a 'simulation year', this report costed aggregate county needs at KES 169 billion, just to maintain services at their existing levels. This is above the amount foreshadowed in the 2012/13 Budget Policy Statement (KES 160 billion including devolved funds), but below the Commission on Revenue Allocation (CRA) estimate (KES 203 billion excluding devolved funds). Moreover, the figure needs to be understood with all its limitations: it is based on a number of assumptions, which ultimately the Government and Parliament will need to determine (such as the fate of devolved funds like the Constituency Development Fund (CDF), Local Authorities Transfer Fund (LATF), Road Maintenance Levy Fund (RMLF), etc.); it only costs existing services (and does not provide for additional equalization, for running the new administrations, or financing urban services), and; it does not account for the transfer of functions over time, or the possibility that seconded staff may remain on the national government’s payroll. Figure 2: Flows of revenues from different sources for county governments. National Revenue County own-source revenues (property tax, entertainment tax etc.) County equitable share (minimum 15%) Should be decided together Conditional or unconditional grants County Revenues Equalisation Fund Conditional grants (0.5%) Source: World Bank (2011b). viii Executive Summary 12. Because the total county financing needs will be in excess of 15 percent of audited national revenue, Kenya’s policy-makers should consider the full set of options for resourcing counties. Figure 2 shows the various ways in which county governments will be resourced, including the equitable share, but also own revenues, the Equalization Fund, and additional transfers from the national government (conditional or not). 13. It would be misguided to seek to meet all of county needs through the equitable share. The equitable share transfer, because it is unconditional, is critical in giving future counties substantial autonomy, but it also imposes significant constraints. Because the equitable share transfer is untied and formula-based, the central government should consider using additional conditional instruments to: (i) fund regional services which will be an additional cost burden to the counties hosting them; (ii) secure funding for critical programs at the local level; (iii) reward performance; and/or, (vi) correct for imbalances created by applying a crude formula to highly-diverse counties. Given these constraints and additional objectives, there may be a case for limiting the equitable share transfer at―or close to―the constitutional minimum of 15 percent. 14. Own-revenues will be critical for resourcing county governments and also―critically―for fostering accountability at the local level. Yet the Constitution only grants limited revenue-raising powers to counties (broadly those currently enjoyed by Local Authorities (LAs), which will remain highly transfer- dependent, unless new revenue-raising powers can be granted to them (see Figure 3). Figure 3: Vertical imbalance: Applying the CRA formula, only two counties would receive less than half of their resources from transfers. 100 90 80 70 60 50 Percent 40 30 20 10 0 Kisii Tharaka Turkana Mandera Wajir Nyamira Tana River Lamu Pokot Homa Bay Marakwet Marsabit Vihiga Garissa Bomet Kwale Makueni Nandi Trans Nzoia Baringo Samburu Kirinyaga Bungoma Isiolo Kitui Siaya Nyandarua Meru Busia Migori Murang'a Taita Taveta Laikipia Kericho Kisumu Embu Nyeri Kakamega Kajiado Nakuru Uasin Gishu Kiambu Mombasa Machakos Nairobi Narok Kili Transfers Own Revenue Source: World Bank staff calculations. 15. To maximize the own-revenue potential of Kenya’s counties, existing sources should be reformed and new ones found. The collection of property taxes should be strengthened (as Kenya under-collects by a wide margin compared to other countries) by updating revenue base information, updating ix Devolution without disruption – Pathways to a successful new Kenya rates, and minimizing applicable exclusions. There is also scope to revisit the Single Business Permit (SBP). However, increasing county fiscal autonomy would almost certainly require creating additional revenue sources, such as piggybacking county taxes on existing national taxes. In addition, it is likely that discussions over the sharing of taxation related to national resources will be brought to the fore. 16. In the short term, the priority is to ensure that county governments are legally entitled to collect revenues as they come into existence. This is because the constitutional base for county tax collection has not been converted into legal instruments to enable counties to collect revenues on day one. In order to address the current legal vacuum, a new legal instrument is required to allow counties to collect revenues immediately after the elections through local authorities staff and systems. Over time, a decision will be needed on tax administration at the county level, possibly involving Kenya Revenue Authority (KRA) acting as an agent of the counties. 17. There are powerful reasons to consider conditional grants to county governments, although to date, the debate has focused exclusively on the equitable share. Conditional funding will be important in at least five ways to support devolution: (i) to finance services that are hosted by one county but benefit several; (ii) to ensure that national priorities continue to be supported under devolution; (iii) to address region-specific needs (including marginalization); (iv) to mitigate possible shortcomings of the equitable share transfer formula in specific counties; and, (v) to support a subnational performance management system. 18. Provincial hospitals are the most important regional service to consider financing through conditional grants. If the entire cost burden of provincial hospitals falls on the eight counties that host them, with no additional financing to support them, there is a risk that services will be under-funded. Other regional services include livestock training facilities, and the capital works programs of eight regional water service boards. 19. In the short term, Kenya’s policy-makers will need to decide on the fate of three earmarked programs, which, if maintained─would constitute conditional grants. Alone, these three programs― the Road Maintenance Levy Fund (RMLF), the Constituency Development Fund (CDF) and the Local Authorities Transfer Fund (LATF)―accounted for KES 38 billion or 8 percent of 2010/11 national revenues. Maintaining them at their existing level would greatly affect the fiscal space available for other conditional grants, because the unconditional equitable share will still need to meet the constitutionally mandated 15 percent minimum. 20. A system of capital grants may be required to protect development spending under the new dispensation. Worldwide, subnational governments tend to spend a high proportion of their budgets on wages and salaries and relatively little on investment. In order to ensure minimum standards of service delivery throughout Kenya, and to promote catching up in capital-poor areas, the national government may consider implementing a capital grants scheme, based on a transparent formula. 21. Finally, conditional grants could provide the backbone of a county performance monitoring system. As counties come into existence, there is much uncertainty as to how well and how fast they will be able to perform their missions. As part of a subnational performance monitoring system, grants can help to: (i) spur healthy competition between counties, and provide the incentives for county governments; (ii) regularly monitor and report on their performance; and, (iii) identify key service delivery bottlenecks. x Executive Summary KEY RECOMMENDATIONS ON FINANCING COUNTY NEEDS. The process of clarifying function assignments should begin as soon as possible,led by the TA and capitalizing on work already carried out by Treasury and the CRA. There should be a roadmap to streamline and simplify the initial asymmetric transfer of functions, possibly through a three-phase approach with functions transferred in groups. Costing devolved functions should begin with historical costing, but a more detailed and thorough costing methodology will eventually need to take into account gaps peculiar to the Kenyan budget, and any equalization objectives, which will require significant redistribution or additional budget. Designing transfers to counties should consider all possible sources of revenue available to them and the full range of transfer instruments. The tradeoff between unconditional and conditional funding deserves particular attention, as the scope for one will constrain the use of the other. As a matter of urgency, the GoK should decide the fate of existing earmarked programs. In the immediate term, a decision should be made on financing a number of key ‘big ticket items’ which will greatly influence the estimation of total county needs, including provincial hospitals, CDF, LATF, RMLF, construction of new roads and the remuneration of seconded public servants. Legal instruments ought to be put in place to allow counties to continue collecting revenues and charges currently have collected by local authorities until taxation laws can be passed by county assemblies. A new interim national law to govern county own-revenues could provide an opportunity to revise existing taxes, particularly property rates and SBP. Additional local fiscal revenues should be sought among a range of possible options, including PIT surcharges, taxes on the use of motor vehicles, payroll taxes, etc. Particular consideration should be given to the sharing of taxation related to natural resources as their potential may increase vastly in years to come, and this issue often constitutes a source of conflict. Given the radical and experimental nature of Kenya’s devolution, it may make sense to maximize the scope for conditional funding (and minimize that of unconditional resourcing) if only to ensure that essential programs will remain funded. Conditional grants programs could be considered specifically for: o Regional services including provincial hospitals. o Capital projects. o Supporting urban service delivery. o Providing temporary stop-gap support to counties at risk of experiencing severe fiscal shortfalls during the transition. o Fostering inter-county performance and promote a system of performance monitoring at the local level. xi Devolution without disruption – Pathways to a successful new Kenya Promoting greater equity while safeguarding service delivery 22. Promoting greater equity in the allocation of spending and services is at the heart of Kenya’s new Constitution. Moreover, expectations are high that devolution will rapidly bring about spatial equalization, after years of unequal development across Kenya. And yet, seeking to equalize too much too quickly could be a risky strategy. The goal of equalization will need to be pursued in a tight fiscal environment―limiting the scope to increase spending―and without undermining existing service delivery, which is unequally distributed to start with. This has two major implications: first, existing imbalances may only be tackled over time second, equalization should target the people who will benefit from services, rather than trying to achieve equalization across geographic locations. 23. The CRA-proposed formula for allocating the equitable share is highly redistributive. While the formula places heavy emphasis on county population (45 percent), this is logical since population is the main driver of service needs. Yet taken together, all of the other components in the formula (amounting to 55 percent) favor counties that have been historically underserved (see Table 2). Table 2: Components’ source, weights and objectives in CRA's formula. Components Population Poverty Equal Share Land Area Fiscal Responsibility Year / Source Census (2009) Year / Source TBD (allocated equally initially Weight 45% 20% 25% 8% 2% Objective Resource counties Promote Provide each Factor in the Provide incentives to deliver services redistribution county with higher cost for prudent fiscal equally on a per in favour of resources to cover of delivering management capita basis historically lagging the fixed costs of services in areas running county remote, sparsely administrations populated areas irrespective of population or size Source: World Bank based on CRA report to Parliament, August 2002. 24. The simulated county transfers, using this formula, display a strong equalization bias. This is not obvious at first because, on an absolute basis, the better-off counties (such as Nairobi and Kakamega) will be receiving the lion’s share of the funds. However, the picture is almost entirely reversed once one looks at allocations on a per head basis, with Isiolo, Lamu, Marsabit and Tana River coming out the big winners. In other words, once population is “netted out� the poorer and smaller counties receive disproportionately big allocations (see Figure 4). 25. Because it promotes substantial re-distribution of resources across counties, the CRA formula creates risks for both “winners� and “losers� on day one. This can be seen by contrasting the existing costs of delivering devolved functions with the allocations that the proposed CRA formula would generate, if applied to the same total amount (see figure 5). There are two twin challenges. Areas that were historically privileged will inherit service delivery obligations that will require substantial funding (by definition above and beyond what a strictly population-based formula would provide). In these counties, the challenge will be to rapidly streamline service delivery without interrupting key services or making inefficient reallocations. By contrast, counties that were hitherto underserved will xii Executive Summary receive substantial extra cash (relative to the cost of their current service delivery obligations) and the challenge will be to manage these resources well, and to scale-up service delivery with limited capacity. Unfortunately, increasing the share of the equitable transfers to alleviate one bottleneck would only exacerbate another one. Figure 4: Per head equitable share allocations to counties. 7,000 (Assuming 15 percent equitable share, i.e KES 91.2 billion for 2012/13). 6,000 5,000 KES per head 4,000 3,000 2,000 1,000 0 Lamu Isiolo Marsabit Tana River Samburu Taita Taveta Wajir Turkana Elgeyo-Marakwet Busia Laikipia Garissa West Pokot Tharaka Nithi Baringo Kwale Migori Embu Kitui Vihiga Makueni Nyamira Nyandarua Bomet Kirinyaga Kajiado Nandi Trans Nzoia Nyeri Machakos Mandera Kericho Siaya Homa Bay Uasin Gishu Mombasa Murang'a Kisumu Kakamega Kisii Meru Bungoma Nakuru Kiambu Nairobi Narok Kili Source: World Bank staff calculations. Figure 5: A massive reallocation of funds across counties under a �scally prudent scenario. Difference between current distribution of spending and projected allocation under a KES 97.5 billion equitable share scenario 3 2 1 KES billions Nyeri Nairobi Homa Bay Murang'a Kirinyaga Embu Kiambu Mombasa Nakuru Garissa Kisumu Kericho West Pokot Isiolo Elgeyo-Marakwet Laikipia Nyandarua Lamu Baringo Meru Taita Taveta Migori Tana River Machakos Siaya Wajir Tharaka Nithi Kajiado Makueni Busia Marsabit Samburu Nyamira Kisii Vihiga Kakamega Uasin Gishu Nandi Kitui Kwale Bungoma Trans Nzoia Bomet Mandera Turkana Narok Kili -1 -2 -3 -4 Source: World Bank staff calculations. xiii Devolution without disruption – Pathways to a successful new Kenya 26. These likely imbalances call for three short-term actions to phase in equalization over time. First, it is urgent to model inherited costs of service delivery at the county level. This would require compiling detailed geographically-disaggregated spending data on services to be devolved, from which to estimate the future cost of service delivery at the county level. Second, with this information, the GoK may need to consider complementing the CRA-formula with temporary bridging grants that would be limited in time and specifically targeted to allow cash-strapped counties to maintain existing services at least at their current level. Third, a strategy is needed, to build capacity rapidly and systematically in lagging counties to develop adequate Public Financial Management (PFM), procurement, project management, Human Resources (HR) and service delivery capacity. Additionally, this may require only phasing in the transfer of functions over time, following predefined capacity benchmarks. 27. The Equalization Fund, provided for in the Constitution, will be too small to address deep spatial inequalities. At 0.5 percent of audited national revenues (and even increased to 0.8 percent as per the current Budget Policy Statement (BPS), the Equalization Fund will represent a fraction of the resources currently channeled through the CDF. Moreover, if the Fund crowds out existing funding for lagging areas, such as programs currently under the Ministry for the Development of Kenya and other Arid Lands, or the Ministry of Special Programs, its net impact could be nil or negative. Therefore, as a matter of urgency, policy-makers should clarify the Fund’s objectives, and target it as much as possible to specific areas and communities. Given its modest size, it might be better used as ‘seed money’ to leverage additional resources into a bigger, more effective fund. KEY RECOMMENDATIONS ON PROMOTING EQUALIZATION THROUGH TRANSFERS. It may make sense to limit the equitable share transfer at –or close to─15% during a transitional period, given the experimental nature of the CRA formula and lack of clear estimate of actual county needs and capacity. Unmet needs at county level could be filled via other unconditional or conditional instruments. In counties that have been historically over-serviced (relatively) the key would be to avoid service delivery interruptions, by extending ‘stop-gap’ funding to be phased out overtime. In counties that have been historically under-serviced and marginalized, it will be vital to build up the capacity to handle vastly increased funding efficiently and transparently. A priority should be to clarify the object of the Equalization Fund and particularly: o The definition of ‘marginalized areas. o The scope of interventions which should be focused on alleviating key service bottlenecks. It may not make sense to dedicate the Fund to financing infrastructure. Instead catalytic interventions could consist in addressing: o Staff incentives problems in working in remote locations. o Capacity constraints in applying for capital funding (under a capital grants program) and for managing capital projects. xiv Executive Summary Managing money efficiently and transparently for county development 28. Counties will need sound PFM systems to make effective use of their resources. At county level,this may be particularly challenging as many PFM related functions will need to be established, more or less from scratch. In turn this raises the question of the degree of oversight that the national government should be allowed to exert over counties; sufficient to lead capacity building and promote application of standards, but limited enough to prevent abuse and excessive control by the center. 29. Kenya’s new Constitution has provided the basis for a more coherent PFM legal framework. As well as replacing or consolidating a number of existing laws, the new Public Financial Management Act provides a framework for PFM in the new county governments as well as urban areas and cities and, for the first time in Kenya, there will be a single PFM legal framework for all levels of government. 30. Budgeting for the transition will be difficult given the asymmetric transfer of functions. The transition period―during which functions will be gradually transferred to counties―raises challenges for the annual budget process: the new counties may simply not have the capacity or the time to budget for 2013/14, and there is the additional question of how to budget for devolved functions temporarily executed at the national level. It is likely to take some time to establish the payroll systems in counties, to allow them to process salaries of seconded public servants locally. Some ‘bundling’ of functions to simplify the functional transfer process into manageable “chunks� (see Table 1) would also simplify the financing arrangements for these functions. 31. Even if the national government continues to manage funds for some devolved functions during the transition, allocations to counties could be clearly shown in the budget under ‘county votes’. While counties may not have the responsibility initially for all devolved powers and functions, specific resource allocations will be made to counties on day one. Grouping all these allocations under county votes clearly sets out in the national budget,would reassure county governments that funds are provided for them. The budget would show explicitly the total funding allocated to each county, and specify how much is to be spent directly by the county, and how much will be spent on its behalf, through national government systems. Alternatively, a transitional budget law passed at national level for 2013/14 could be used to achieve this. In any case, counties will need legislative authority―whether from their own laws or under a transitional national law―to spend funds from day one. 32. The new Constitution greatly increases the power of the legislature in the budget process, creating the risk of gridlock. In particular, because the Division of Revenue Bill (DoRB) is not defined as a money bill (which must be passed or amended by the Assembly only on the recommendation of the relevant committee of the Assembly) there is a risk that delays in approval of the DoRB could derail the budget process at both national and county levels. One approach adopted elsewhere is a two-tiered budget process in Parliament, where the legislature at national and county levels bind themselves to a fiscal framework for the duration of the fiscal year prior to the consideration of the DoRB, County Allocation of Revenue Bill (CARB), and budget estimates. The PFM Act seeks to implement such an approach, but this will require significant cooperation from parliament to ensure the DoRB is approved on time and in conformity with the BPS. xv Devolution without disruption – Pathways to a successful new Kenya 33. Building-up the capacity to budget at county level will be a tall order. Counties will be comprised of staff from Districts, who have very limited experience of preparing and managing budgets, and from former LAs, who have only limited experience in budget management. Therefore, in anticipation of the difficulty of establishing budgeting functions from scratch at county level, guidelines and associated templates should be developed to guide the formulation of county budgets. This would also provide an opportunity to integrate the planning and budgeting functions into a single process, to be overseen ideally by a single county institution. Lastly, it will also be essential to ensure that county budgets are prepared, executed and reported using a single country-wide chart of accounts. 34. Kenya’s counties may not manage to spend the funds available to them, without determined capacity building and monitoring. Experience among Kenya’s local authorities suggests that this under spending is an issue that many counties―particularly those historically marginalized―may face (see Figure 6). Specific capacity needs to be developed for resource allocation, cash management―in line with the principle of a treasury single account―and budgetary oversight through the Controller of Budget. The national government will also have a key role to play in setting standards and monitoring performance, for instance through league tables and rankings and a system of early warning to identify service delivery problems in devolved functions. One way of promoting this would be through a ‘county performance assessment tool’. A county performance monitoring system (consisting of an assessment and monitoring tool, and the administrative arrangements for implementing it), will establish as basis for central government county assemblies and citizens to hold county governments accountable, signal capacity gaps that require support, foster positive competition between counties, and focus leaders’ attention on key indicators of success. Figure 6: Counties where Local Authorities ran budget surpluses in 2008/09. 50 45 40 35 30 Percent 25 20 15 10 5 0 Tana River Wajir Marsabit Tharaka Kajiado Taita Taveta Kiambu Trans Nzoia Uasin Gishu Meru Garissa Kericho Kwale Mandera Migori Busia Nyandarua Kakamega Nyamira Laikipia Bomet Nakuru Turkana Kisumu Pokot Murang'a Nandi Kitui Bungoma Samburu Nairobi Narok Kili Source: World Bank staff calculations. xvi Executive Summary 35. PFM is only one dimension of promoting accountability at the local level―and there would be important payoffs from setting up accountability systems at the local level from the get-go. Contrary to the common expectation that decentralization enhances accountability and efficiency in service delivery, experience suggests that the contrary often holds true. Devolution presents a particular challenge for service delivery, by breaking apart existing (more centralized) accountability relationships, and requiring new ones to be established. This is also an opportunity to start from a clean slate and Kenya has all the assets to be a leader in Africa with respect to transparency and subnational government accountability. 36. Transparency, participation and accountability are clear requirements in the Constitution, but laws alone are insufficient to implement these principles. The experience of devolution in other countries indicates the importance of building mechanisms that ensure subnational governments are not only accountable upward, but also downward to citizens. Establishing formal government systems is a necessary but not sufficient condition for good governance at the devolved Figure 7: Elements of social accountability systems. level. Similarly, demand-side social accountability initiatives by themselves face major challenges of sustainability Government and scale, and are typically heavily reliant on donor financing. Integration Transparency: Participation of demand and supply side systems is information Accountability and Feedback: for citizens information therefore needed. An effective social from citizens accountability system, embedded in county governments, requires three core elements: (i) fiscal transparency; Citizens (ii) participation mechanisms; and, (iii) accountability mechanisms (see Source: World Bank. Figure 7). 37. Kenya can draw on its extensive experience with devolved funds to reinforce both upward and downward accountability. For instance the LATF system has been largely successful in ensuring reporting by local authorities, even though the information they produce is not usually made public. As for the CDF experience, it shows that there is an important role for civil society in auditing projects. Building on Local Authority Services Delivery Action Plan (LASDAP), county governments could be required to involve citizens in the budget and planning process in line with the PFM Act. 38. For social accountability systems to function, supporting institutions need to be put in place. Key elements include accurate and timely financial and performance information, and in turn this would require both capacity building and enforcement mechanisms (such as penalties for failure to produce accurate information). A second priority is to strengthen the relation between planning and the budget process, since public participation in planning will only be meaningful if the choices made are translated into spending. A third priority is to incorporate transparency and participation mechanisms in county government systems―systems to share information (disclose, simplify, and disseminate) on budget and spending, and to enable effective citizen participation in setting social service delivery priorities and monitoring performance. In addition to creating an enabling environment for citizen participation xvii Devolution without disruption – Pathways to a successful new Kenya and oversight, county governors and administrators can make use of civil society capacity to use social accountability tools, such as participatory budgeting, scorecards, social audits and procurement oversight committees. KEY RECOMMENDATIONS FOR PROMOTING SOUND AND TRANSPARENT FINANCIAL MANAGEMENT IN COUNTIES. During the transition period, the budget could include “county votes� showing total county allocations as well as the portion that remains executed by the national government on behalf of counties. This could include a transitional budget law, passed at national level for 2013/14 with forty-seven county budget schedules showing county allocations disaggregated by function. A two-step budget process could be introduced with parliamentary votes at each stage to generate consensus around the fiscal framework and mitigate the risk of Executive-Legislative gridlock. Building county PFM systems from scratch will require central support to capacity building, monitoring of county progress against clear benchmarks and standardized guidelines, and templates developed nationally. Public financial information (including results) should be made public in a way that allows citizens to assess the efficiency and effectiveness of national and county spending. Build transparency and participation measures into regulations and county systems for planning, budgeting and monitoring of expenditures and service delivery. A key element for promoting good financial management will be to quickly put in place and M&E system to benchmark county starting pointson a range of development and service delivery performance indicators, and to monitor and disclose county performance over time. Citizens should be actively involved in planning and budgeting at local level and equipped with the tools to make significant contributions. Accounting systems at county level should allow to track spending on individual projects and by service delivery unit. Social accountability toolkits should be developed for the benefit of local administrators. Protecting the urban growth engine 39. Unless corrective action is taken, Kenya’s cities may well be the big losers of the devolution process. This would be dramatic as Kenya’s cities are growth engines for the entire country, and will be the main source of own-revenues for many counties. Kenya is experiencing a demographic transition that will see an increasing number of people moving to the cities to seek the opportunities that urban areas have to offer. Cities are also increasingly useful and meaningful to rural residents, as they they provide services and infrastructure that reach well beyond city boundaries. 40. Kenya’s devolution is unique in that it involves simultaneous decentralization of key services and resources from the national to county governments, but also recentralization of urban management. Existing arrangements will be profoundly affected in the new dispensation, as the current system of local authorities, with elected management and significant discretion over resources and functions, will be replaced by a new system that gives more power to county executives. In the new dispensation, Kenya’s cities will be managed by appointed boards, with far less power and autonomy than the local authorities they replace: most of their functions will be delegated by county governments, and they will have no guaranteed funding. xviii Executive Summary 41. Moreover, only three urban centers will have Table 3: Twenty-one urban centres with more than municipal or city boards. The Urban Areas and 80,000 residents will not have municipal boards. Cities Act (UACA) sets the population threshold Urban Centre Population for city status at 500,000, and for a municipality Ruiru 240,000 at 250,000, and only cities and municipalities are Kikuyu 230,000 entitled to boards. Yet at present there are only Kangundo-Tala 220,000 five urban areas in Kenya with a population of over Naivasha 170,000 250,000 (Nairobi, Mombasa, Kisumu, Nakuru and Machakos 150,000 Eldoret), two of which will be city counties, and Mavoko 137,000 will be governed as county governments (without city boards). In all other urban centers, a rather Thika 135,000 uncertain arrangement for “town committees� Vihiga 120,000 will apply. A number of substantial urban centers Nyeri 120,000 with viable existing local governments, including Malindi 120,000 twenty-one urban centers with more than 80,000 Ngong 110,000 residents, will thus be effectively recentralized into Kitui 109,000 the county administration (see Table 3). Karuri 107,000 Mumias 100,000 42. Looking forward, there are three ways in which the Kitale 100,000 management autonomy of urban centers could Kericho 100,000 be expanded. One is by lowering the population Kimilili 95,000 threshold for the definition of municipalities. Awasi 93,000 Alternatively, the Act could deem all county capitals Kakamega 90,000 to be municipalities (which would still leave out Kiambu 85,000 important towns like Ruiru, Naivasha, Ngong, etc.). A third option is to enhance the managerial Kisii 80,000 Source: KNBS (2009). autonomy of town committees to make them function in the same way as municipal boards. 43. With all urban functions and resources vested by the Constitution in county governments, specific functions and resources will need to be delegated to city and municipal boards. Yet there is no clear process or framework for such delegation, and no transparency requirement concerning the delegation by county governments to city and municipal boards. A useful measure would be to require county governments to publish the functions and revenue streams assigned to urban boards, so that urban residents understand what services they are entitled to receive, and as services from which level of authority. 44. Kenya’s Constitution makes county governments responsible for financing urban service delivery, and as a result, there is a risk that urban services may be under-funded. Because rural residents will dominate most counties (see Figure 8), county governments may chose to preference rural services and to redistribute revenues raised from urban residents to their rural constituencies. This may jeopardize the economic development potential of Kenya’s urban areas, and violate the fiscal federalism axiom that revenues and expenditure responsibilities should be aligned to the extent possible at the local level. xix Devolution without disruption – Pathways to a successful new Kenya Figure 8: Despite rapid urbanization, most counties are still predominantly rural. 100 90 80 70 60 50 Percent 40 30 20 10 0 Mombasa Nairobi Kiambu Kisumu Machakos Nakuru Isiolo Kajiado Uasin Gishu Migori Vihiga Kericho Laikipia Nyeri Garissa Taita Taveta Marsabit Bungoma Kisii Trans Nzoia Lamu Nyandarua Bomet Kwale Mandera Samburu Busia Murang’a Embu Kirinyaga Kakamega Tana River Wajir Marakwet Homa Bay Turkana Nyamira Kitui Nandi Meru Makueni Baringo Siaya West Pokot Tharaka Narok Kili Urban Rural Source: World Bank staff analysis. 45. Urban service delivery will depend largely on the priority which county assemblies give urban issues, but the national government could also consider earmarked urban grants. The UACA envisages that county governments will provide transfers to city and municipal boards, but not to town committees and it offers no guidance as to how these amounts ought to be calculated. In this context, the national government could help to ensure that urban services are adequately provided through earmarked urban services grants, which could be paid either to county governments or to the urban boards directly. To ensure that counties maintain their own levels of funding for urban services, the transfer could include an additionality clause, binding the county to maintain a certain level of funding from their own sources. 46. However, even additional transfers may be insufficient to finance the type of infrastructure that Kenya’s cities need. While counties may be able to borrow, this will be subject to national government review and guarantee, helping to reduce the risk of uncontrolled borrowing at the subnational level. However, the flipside may well be insufficient access to capital for infrastructure, particularly in Kenya’s larger cities, as the national government may well be reluctant to encourage subnational borrowing that would end-up on its balance sheet. xx Executive Summary KEY RECOMMENDATIONS ON PROTECTING URBAN AREAS UNDER DEVOLUTION More urban centers should have corporate bodies to manage them. This could be achieved either by lowering the threshold for ‘municipality’ status, or by amending the powers and functions of towns under the current law. Functions of city and municipal boards should be clarified, as well as the formal process for counties to delegate additional functions to them. Standards ought to be set for urban service delivery, and urban boards should be required to report against those to the county assembly. City and municipal boards should be given own revenue powers through delegation of additional national taxing power. Cost benchmarks are needed to monitor future resourcing of urban functions. The current cost of urban services should be calculated, based on past spending and monitor that funding is adequate to at least maintain resourcing for services at current levels. The role and functions of local authorities should be extended during the transition until their staff, functions and assets can be accommodated by the new counties. The national government could help to ensure that urban services are adequately provided through earmarked urban service grants. To ensure that counties maintain their own level of funding for urban services, transfers could include an additionality clause, which binds the counties to maintain a certain level of funding to urban services. Decentralizing Public Service in a sustainable way 47. Kenya’s existing public service structure is currently highly centralized, so devolution will bring about major changes. Forty-seven new county civil services will be created, with two immediate sets of challenges: (i) to define the overall governance framework for county civil services; and, (ii) to manage the HR transition implied by their creation. 48. The new devolved arrangements offer the opportunity to rationalize the current highly fragmented arrangements for service delivery. By contrast, counties will have considerable autonomy over public service management: the county governments will be responsible, within a framework of uniform norms and standards, for establishing and abolishing offices, appointing public servants and exercising disciplinary control over them. In each county, a Public Service Board―whose members will be nominated by the Governor―will exercise these powers on behalf of the government. 49. The County Governments Act (CGA) provides the regulatory framework for counties to engage their own public servants but it leaves a number of important gaps in the policy framework. While the Constitution provides clear authority for the national government to set up a framework of national standards, this could still leave room for counties to regulate some aspects of civil service management. Going forward, there should be a clear decision as to which issues national laws and policies should cover, and which should be left to counties. At present, for example, it is not clear who decides county public service remuneration, and there is no uniform discipline procedure applicable to all counties. The new framework should set national standards and guide counties in their day-to-day management of public servants, by setting the general framework for management, specifying who is responsible for regulating the detail of how county public servants should be managed, and providing interim statutory instruments to specify the detail of the management arrangements, until county governments develop capacity to pass their own. xxi Devolution without disruption – Pathways to a successful new Kenya 50. It is unclear which agency should take the lead in proposing laws to fill existing regulatory gaps. The most immediate issue is to determine which agency in government should be responsible for proposing the laws to cover county civil servants. From a legal perspective, the missing elements of the regulatory framework could be provided under the CGA, or they could be in separate legislation administered by the Ministry of State for Public Service. If the former approach is adopted, two ministries will regulate public service matters at different levels of the government (since the CGA is the responsibility of the Minister responsible for intergovernmental relations). If the latter is chosen, the two ministries will administer policy on county public service concurrently, with scope for overlap and contradiction. 51. Devolving public service management involves balancing the risks of too much central control against excessive local autonomy. With too much control, county autonomy and accountability may be undermined. With too little oversight, local controls could be ineffective, undermining service delivery. This is a paradox of decentralization: effective devolution actually requires a strong central government. Failing local controls could result, for instance, in elite capture and politicization of appointments. Establishing independent county Public Service Boards has the potential to mitigate these risks, but they arguably have been given too much power. In particular, the concentration of human resource management creates the potential for fiscally unsustainable recruitment, and excludes direct supervisors and other stakeholders from an appropriate role in recruitment. In addition, the Governor will appoint board members with the approval of the county Assembly, but without a specific process to ensure their competence and political independence. Regulations under the CGA could set out the process for selecting county Public Service Board members and the national government should provide support (training, mentoring, procedures, etc.) to the boards as they come into existence. 52. Uncontrolled spending on personnel expenditure is a common side effect of decentralization and will be a major challenge for Kenya at both levels of governement. The reasons for excessive local spending on wages are complex: they often reflect patronage, but also the fact that it is easier to spend money on salaries than on investment projects. Spending on personnel alone does not contribute to better services per se, if the recruited staff have insufficient access to funding for operating costs and facilities. In Kenya, there is already anecdotal evidence that uncontrolled salary spending is rife in local authorities, despite national controls. The public service provisions of the CGA emphasize control over recruitment of public servants, but leave employment of non-public-service staff (short-term and consultants) relatively unregulated. 53. Paradoxically, devolution could widen capacity gaps across Kenya. There is already gross inequity in the distribution of public service skills across the country. For instance, the ratio of doctors to population varies across counties by a factor of 90! Yet the counties that currently have the lowest levels of public service skills are also likely to be those which will have the hardest time attracting skilled personnel, given their remoteness, lack of mobility prospects, and weak systems for incentivizing staff. In turn, this could put undue upward pressure on remuneration, crowding out other important types of expenditures. These risks call for incentives―especially non-financial―within the framework of a single public service where staff can be offered greater access to training and promotion if they serve in remote counties, and where mobility from one county to another can be guaranteed. 54. In the coming months, managing the transfer of staff to counties will be the number one challenge. Most of the public servants needed to run county functions are already there but the conditions under which they will remain are unclear, and the fiscal implications have not been properly assessed. The biggest risks are associated with the legal framework for staff to move across to county governments under the CGA. xxii Executive Summary 55. Initially, national public servants will be seconded to county governments but the “zero-basing� approach is unusual and risky. Most commonly, countries that establish a new level of government simply transfer existing staff performing devolved functions to the new governments. In Kenya, in order to maximize county flexibility, a secondment approach has been chosen. Counties will then be free to develop their own staff structures and recruitment processes. The secondment of national staff will come to an end when the seconded officer is either appointed to the county public service, or handed back to the national government. Moreover, if public servants feel they have the right of return to the national government, they may also opt ex ante not to apply for a position in the county civil service. But if a large number of secondees are returned to the national government, it may not have the jobs or the funding to absorb them. 56. It is unclear how the salaries of the seconded staff will be managed. Initially staff will probably need to be paid at the national level, since it will take time to establish payroll systems at the county level. But it is not fiscally realistic to expect the national government to transfer the full amount of the county revenue share, while continuing to pay the salaries of seconded staff. Therefore, some arrangement will need to be made to fund salaries of county public servants from the county equitable share (although no conditions may be attached to it), or the national government could be left with an unsustainable fiscal burden. 57. The fate of local authorities’ staff needs to be addressed. Existing public employees at the county level include some 33,000 LA staff of whom 30,000 are not civil servants. They are employed and managed by the LAs themselves outside of the state civil service. But LAs will cease to exist on the day of the next general elections, and further laws called for under the UACA to cover what happens to the staff, assets and liabilities of the former LAs have not yet been prepared. KEY RECOMMENDATIONS FOR MANAGING THE PUBLIC SERVICE TRANSITION. National policy on county public services should be developed but first a national law should make it clear who is responsible for it. At the very least, the framework of regulation should provide uniform procedures and a comprehensive regulatory framework to apply until the counties pass their own laws. National regulation of some aspects of county public service management is needed, in order to maximize career progression opportunities and encourage public service mobility. Some national standardization of county pay policy would be beneficial, but care needs to be taken not to impose unaffordable fiscal burdens on county governments. The national government should support the establishment of county Public Service Boards to ensure that they are fully independent and competent and a regulation under the CGA should set out the process and criteria for appointing Board members. As a matter of urgency the fate of Local Authorities’ employees should be clarified through a law to deal with transition issues involved in abolishing local authorities. Priority should be to tighten the loopholes that may allow uncontrolled non public service employment at the county level. As a matter of priority, the Government of Kenya (GoK) should decide which level will be responsible for paying the salaries of seconded staff, and if these will be deducted from the counties’ equitable share transfers. A plan should be developed to absorb redundant staff at the national level should a significant number of county secondees seek to return to the national civil service. xxiii Devolution without disruption – Pathways to a successful new Kenya Promoting intergovernmental coordination between national and county governments 58. Devolution complicates the management of government everywhere but intergovernmental coordination will be particularly important in Kenya. Devolution potentially diffuses accountability between levels of government, and introduces the possibility of mismatched resources, responsibility and authority. In Kenya, the Constitution mandates shared responsibility for some important aspects of service delivery, with the national government generally being responsible for policy, and counties in charge of implementation. But it gives the national government limited fiscal or supervisory levers with which to influence the achievement of national policy priorities. In the absence of such levers, given Kenya’s long standing history of distrust between government and local stakeholders, and the fact that no level of government is superior to the other, the institutions of cooperative government become all the more important. 59. The current framework for devolution focuses mainly on the relations between governors and around the budget process, overlooking issues in sectors. International experience suggests that frictions between levels of government undermine service delivery. In Kenya, the main devolved sectors are health, agriculture and livestock. Coordination and decision-making will be complicated by the sheer number of counties, and creative approaches will be required to ensure that the coordination bodies can still make decisions effectively. 60. Effective intergovernmental coordination will require both a change of culture and capacity building at both levels of government. Line ministry staff will have to reorient from a service delivery role to a policy role, and understand how to use new tools at their disposal to influence policy outcomes (like standards and expenditure norms). At the county level, sector staff will also have to adapt to new roles, including resource constrained budgeting and independent formulation of policy or legislation. 61. Two laws establish the framework for future relations between levels of government. The Intergovernmental Relations Act (IRA) establishes intergovernmental mechanisms (such as the National and County Government Summit, the Intergovernmental Relations Technical Committee and the Council of Governors) and the Public Financial Management Act (PFMA) establishes an Intergovernmental Budget and Economic Council (IBEC). 62. A particular challenge will be to get the system working with forty-seven counties, and to effectively coordinate implementation. Many of the international models for intergovernmental coordination involve much smaller numbers of subnational units. Moreover, in some countries the main objective of coordination is joint policy development. In Kenya, the real challenge will be to coordinate at the level of implementation in areas where both levels of government share service delivery responsibilities. The first area of policy coordination that will be required is the unbundling of the constituent elements of the different functions, to ensure clarity as to who will do what. As part of this exercise therefore should also be a joint planning process to work out day-to-day coordination around implementation. 63. It is not yet clear how the existing system of provincial and district administration will relate to county governments. The Constitution requires the system of provincial administration to be restructured in line with devolution within five years, but how exactly remains to be worked out. A key question is how the remaining national functions not devolved to county governments will be organized: In some cases functions that are best performed at local level (like social welfare or children’s and gender programs) xxiv Executive Summary have not been devolved by the Constitution, and in others functions of existing department heads might be split (like education). Therefore it will be important to define a new architecture both for day-to-day relationships between national and local agencies in each county, and also for arranging the functions that remain national. 64. Given the resilience of provincial administration over the years, Kenya will need to manage the risk that parallel and competing structures could undermine the effectiveness of county government. Administrative parallelism, under which staff of central government either continues to manage subnational functions or jointly administer them alongside national staff, would weaken the accountability of county governments. KEY RECOMMENDATIONS FOR PROMOTING INTERGOVERNMENTAL COORDINATION UNDER DEVOLUTION. It is important to focus on getting intergovernmental relations bodies functioning early on. The first priority is to get sector bodies working on issues of function assignment, service delivery standards, and performance monitoring. The second priority is to focus on intergovernmental relationships at the county level, particularly by resolving the role of provincial and district administration staff, and designing the new arrangements for remaining national staff at county level. Managing the transition to hit the ground running 65. Kenya’s transition to devolved government will be as complex as it is ambitious. Managing such dramatic transition will be key since it will set devolution back or forward for years to come if implementation is chaotic. Citizens will quickly lose their faith in county government if its inception results in an immediate negative impact on service delivery. Immediate challenges include the movement of public servants from line ministries to county governments, setting up sound PFM systems in counties, and ensuring continuity of urban services. 66. For the transition to be as smoothas possible, Kenya needs institutional arrangements for managing the transition itself. There is a clear legal framework but much work remains to be done. In February 2012, Parliament enacted three laws to implement Chapter 11 of the Constitution. Together with the provisions of the Fifth Schedule of the Constitution, they provide a set of institutional arrangements for managing transition through the TA, an independent body with broad membership and powers to coordinate implementation, by the various organs of government. 67. The Transition Authority has a great deal to accomplish in a very short time. It faces parallel challenges of establishing its own internal systems, recruiting staff and developing government-wide networks at the same time as its workload is heaviest and most urgent. The CRA and the Commission on Implementation of the Constitution (CIC) were established in early 2011, and found their operations were constrained by not having their own budget appropriation until July 2011, when the 2011/12 fiscal year began. 68. It will need to command respect across government. The TA cannot make devolution happen alone. Devolution will occur through the actions of line ministries in each sector, and the TAs job will be to ensure that they all do their part. Although the Transition to Devolved Government Act (TTDG Act) xxv Devolution without disruption – Pathways to a successful new Kenya gives the TA power to make regulations, there is no clear sanction if they are not followed. The best way the TA can ensure that the rest of government follows its lead is by: (i) ensuring that it gets high level government sign off on its strategic direction; (ii) meaningful involvement of the implementing ministries; and, (iii) using transparent reporting of expectations and progress in meeting them to create an environment in which line ministries feel public pressure to meet their planned obligations. 69. A detailed planning process for the transition should also be considered. Under the TTDG Act, the CIC is responsible for requiring individual ministries to submit implementation plans and for monitoring their implementation, but the TA issues the guidelines about what the plans should contain. The guidelines issued by the TA could play a crucial role in setting the agenda of issues the line ministries should address, but ideally they should be determined in dialogue with line ministries. This is why it would be useful to develop an overarching strategy as well. The strategy should guide the decision about what to do first. It should highlight the respective roles of the TA and line ministries, and the relationship between the bureaucratic machinery and the political leadership. 70. The process of developing the detail of the system of devolution has so far not been well integrated with line ministries. It is the line ministries who are devolving their functions, seconding their staff, and reorienting their own roles to reflect a new focus on policy, standards and service delivery performance, and supporting the development of county capacity. Ministries are most likely to undertake effective change if they are actively responsible for planning it. Some further ways in which the TA can enhance coordination could include: seconding an official from each main ministry to the TA, providing line ministries with standard toolkits that they can use, and holding regular structured discussions with the Principal Secretaries (Permanent Secretaries) of the key ministries in devolved sectors. 71. The TA will also need to engage county governments on day one. A great deal will be expected of the forty-seven new governors, and they should have the basic systems in place that allow them to achieve some early wins. The TA could appoint teams for each county to help get these systems established. It is helpful to think about what decisions the county governments will want to make first, and focus on putting in place the systems they will need for those. For example, the county assemblies will need clerks and other staff to make the work of the assembly and its committees effective. Those staff can only be appointed by the County Public Service Board and therefore appointing the County Public Service Boards and establishing the systems to make them functionally effective, are first-order priorities. 72. There will be big differences in capacity between weaker and stronger counties. One of the main risks of devolution is that it actually exacerbates these gaps. The TA should develop a strategy for ensuring that weaker counties receive special assistance. This might involve targeting donors to give priority to supporting specific counties, or developing special programs of assistance for them. xxvi Executive Summary KEY RECOMMENDATIONS TO MANAGE THE TRANSITION PROCESS. The Transition Authority will need to be empowered financially and statutorily to lead the transition process with clear authority over other organs of government. The TA should develop an overarching strategy for implementing devolution, and provide uniform guidelines for line ministries to follow when preparing their implementation plans. The TA should appoint teams in each county, to help establish basic systems in a sequence prioritizing the first decisions county governments will have to make. The TA should develop a strategy for filling major capacity gaps in Kenya’s most disadvantaged counties, so as to pre-empt a widening of these gaps under devolution. Conclusion 73. Managing expectations will be a big challenge because decentralization is no silver-bullet. It will take time to balance resource allocations, let alone improve equity of access to services. Many Kenyans believe that devolution will bring dramatic change overnight: upgraded infrastructure, more jobs and opportunities and better services (see Box 1). But they have different and often contradictory views of how this will happen. For instance, lagging areas are counting on redistribution of national wealth to help them catch-up, while leading regions see devolution as a chance to run their own affairs un- impeded. Box 1: What Kenyans hope devolved government will do for them. Alice Vutage – Housekeeper in Nairobi. Born in Western Kenya 32 years ago, Alice Vutage migrated to Nairobi in 2000 in search of employment. With little education and great determination to support her family back home, she found a job as housekeeper. Alice is not conversant with the new Constitution. All she knows is that it will improve the livelihood of Kenyans, a fact gathered from her daily interaction with friends and relatives. “People say that the new Constitution will bring a lot of development in the country, and this makes me happy, because I would like to see people in my village leading a better life,� she said. Alice who is a single mother of a two year old daughter, hopes that the new Constitution will help to create job opportunities in her rural area, so that people can engage in economic activities, and become less dependent on financial support from relatives who work in big cities. “I am really eager to see how life will improve for my daughter and I when the new Constitution is implemented,� she said, with a hint of apprehension in her voice. Source: World Bank interview. 74. The transition to devolved government will be long, complex and risky but the potential payoffs are commensurately high. This report addresses both technical and transition challenges that Kenya will be facing in the months ahead. Part I, sets the stage and provides contextual elements. It retraces the historical and political-economy context against which the devolution project has been developed (chapter 1), and emphasizes the ambitious character of the administrative reorganization it implies in a country that remains highly unequal (chapters 2 and 3). Part II, focuses on the main parameters of the devolved intergovernmental fiscal architecture, starting with an estimation of future county needs (chapters 4 and 5), and laying out the overall funding architecture through which they will have to be addressed (chapter 6). Part III, examines each of the possible funding streams for decentralized services including own revenue sources (chapter 7), the unconditional equitable share (chapter 8), additional xxvii Devolution without disruption – Pathways to a successful new Kenya conditional instruments (chapter 9), and the Equalization Fund (chapter 10). Finally, Part IV, tackles implementation challenges in translating into reality the vision of a more responsive and accountable government through devolution. Chapter 11 addresses intergovernmental coordination challenges, while chapters 12 and 13 focus on the systems to put in place in order to ensure that sound PFM and effective social accountability are effectively promoted in the new counties. The final chapters relate to key issues in the transition namely the management of urban areas (chapter 14), the public service implications of devolution (chapter 15), and the management of the transition itself (chapter 16). xxviii PART I Seeing the future from the perspective of the past CHAP TE R O N E Kenya’s devolution in context 1.1 The enormous aspirations driving the new Constitution, and especially the devolution provisions, derive from three major expected changes. First, a significant share of government spending will reach the local level, instead of being consumed at the higher levels of government in Nairobi. Second, the distribution of resources between regions and localities will be more equitable, thus leveling the social and economic opportunities of Kenyans. Third, government will be more accountable, open and participatory, leading to reduced corruption and increased responsiveness to citizen needs. As captured by the phrase ‘our turn to eat’, many Kenyans believe that the dominant dynamic in their political system since independence, has been the capture of resources by particular ethnic groups through centralized control over the organs of the government. In the wake of the worst ethnic violence in Kenya’s history following the 2007 election, the power-sharing coalition government led by President Mwai Kibaki and Prime Minister Raila Odinga, gave birth to a constitutional revolution on 27 August 2010. 1.2 The practical changes introduced by the Constitution are dramatic and wide-ranging. Independent and accountable institutions, transparent appointment of public officials, and the complete separation of Parliament and the Executive under a presidential system of government, characterize the new Kenyan state. The centerpiece of the changes is the devolution of powers to a new tier of constitutionally entrenched county governments. In terms of achieving the promise Figure 1-1: Number of bills considered by parliament per year. of greater equity and participation, 70 devolution lies at the heart of 65 the new system of government. 60 Bills considered per year Devolution will be one of the 50 most complex aspects of the 40 new constitutional arrangements 32 30 26 to implement, but it holds the 24 20 20 promise of having the greatest impact on the lives of ordinary 10 Kenyans, particularly those living 0 in traditionally marginalised and 2007 2008 2009 2010 2011 peripheral regions of the country Source: Calculation from data on KenyaLaw.org (see Figure 1-1). A turning point in a long historical process 1.3 Devolution has always been a highly passionate and also divisive issue in Kenya. While the state that was ushered into place at independence was characterised by some regional autonomy, President Kenyatta moved rapidly to centralize and consolidate state power, and President Moi furthered this trend, influencing key decisions, including the formation of the Judiciary and the Parliament. 4 Devolution without disruption – Pathways to a successful new Kenya 1.4 A decade of relatively piecemeal decentralization began in 1999, lasting until the passage of the new Constitution. This decade saw the introduction of devolved (geographically earmarked) funds in an attempt to address spatial inequality. The most notable were the Local Authority Transfer Fund (LATF) created through the LATF Act No 8 of 1998; the Road Maintenance Levy Fund, (RMLF) created through the Kenya Roads Act, 2007; the Rural Electrification Fund, created through the Energy Act of 2006; and, the Constituency Development Fund (CDF), created through the CDF Act of 2003. Despite these piecemeal efforts to address inequality in resource distribution, political tensions remained highand continued to fuel demands for more local autonomy. 1.5 The defeat of the Kenya African National Union (KANU) in the December 2002 general election opened the possibility for change. The united National Alliance of Rainbow Coalition (NARC) promised a new Constitution soon after it was installed in power, but the initial draft produced by the Constitution of Kenya Review Commission, while containing some provisions for devolution, also concentrated powers in the Presidency, which was a key point of contention. The coalition government that came to being after the 2007 election was finally able to present a new constitutional draft, which was approved in 2010, with devolution at its core, as well as numerous checks and balances in the national government. This Constitution was approved by 67 percent of voters in a referendum, and officially promulgated on 27 August 2010. The Constitution establishes forty-seven county governments, which are based on the districts established by the Provinces and Districts Act of 1992. The guarantee of 15 percent of national revenues to be transferred to county governments on a purely unconditional basis, and the assignment of such major state functions in the areas of health and agriculture to county governments, go far beyond the historical precedent. 1.6 The strong provisions for devolution in the new Constitution were a key source of public support for the draft of the Constitution. According to survey results presented by Kramon and Posner1, just under 20 percent of supporters of the new Constitution (self-described ‘yes’ voters) did so because of the strong provisions for devolution; this is the second most common response after a simple desire for change, which measured at just over 20 percent. In other words, devolution featured prominently in voters’ decision to support the new Constitution, and this support will play a key role in the roll out of the county governments. 1.7 The provisions for devolution emerged from a history of debate, and even struggle regarding the proper balance of power between the center and locally representative governments. This issue dates to colonial times, and forms a central theme in the design of colonial government, the Mau Mau rebellion in the 1950s, and throughout the various efforts to reform the Constitution of Kenya. The proponents of a strong center argue for the need for control from the center, national solidarity and the exploitation of economies of scale provided by centralisation. Devolution advocates stress local rights, enhanced accountability, participation and―above all―equity. The issue of equity is of particular concern for the pro-devolution advocates, given the large disparities of wealth and social outcomes between regions and communities. However, it is as yet unclear whether the new devolved system delivers the benefits which its proponents hope for, as much will effectively depend on the actual process of implementation. 1 Kramon and Posner (2001). 5 Chapter 1: Kenya’s Devolution in context Deep-rooted political economy dynamics 1.8 Kenya’s efforts to devolve come in the context of centralized power in the Executive, which is a challenging starting point. While many countries, such as the United States, started with a set of states and then built a central government through a federal system, Kenya starts from a highly centralized system, then seeks to build up the local level. This form of devolution therefore relies on a contradiction because the center must willingly give up power to the local level. In practically every devolution experience, this contradiction haunts the process. At times, it leads to vacillations in devolution, as in Pakistan and Uganda; in others, this limits the level of devolution, as in China. 1.9 A related dynamic in the Kenyan context is the risk of conflict. The dominance of the Executive in distributing services, projects and other resources, which often operate under a patronage system, is commonly perceived to provide an incentive for conflict.2 Politicians reinforce this perception when they campaign for their community, so that they can take their ‘turn to eat’. Many communities and regions experience a highly unequal distribution of state resources, as presented in the subsequent section. 1.10 The push for devolution stems from an effort to reverse these trends, through diffused power and reduced conflict. By providing county governments with substantial resources and responsibilities for many service delivery functions, the Constitution seeks to counter the above trends by both equalizing the distribution of resources, and also reducing the all-or-nothing view of the presidency, which is driven by the idea that the group which controls the Executive, has near exclusive access to the state’s resources. 1.11 The new Constitution―with a highly ambitious plan for reform―emerged at a time of great division in Kenya. The post election violence of 2007-2008 represented a low point for relations between several of the highly diverse communities in Kenya, with estimates ranging from 800 to 1,500 dead, and hundreds of thousands displaced. The political leadership, however, came together to form the coalition government that would seek to unite Kenyans and stop the violence, with the Government of National Unity finalizing its agreement on April 13, 2008. Under this coalition, the government has successfully drafted and facilitated public approval of the new Constitution. The government, however, clearly maintains political divisions that lead to conflict, and a lack of coordination in the implementation of the Constitution. 1.12 Further divisions are apparent in the highly complex and multi-faceted devolution process, that is squeezed into a short time frame. The process of implementing the Constitution provides a schedule by which important acts must be approved by Parliament. The number of bills considered by Parliament has more than doubled from previous years, with 65 bills being considered in 2011 (see Figure 1-1). The ambitious legislative program has led to a number of frictions in the process, including political fighting between various constituencies. 1.13 International experience suggests that there are a number of serious political economy risks to devolution. The numerous experiences with devolution show that it does not follow a singular path, but rather, can follow a number of routes. Much of this variation depends on the motives for devolution; whether to increase democracy or to appease local separatist movements. Moreover, 2 Wrong (2009). 6 Devolution without disruption – Pathways to a successful new Kenya the type of devolution is quite different, whether to a quite large unit as in states in Brazil, or local governments in Uganda. However, despite this variation, in none of these cases was the process of devolution simple and without conflict. “Decentralization comes in many shapes and sizes, but in every instance, it involves changing the institutional rules that divide resources and responsibilities among levels of government. Politicians and bureaucrats thus fight over decentralization for the same reason that they fight over the design of state institutions more generally: their power and authority are at stake.�3 1.14 First, many officials in the center will inevitably resist relinquishing power and authority,often for good institutional reasons. Naturally, officials who are accustomed to a certain level of authority, as well as a particular oversight role, will be hesitant to give up that authority. Such resistance will arise in both the policy process, as well as the actual process of devolution, when officials will be asked to substantially change their day-to-day work, by the official who is their administrative superior. 1.15 Such resistance may be encountered in the transfer of functions, from the national to county governments. Different interpretations of Schedule 4 of the Constitution lead to vastly different functions for county governments, with huge cost implications. While legislation will clarify some of these ambiguities, the process of devolution will show that other ambiguities remain. Many officials, in this context of ambiguity, will seek to protect their authority (partly out of genuine concern for service delivery continuity and quality), and conflicts could arise in the process between these officials and others who feel that their constitutionally mandated responsibilities (and associated funding) are being denied. 1.16 The backdrop to these dynamics is the need for the national government to operate a ‘culture change� and redefine its role vis-a-vis local service providers. With devolution, the Executive at the national level must learn to transition from an implementation role to a policy setting function. At the same time, the Executive may have the tendency towards excessive legislation as a product of distrust, which may lead to resentment and even non-compliance of county governments. 1.17 Overall, there is need to build trust between the national and county governments. A productive dialogue between representatives of both levels of government, will be critical to ensuring the trust that will enable the system to function. The intergovernmental mechanisms are key to this process, and especially the institutions bringing together national and county representatives. At the same time, national government must respect the need for county governments to sort themselves out, based on certain guidelines, in terms of the day-to-day functioning of county government administration and finances. 1.18 Trust and dialogue are critical to proper coordination of activities, and to prevent duplication of functions. With devolution, there is an enormous risk that national and county governments will hire officials whose responsibilities are essentially the same, leading to not only lack of accountability, but also the fiscal risk of an explosion in the wage bill. Again, the intergovernmental mechanisms, and especially representative institutions, are key to preventing such a situation. 3 Eaton et al. (2010). 7 Chapter 1: Kenya’s Devolution in context 1.19 Capacity concerns often drive the effort for re-centralization. This means ensuring administrative and financial functioning at the county level, will be key to both service delivery and the continuity of devolved functions. In Uganda, the push for decentralization lost steam once the government realized that many local governments were not performing. For example, in the education sector, some local governments were wasting substantial resources, which prompted a process of recentralization, thatincluded conditional grants, and reduced fiscal autonomy. Given that national government officials are often hesitant to cede authority to local government, mediocre or poor performance of local governments provides a reasonable argument for these officials to seek re-centralization. 1.20 Second, local power holders resist the presence of a new government that competes for their authority and power, or seek to co-opt the new government. Local governments are not established in a vacuum of power, but rather formed or strengthened in the context of local power structures. These structures can be formal, such as provincial administration, or informal, such as traditional power hierarchies, like village leaders or the landed elite. Once a prominent local government is established, these power holders seek to take over the new structure or undermine its authority. 1.21 Moreover, these local interests are quite varied and complex, even in the formal government structures. Figure 1-2 shows the various government representatives present at the local level in Kenya. Except for rare cases, with the creation of county governments, these various bodies will understandably either seek a prominent position within the county government, or to retain power at the expense of the county government. Figure 1-2: Various government systems at the local level in Kenya. District Sector Local Government NGO/Private Constituency Level system system System sector system Central Office of the Sector Line Ministry of Local National Coordinating National Management Government President Ministries Government Committee Committee (Evolving) Sector Utility / PMU (eg. CDTF) Commission Province Provincial Sector Provincial Provincial Nairobi City Commissioner Offices LG Office Council PMU Prov. Group District District Using District Commissioner Treasury / Planners District Sector District Development Committees Committee District Sector County Mombasa City Offices Councils Council Division Development Sector - facility level Municipal Constituency Division Committee / Officer Committees and committees Councils - Self-help Water supply Officers schemes Town Location Chiefs - Hospital Councils - Health centers Sector - Rural dispensaries Facilities Sub-Location Asst. Chiefs - Secondary schools - Primary schools Community Citizens Community Groups Source: World Bank (2002). 8 Devolution without disruption – Pathways to a successful new Kenya 1.22 The provincial administration, under the Ministry of Provincial Administration, represents a particularly strong set of institutional intereststhat stand to lose out from the establishment of strong county government. In particular, powerful officials within Provincial Administration, including District Commissioners (DCs) and Chiefs, often represent the most prominent official at the local level. Serious questions remain as to the new position of DCs and Chiefs, including their new responsibilities, their organization and their relationship with the county government. The new Constitution does not directly address the problematic role of Provincial Administration vis-a-vis the new county governments, and instead, Article 17 stipulates that the national government has five years to restructure it “to accord with and respect the system of devolved government established under this Constitution.� 1.23 Third, national interests often dominate local interests, contrary to the theory of decentralization. Even with formal administrative and fiscal autonomy, national political interests sometimes dominate local government. In one clear scenario, which is common in Nicaragua, national political parties appoint party favourites in elections for mayor, selecting politicians who will please the national political leaders, and not necessarily respond to their constituents’ needs. Because of the importance of party name in elections, much of the population will vote simply based on party affiliation. As a result, local political representatives often represent national party leader interests, rather than simply the local interests. 1.24 Key to this dynamic is that county governments have not yet been established, meaning that they lack a clear interest group which may represent them vis-a-vis other competing interests. In most decentralization processes, the decentralization process is to a local government that has been functioning previously; as such, the local governments have an established political and administrative leadership, who is eager to receive resources and responsibilities. Kenya is a quite unique case, in that there are few local actors who represent such a direct local interest. 1.25 Similarly, the uncertainty that devolution poses for both national government and local authority staff will surely play an important role in determining the path for devolution. This is a key question for thousands of civil servants, who represent quite powerful positions at the local level. Moreover, the fiscal health of the government can be compromised, if the process leads to heightened levels of recruitment without a commensurate shrinking of the national civil service. In other words, if local authority staff are not hired by county governments, and simply transferred to the national government, the wage bill of the national government will increase substantially. 1.26 Another potential source of conflict is the design of elections for governors. Once elections for governors are held, the plurality voting system presents a further risk. According to the Constitution, a candidate who wins the most votes wins the governorship, with no required threshold. In order words, if there are five candidates, it is possible for a candidate to win with just 21 percent of the vote, which is quite a weak mandate for a governor. Such a selection process could potentially lead to conflict, if the wider public strongly rejects this candidate, or if the governor seeks to channel resources to the relatively small group which voted for him or her. 1.27 Finally, a key question for county governments, is how they will manage relationships between different communities at the local level, an issue that is highly relevant to setting ward boundaries. Some counties will maintain a majority and a minority community, and it is unclear how minority 9 Chapter 1: Kenya’s Devolution in context representation and rights will be ensured. There is a risk that inequality will be devolved to the county level. One possible step for ensuring minority rights, is promoting minority representation in government, either through designated seats or specifically designed wards. Managing such a process may become contentious, however, as numerous local groups seek to maximize their representation. For this reason, the setting of ward boundaries is significant, yet it is a controversial issue that will require a delicate negotiation within Kenya. CHAPTER 1: KEY INSIGHTS / RECOMMENDATIONS. Devolution is at the core of the constitutional vision of a new, more inclusive and equitable government. This follows decades of state capture, real or perceived, by ruling factions and a culture of “all or nothing� that dominated Kenya politics. It was born out of a long divisive history, and at times violent struggle, which will shape to design choices, and dormant conflict may well resurface, if devolution is not prepared carefully. Devolution is particularly difficult to implement―especially on such radical scale―because those in power are the ones who must give up functions and resources. The challenge is compounded in Kenya by the magnitude of the proposed transformation, the very short time frame for design and implementation, and the lack of unity in government to date. Policy-makers will need to walk a fine line between seizing the momentum, to secure far reaching devolution, and avoiding the risk of overcrowding the process, by requiring counties to take on too-much-too-soon. Failure to deliver on the promise of greater equity through devolution―particularly for Kenya’s poor and marginalized communities―could lead to disappointment, and county failure could lead to creeping recentralization. *** It is important to clarify the division of responsibilities between the national and the county level in order to avoid the temptation of ‘claw-back’. Likewise the national government’s role vis-à-vis local service providers, must also be clarified, and it will need to operate a culture change, from implementation to policy setting roles. An intergovernmental coordination mechanism is needed to build trust between the national and county governments, and provide a forum for conflict identification and resolution. The roles and responsibilities of key institutions will need to be redefined, in light of devolution including: the Ministry of Local Government and Provincial Administrations. How devolution will affect relationships between different communities at the local level must be factored-in so that risks can be mitigated early on. 10 CHAP TE R T WO The scale, scope and complexity of Kenya’s devolution 2.1 The scale, scope and complexity of Kenya’s devolution entails major challenges and risks inherent in the transformation process, but creates new opportunities for improving service delivery at local levels, and strengthening overall accountability to communities. This is the most significant public reform effort since independence, and the expectation is that the move will lead to better services and more equity, open up government to public participation and scrutiny, and improve the usage of public funds. Yet, precisely because of its enormity, the reform effort carries significant risks of disruption to existing service delivery mechanisms (however dysfunctional these might be and are in need of reform) and consequently, of possible reversal of the devolution project. Whether successfully implemented or not, the reforms will have unparalleled implications for the functioning of Kenya’s public sector and political governance, including the role of key institutions. They will leave a footprint in the lives of ordinary Kenyans. How the public sector functions today, and where it has failed 2.2 Kenya’s public sector is characterized by parallel mechanisms of services delivery at the local level. There are at least four distinct systems of which central government is the most dominant. Central government systems include an ‘administrative’ district system responsible for law and order, and a ‘de-concentrated’ sector system, through which line ministries provide public services. The district system includes various administrative units (district, division, location and sub-location) grouped under eight provincial administrations. Officials of both the district and sector systems are appointed by central government, and both are funded through the national budget. 2.3 Line ministries directly allocate and disburse funds. These funds received under the central government budget system, are disbursed to ministries’ field departments, to finance annual work plans, projects and programs. The resources reach the community level through various public service delivery units, such as schools, health facilities, water schemes and roads. All payments under the sector system are processed through the office of the district accountant, upon request of the district offices of the line departments. The District Accountant reports directly to the Ministry of Finance. 2.4 Kenya’s public sector, today, is moderately decentralized, through the local government system. Established during colonial times, the Local Government system has three elements, namely Municipalities (Cities and larger towns), Town Councils (small towns) and County Councils (rural authorities). Today, there are one hundred and seventy five Local Authorities (LAs), whose decentralized authority (administrative and legal powers, functions and responsibilities) derives from the Local Government Act Cap 265 (1963), which also stipulates revenue sources to finance the delegated functions. 2.5 The functions delegated to LAs are financed through a variety of local taxes, fees and charges, as well as the Local Authorities Transfer Fund (LATF), introduced in 1999/2000. Today, LATF covers both a 12 Devolution without disruption – Pathways to a successful new Kenya substantial part of LA operational costs, as well as some development expenditure. In addition, some LAs receive resources from donors through various projects. 2.6 Local Authority services are provided through the LA structures for planning, budgeting, implementation, monitoring and accountability. The services mainly include: maintenance of rural access roads, establishment and maintenance of public markets, bus parks and slaughter houses, housing and implementation of social welfare programs, including support to and burial of destitute people. 2.7 In practice, Local Authorities lack the resources to effectively deliver services. In addition, their planning and resource allocation systems are largely inefficient. Most LATF funds are used to finance operational costs. The meager portion of the LATF funds allocated to development activities is fragmented equally to councilors, to support development activities in their respective wards, making the resources ineffective. 2.8 The final service delivery level is the constituency (corresponding to a parliamentary seat). The Constituency system has various types of funds, including the HIV/AIDS fund, the Bursary Fund, the Road Fund, and since 2003/04, the Constituency Development Fund (CDF). These funds are made available at the constituency level, through a system that is a blend of both the government “top-down� system, as well as the “bottom-up� NGO-type system. While the system of funding is uniform across constituencies, with differences being only in the actual amount received by each, the actual service delivery experience varies widely; some constituencies have managed to successfully implement community projects through participatory processes, but others continue to report average or poor results―delayed or stalled projects, which are sometimes overpriced, and misappropriation of funds by local officials. How the public sector will function after devolution 2.9 Kenya’s devolution program is one of the most ambitious in the world, because it transfers a substantial amount of power and resources to an entirely new level of government. At once, Kenya’s eight provinces and over two-hundred and eighty districts will be replaced by forty-seven brand new counties. In many countries, devolution is a process of giving political autonomy to administrative units that are already in place. By contrast, in Kenya, devolution will entail creating new political and administrative units at once. 2.10 Each of the forty seven new county governments will have an elected Assembly, a Governor and an Executive Committee. Both the Executive Committee and the Governor are not members of the Assembly, meaning that there will be full separation between the Legislature and Executive at county level, just as there will be at national level. Counties will also have a voice in the national Parliament through the Senate―a new upper house which mainly comprises of forty-seven directly elected county representatives. 2.11 The new county governments will take over functions, staff and (where applicable) revenue raising powers from existing administrative units namely local authorities and districts. County governments will inherit the responsibility for service delivery in four key sectors (health, agriculture, transport and water). This means that county governments will be responsible for both personnel currently engaged 13 Chapter 2: The Scale, Scope and Complexity of Kenya’s Devolution in these functions, as well as specific Figure 2-1: Kenya’s devolution presents massive challenges for political and administrative restructuring. tax revenues (e.g. liquor licensing) used to fund devolved functions. This transfer process from local authorities and districts to counties, reflects the massive scale of Kenya’s devolution, its unprecedented scope and the intricacies surrounding of its successful implementation, especially in view of the relatively tight constitutional deadlines to complete the takeover. Figure 2-1 illustrates the political and administrative restructuring in Kenya’s devolution. Source: World Bank (2011b). 2.12 Each county government will have its own public service, and will be able to appoint its own public servants, within a “framework of uniform national standards�. Significant numbers of staff who are currently performing devolved functions, are expected to move across from national to county governments. County governments will be responsible for establishing and abolishing offices in their public services; appointing bearers to these offices; and, exercising disciplinary control over individuals occupying them. Through an Act yet to be drafted, Parliament will be expected to homogenize the public services of different counties. 2.13 Kenya’s devolution represents a bold and globally unprecedented experiment in urban management and financing. According to the Urban Areas and Cities Act, the current system of local authorities will cease to exist, and only urban areas with over 250,000 inhabitants will have corporate bodies to manage urban services. Only five urban areas currently meet this threshold. In other urban areas with less than 250,000 residents, the executive responsibilities of local authorities will pass to county governments, although a town committee will be advising it. In both cases, county governments will be responsible for financing urban service delivery. Effectively, the Urban Areas and Cities Act recentralizes urban service delivery for most urban areas; from local authorities to county governments. Implications of the transformation for public administration and service delivery 2.14 Kenya’s public administration and service delivery systems are complicated, mainly because local authorities, districts and constituencies have overlapping jurisdictions as well as functions. In many instances, local authorities, districts and constituencies deliver similar services, making it difficult for citizens to identify their primary service provider, or hold them to account. This problem is heightened in urban areas, where all three agencies have some level of responsibility (such as in road maintenance). The overlaps complicate any attempt to compare spending and performance over time, for the purpose of informing resource allocation, decisions and policy improvements. For districts (de-concentrated national government units) comparison over time is further complicated by their proliferation, from forty at the time of Independence to more than two hundred and eighty today. 2.15 The existing public administration and service delivery systems also have unrelated institutional frameworks, so a clear strategy (and much time) will be needed to blend them. As Local Authorities 14 Devolution without disruption – Pathways to a successful new Kenya disappear on day one after the election (while their functions are transferred to the counties) LA personnel will need to be integrated in the county civil service (although over 90 percent are not currently civil servants). Over time, district staff will also transition over to the county administration. As illustrated in Figure 2-2, on day one, Kakamega county government will be responsible for staff and functions of three local authorities and eleven former districts (in totality for the former which will disappear, and in relation to new division of responsibilities for the latter which will continue to exist in parallel to the counties). A key challenge will be to manage possible redundancy of LA and district staff skills, as they are unified into a single county administration. Clearly, much innovation will be required to come up with a public administration and service delivery system, that retains the best features of the old systems (e.g. strong presence at the local level) while overhauling areas with weaknesses (e.g. poor staff supervision, inadequate focus on outcomes, and rigidity to change). In any case, this process will take time, and will certainly be subject to constant improvements in the future. Figure 2-2: How Kakamega will integrate former districts and Local Authorities in a new county administration. Kakamega County day 1 … Kakamega county government will be responsible for staff and functions of former local authorities, which will no longer exist (Section 134 County Governments Act) . DISTRICTS 90% of these local authority staff are not civil servants Matete Likuyani Lugari Kakamega County today… Mumias Matungu Butere Kakamega Kakamega Khwisero Central North DISTRICTS Kakamega Kakamega Matete Likuyani Lugari East South Mumias Matungu Butere Kakamega Kakamega Khwisero Central North Kakamega Kakamega Kakamega County in the future … East South Eventually some (all?) LOCAL AUTHORITIES district staff will also KAKAMEGA COUNTY ADMINISTRATION be integrated into the county administration Mumias Municipal Council Malava Town Council DISTRICTS Kakamega Municipal Matete Council Likuyani Kakamega Mumias East Khwisero Kakamega Central Matungu Kakamega North Lugari Kakamega Butere South 5 Source: World Bank staff analysis. 15 Chapter 2: The Scale, Scope and Complexity of Kenya’s Devolution 2.16 An important outcome of the public administration transformation process will be to make county governments the hub of organizing services at the local level. Following Kenya’s lengthy history of disjointed public administration and service delivery frameworks, the expectation is that the devolution- induced transformation will: (i) clarify the roles and responsibility of each administrative (and political) unit; and, (ii) align administrative boundaries to locate them within a county boundary, so as to simplify coordination within and between counties. The outcome of this transformation will be to position the county government squarely at the center of public service delivery, at the local level. This will have huge benefits: (i) it will streamline funding; (ii) lead to efficiency (and thus cost savings) in service provision; and, (iii) enable effective monitoring and evaluation not only to track progress, but also to “loop� back into policy improvements. 2.17 Sectoral service delivery arrangements will be considerably redesigned, although majority of the personnel will remain in their present locations. In the health sector, nine out of ten personnel in both the Ministry of Public Health and Sanitation (MoPHS), and the Ministry of Medical Services (MoMS) are currently engaged in activities that will be devolved.1 The activities―mostly district and rural based operations, as well as environmental health services―are currently funded directly from the national budget, but they are operationally delinked from the central government. In this case, the implication of devolution is that financial flows and reporting lines will change, but personnel are unlikely to move, except in cases where inter-county reallocations are deemed necessary for the purpose of improving equity (Figure 2-3). Figure 2-3: In the heath sector, 9 out of 10 personnel will be moved to county governments, even though they are likely to remain in the same pay station. National vs. County Health Personnel; Staff attached to devolved function by main Likely Split Percent budget heads Percent 100% = 49,017 Personnel* 100% = 44,574 District Health Services 40% National Rural Health Services 9% 18% County Environmental Health Services 5% 91% 1% Nutrition Others 36% Source: World Bank staff analysis. * As of January 2012. 1 Using Schedule IV of the Constitution as a guide, all 49,017 personnel from the two health ministries (a total of 49,017) were assigned to either national or county governments according to each personnel member’s: i) district; ii) “pay station�; and, iii) budget sub-head. While most personnel fit neatly into either national or county governments, some do not, and assumptions were made about where to categorize these. 16 Devolution without disruption – Pathways to a successful new Kenya 2.18 It is not only devolved sectors that will be transformed. Sectors with minimal devolution (e.g. education in which only early childhood development becomes a county function) as well as those that will remain national government functions ought to be transformed, so as to align them with the new overall public administration and service delivery framework.2 Failure to achieve this alignment will mean that the existence of parallel (or unsynchronized) public administration and service delivery frameworks, will undermine the success of the entire constitutional implementation process. Implications for governance and the role of Parliament 2.19 The National Assembly will grow sharply, from two hundred and ten members to three hundred and forty nine, excluding the Speaker. The Assembly will comprise of two hundred and ninety Members of Parliament (MPs), and forty seven elected women representatives. There will be twelve additional National Assembly members―six male and six female―nominated by the political parties according to their numerical strength. 2.20 New features of Kenya’s governance sector include the Senate and the forty-seven County Assemblies, the former a re-introduction, the latter for the first time. The Senate, which existed briefly following Independence, will consist of sixty seven members elected directly from the counties, and the Speaker. The Senate will also include: (i) sixteen women elected from the counties; (ii) two youth representatives ―one male and one female; and, (iii) two representatives of people with disabilities (PWDs)―one male and one female. County Assemblies will consist of members elected from the wards, an additional six nominated representatives of marginalized groups, and the Speaker. 2.21 Changes in structure and composition of Parliament are not only dramatic; they have major implications. Expansion of the National Assembly and introduction of the Senate will lead to an increased budget for civic expenses, and for the necessary infrastructure e.g. construction of Senate and County Assembly buildings. Parliament’s additional roles (e.g. vetting of state appointments) might also lead to increased budget and time constraints. 2.22 The changes will transform the budgetary process, and the relations between Parliament and the rest of Government. Under the constitutional changes, Parliament becomes the prime mover of the budget process─additional responsibilities of the National Assembly, as well as those of the Senate will lead to greater public purse powers. The Legislative process will be streamlined (e.g. Bills will now originate from Parliament, not the Executive), but it also becomes more complex (e.g. Bills moving between the two houses will lead to scheduling and time constraints) and generally places greater demands on Parliament. Increased oversight on the work of cabinet secretaries by Parliamentary Committees will mean that State and public officers become directly answerable to Parliament for their actions, not to Ministers. And after transition (three years after next elections), the Attorney-General will not be a Member of Parliament, meaning that Parliament will rely wholly on its own legal service. 2 There is particularly a strong case for restructuring government activities falling under the Governance, Justice, Law and Order sector (GJLOS). In actual fact, some of these activities are currently undertaken by the system of Provincial Administration which, according to the Constitution, should be restructured within a prescribed timeframe, to align it with the new system of county governments. 17 Chapter 2: The Scale, Scope and Complexity of Kenya’s Devolution Conclusions 2.23 The enormous scale, scope and complexity of Kenya’s devolution underscore the challenge, not impossibility of its implementation. At the centre of this implementation cycle is the impending restructuring of the now―inadequate public administration and service delivery mechanisms at the local level. These mechanisms have been in existence for decades―they have deep roots. For this reason, restructuring will not be easy, especially because it involves people, some who will benefit, even thrive, from the status quo. The restructuring also involves the introduction of new value systems, such as the focus on service delivery, the need for closer supervision, and the importance of knowledge, skills and experience. Realizing the desired goals will call for a careful strategy, supported by the necessary policies and legislation―some of these are spelt out in the Constitution, and some are not. Where possible, Kenya will need to learn from the experience of other countries, which have restructured their public administration and service delivery mechanisms. But contexts differ. In Kenya, an additional degree of complexity will be added, because future counties will vary enormously in terms of basic characteristics and needs. CHAPTER 2: KEY INSIGHTS / RECOMMENDATIONS. Kenya’s devolution is dramatic because: (i) it involves transferring substantial powers and resources to entirely new units of government; and, (ii) it implies major changes to political and administrative institutions at the same time. Kenya’s public sector today is highly centralized, with most services delivered through sectoral silos, therefore devolution will imply a major reorganization and ‘culture change’. Establishing counties will involve vertical consolidation of some 175 local administrations, and horizontal consolidation of over 280 districts (within which some twenty or more departments represent different line ministries). Each of the new counties will need to set up institutions from scratch, including an elected assembly and a county executive. At the same time, it will take over functions, staff and revenue raising powers from existing districts and local authorities. Particularly, as local authorities cease to exist, counties will take over the functions of urban management. *** Devolution provides a unique opportunity to rationalize the service delivery framework in Kenya for increased efficiency and accountability, by making counties the hub for organizing services at the local level. But this will require strategies to: o blend institutional frameworks of co-existing public administrations. o clarify the roles and responsibilities of each administration and political unit. o align administrative boundaries to county boundaries. o transition staff currently employed in LAs and districts in a way that avoids redundancies and maintains fiscal discipline. 18 C H A PT E R T H RE E Demographic and geographic diversity of Kenya’s counties 3.1 Understanding the demographic and geographic diversity of Kenya’s future counties will be central to designing the new financing, administrative and service delivery arrangements. Typically decentralization goes either straight to the local/municipal level (to which a discrete set of municipal functions are handed) and/or to larger units (such as provinces) of which there are just a few. The uniqueness of Kenya’s proposed set-up is that counties are in fact in between those two levels. Their mandate is much broader than that of municipal of local authorities, but their reach is smaller than is customary for higher subnational units. Therefore, the challenge will be to devolve substantial powers to a large number of subnational governments. This challenge is compounded by the significant diversity―in terms of endowments, economic texture and social outcomes―across Kenya’s forty-seven new counties, each of which will inherit similar functions at more or less the same time. Understanding this diversity is key, because it is has been a driver of devolution to start with and because it will constitute a major challenge in implementation. Midgets and giants: Wide variations in population, density and urbanization 3. 2 Envisioning the average Kenyan county is not easy given the wide diversities, yet this might help to convey the extent of existing geographic and demographic differences, and the contrasts that underlie these differences. The average county has eight hundred and twenty thousand people, and a total land mass of twelve thousand three hundred kilometers squared. Thus, the average county is fairly sparsely populated with only four hundred people per kilometer squared. Slightly less than half of the population in the average county would be income poor, while one third would be urban residents. The average county has a moderately high level of immunization among children below the age of one (60 percent), and a moderately low malaria burden (30 percent). Out of every ten people, seven have primary schooling, yet only one has secondary education. In the average county, there is better access to improved sanitation facilities (80 percent) than to improved water source (64 percent); and, only 10 percent of the road network is in good/fair condition, even though only 10 percent of all roads are paved. Typically, less than one fifth of the entire population would be having access to electricity. 3.3 In reality, future counties will vary widely in terms of population and density. The ten most populous counties with 40 percent of the country’s population occupy approximately one tenth of the entire land area. Conversely, three quarters of Kenya’s land area is home to only 15 percent of the country’s population. Nairobi and Mombasa, the most inhabited counties have densities of more than four thousand people per square kilometer, which is more than seventy times the average density for the entire country. In contrast, the North-Eastern frontier, representing some nine districts or so, has less than twenty persons per square kilometer, on average. Both population and density dimensions are important, since they have historically influenced geographic distribution of the Local Authority Transfer Fund (LATF). The two variables are also major drivers of service delivery needs and cost differentials. 20 Devolution without disruption – Pathways to a successful new Kenya 3.4 Moreover Kenya is witnessing fundamental demographic changes of which rapid urbanization is a central feature. While Kenya is urbanizing rapidly―much faster than the overall population growth rate―most counties are still predominantly rural. Only five counties are more than 50 percent urban, and a majority of counties have at least 80 percent of their populations residing in rural areas. This is important from a political economy point of view, because in Kenya’s future County Assemblies, most members will be representing rural constituencies: the challenge will be to protect service delivery in urban areas, which traditionally generate the bulk of local revenues. 3.5 Demographic diversity has significant implications for function assignment and the delivery of services. For instance, it is unreasonable to expect that all counties will be able to deliver the same mix of services, at the same level. For instance, X-ray services fall within functions that will be devolved, but only eight counties currently have the necessary personnel (radiographers, radiologists, film processors and radiation protection officers) and equipment to undertake the service. The reasons why the other thirty-nine counties do not offer X-ray services, include prohibitive equipment purchase, installation and maintenance costs; and, the general shortage of trained personnel countrywide. Against such a backdrop, does the mere fact of devolution suggest that Lamu county, which has the smallest population, should begin to offer X-ray services immediately, as health functions are transferred to it? Probably not, because: (i) the function would comprise a disproportionate chunk of Lamu county’s health budget; (ii) Lamu has relatively small land area, population/density (so X-ray services might be underutilized); and, (iii) its population is relatively poor, so majority of the residents may be unable to share the full service costs through user fees (so poor cost recovery). 3.6 Significant diversity between counties also makes it difficult to find a single formula to allocate resources, and hence, to achieve fiscal equalization. Since existing service levels (e.g. number of facilities, staff and operating budgets) are very different, counties will inherit dissimilar service delivery ‘obligations’ (mostly recurrent spending). The disparity in inherited costs is part of the legacy of uneven service coverage, but also a reflection of geographic variations, which drive cost differentials─the remoteness of Turkana and its high temperatures mean that for a successful child immunization campaign, vaccines must be airlifted in specially refrigerated containers, all at an exorbitant cost. So, finding a single formula that captures disparities in expenditure needs across Kenya’s counties presents a major challenge. In fact, the challenge of achieving fiscal equalization in the midst of massive disparities has already been encountered by the CRA. In its recently-launched preliminary formula for the horizontal sharing of the prescribed minimum of national revenue, CRA introduce special bands around the land area variable, so as to achieve “affirmative action� in favor of very small counties like Lamu.1 Poverty disparities, service inequities and social outcomes 3.7 As counties are set up and begin to deliver services, they will face dramatically different starting points in terms of poverty, service gaps and development potential. The richest and the poorest counties (Kajiado and Turkana, respectively) have an 80 percentage point gap in terms of poverty prevalence (in other words almost all households fall below the poverty line in the latter while very few do in the former). Like population, the distribution of poverty has historically influenced allocation of decentralized funds, notably the Constituency Development Fund (CDF), and it will have major implications for the design of fiscal transfers to counties after devolution. As it were, CRA’s preliminary 1 In the calculation of the land area variable (which carries a 6 percent weight), the contribution of very large counties to Kenya’s total land area is suppressed to a maximum value of 10 percent. On the other hand, the contribution of very small counties is amplified to a minimum of 1 percent. 21 Chapter 3: Demographic and Geographic Diversity of Kenya’s Counties formula for the horizontal sharing national revenue has populationas its most important variable, with a weight of 60 percent. Poverty is third, with a weight of 20 percent. 3.8 Inequalities within counties are also Figure 3-1: There are signi�cant inequities within counties. quite pronounced, and may in fact prove to be devolution’s “Achilles’ Heel� if not Kakamega addressed. The most significant intra- Marakwet Homa Bay county well-being inequities exist between Embu rural and urban locations. According to the Kili most recent assessment, overall poverty is Busia not only more prevalent in rural areas; it is Marsabit also declining more sluggishly.2 For example, Migori within Migori county, the largely-rural Kuria 0 20 40 60 80 district has a poverty index that is nearly Poverty index (%) Richest district Poorest district twice that of Migori district, which is mostly Source: World Bank analysis of KIHBS (2005/06). urbans (see Figure 3-1). Also, the poorest district in Kakamega county (relatively rich) has more or less the same poverty index as Figure 3-2: Poverty is as dispersed between counties as population density the richest district in Kilifi county (relatively 100 Poverty index (Percent, 2005/06) poor), whose population is more rural. 90 80 There are several important implications. 70 First, despite expectations, devolution may 60 50 not result in decreased inequality in Kenya. 40 While resources will be more equally shared 30 y = -0.036x + 60.64 across counties, nothing guarantees that 20 R² = 0.147 10 they will be better spent (from an equity 0 standpoint) at the county level. Second, 0 100 200 300 400 500 600 700 Population per km2 (2009) a radical shift in resource allocation from Source: World Bank analysis, based on KNBS data. urban areas to rural districts (in line with the *Nairobi, Mombasa and Vihiga counties excluded from the scatter due to their outlier densities. new county political reality) may leave the country worse off, by slowing the growth Figure 3-3: There is a strong correlation between poverty potential of cities. and social outcomes, particularly education. 25 Population with secondary education 3.9 Poverty variations have a strong connection 20 to disparities in settlement patterns, and to (percent, 2009) differences in infrastructure coverage and 15 social outcomes. Ten counties with the 10 highest concentration of poverty also have the lowest combined average population 5 y = - 0.109x + 17.55 density─thirty people per square kilometer, R² = 0.341 0 compared with the national average of four 0 20 40 60 80 100 hundred. If the three counties with “outlier� Poverty index (%, 2005/06) Source: World Bank staff analysis, based on KNBS data and GoK (2006). densities3 (Nairobi, Mombasa and Vihiga) 2 According to the Basic Report on Well-being in Kenya (GoK, 2007), which is based on the Kenya Integrated Household Budget Survey, 2005/06… 3 Meaning that they exhibit population densities that are so much above the norm that including them in the sample would bias the results. 22 Devolution without disruption – Pathways to a successful new Kenya are excluded, density disparities have the same dispersion across counties as poverty inequalities. This signals the potentially huge influence that population concentration has on economic activity (and vice versa), including local revenue raising ability. In terms of social outcomes, while the variation in ‘population with secondary education’ is lower than that of poverty, the two variables are also clearly correlated (see Figure 3-3). But infrastructure coverage has by far the closest correlation with poverty─access to improved water and sanitation, as well as the maintenance condition of roads (i.e. whether good, fair or bad) have more or less the same coefficient of variation as does poverty rate. This suggests that that inter-county differences in water, sanitation and “road conditions� are comparable to poverty differences. Poverty is correlated to low population density, inadequate access to service infrastructure, and poor social outcomes. It is precisely these disadvantages that the new financing arrangements aim to overcome. This means that poverty―with all the caveats surrounding the accuracy of its (and population) estimation―provides a good proxy for infrastructure gaps. 3.10 Poverty is a common problem for most counties, and reducing it will remain a shared policy goal, but the strategy to achieving this goal will vary. While each county will have a scope to improve the well- being of its citizens―some more than others―the significant diversity across the forty-seven counties means that no single policy strategy can produce similar results in each county, and certainly not at the same pace. Particularly, in view of the very predictable resource constraints in the early years of devolution, each county will need to choose from a menu of practical and affordable policy strategies. For each county, the best poverty-reducing strategy mix will be the one that matches with its geographic, demographic and “economic texture� milieu. Table 3-1 illustrates possible policy objectives in fifteen selected counties, and strategies that might be selected to realize the objectives. The policy objectives Table 3-1: Illustrating potential differences in policy objectives and strategies among the counties. Selected ▪ Kitui ▪ Wajir ▪ Kajiado counties ▪ Nandi ▪ Mandera ▪ Garissa ▪ Narok ▪ Turkana ▪ Tana River ▪ Lamu ▪ Kilifi ▪ Mombasa ▪ Tharaka-Nithi ▪ Marsabit ▪ Laikipia SECTOR: Policy WATER: Improve the proportion HEALTH: Promote safe deliveries EDUCATION: Enhance school objective. of households with access to of babies, preferably in a health enrolment and completion rates, adequate source of water. centre with the supervision of especially among the 15 – 18 qualified medical personnel. year olds. Strategy (mix) ▪ Subsidize cost of water tanks to ▪ Dedicate some of Equalization ▪ Liaise with CSOs and public examples. promote water harvesting and Fund allocation to top up organizations (e.g. schools) minimize run-off in rural areas. extraneous duty allowances for to establish: i) internship ▪ Sinking of boreholes and wells. nurses and mid-wives to attract programs; ii) work-placement and retain them. programs; and iii) (peer-to- ▪ Initiate mobile health clinics. peer) mentorship program for ▪ Earmark seed money to secondary school leavers. finance a health insurance scheme providing ante- and post-natal services to poor mothers. Source: Author’s construction based on World Bank’s Human Opportunity Index computations, itself based on Kenya Integrated Household Budget Survey, 2005/06. 23 Chapter 3: Demographic and Geographic Diversity of Kenya’s Counties are based on service-access indices, displaying pervasive levels of inequity between the counties― counties selected under each objective appear at the bottom respective indices. 3.11 Poverty is one dimension of pervasive geographic Figure 3-5: Kenya’s spatially unbalanced growth is due to its geographic diversity and inequity. inequities, whose acute perceptions by Kenyans has been at the heart of the devolution momentum. Kenya’s economic development has been concentrated along a narrow corridor, leaving wide swathes of the country behind in terms of economic activity and employment. A few urban areas generate the bulk of Kenya’s total production (see Figure 3-5) and wealth created by Kenyans has not been adequately redeployed through public service delivery, to promote equal opportunities for all. Source: World Bank (2011b). 3.12 The country’s experience with spatially unbalanced growth is not unique, but redistribution through services will be key. Initial natural endowments (like fertile soils) and location factors (like proximity to ports) are reinforced by migration of people to areas where they can make a living more easily―a process known as agglomeration. Firms tend to cluster where there are other businesses that supply inputs or buy their products or services. These agglomeration and clustering effects explain to a large extent why some regions have developed faster than others and remain more dynamic; why cities, as opposed to rural areas will increasingly be driving growth, and why only a few among them have the potential to become major industrial and service hubs. In China, three coastal provinces accounted for over 50 percent of the country’s Gross Domestic Product (GDP) in 2005. 3.13 The main challenge going forward will be to target the appropriate measure of inequality. Kenya’s new county leaders will need to be realistic about the extent to which government intervention in the economy can overcome these historical and geographic legacies. Seeking to geographically rebalance growth and employment would be wasteful and self defeating. However clearly much more needs to be done to make sure that national wealth (wherever it is generated) is redistributed through services to equalize opportunities. Dealing with patterns of Marginalization 3.14 Geographic and economic history has a special significance in Kenya, because it is overlain with cultural differences and an acute perception that areas of the country have been deliberately marginalized. While parts of the country are endowed with natural resources, or are fertile and therefore agriculturally productive, other regions have no natural resources, are arid or semi-arid. Differences such as these―in sheer economic potential―have underpinned the country’s historical inequities, with some regions attracting more labour and capital, and generating higher revenues. 3.15 However, marginalization is particularly hard felt, because historically social services have also been unevenly allocated. Access to services plays a significant role in determining opportunities available to individuals and their welfare, and yet the data on education and health, water and sanitation, access 24 Devolution without disruption – Pathways to a successful new Kenya to electricity and the road network, all tell the story of historical inequities (see Figure 3-6). These inequities also signify the extent of the failure of the state to equalize opportunities for all Kenyans. 3.16 Variations in access to services reflect historical patterns of marginalization. Kenya’s North-Eastern areas―which are the least developed economically―have failed to receive the level of central government attention, and support required to equalize access to services (arguably the cost of delivering services there, and the needs will also have been significantly higher). For instance, while malnutrition is relatively low in Mombasa, where 84 percent of children are of adequate height for their age, in Wajir County only 21 percent of children are of adequate height for their age, implying very high levels of chronic malnutrition. Overall, there is close correspondence between low population density and low social outcomes, and economic activity as well. Hence, the only way that these inequities can be addressed is through redistribution, and the design of adequate fiscal arrangements. 3.17 There remains a strong perception among Kenya’s public that the central government simply can’t be trusted to redistribute resources ‘fairly’. This explains the hostility, among devolution partisans,to Sessional Paper number 10 of 1965,4 which argued for investing in those areas where the largest returns could be achieved, while using the resulting wealth to equalize opportunities―what the World Development Report (WDR) 2009 calls targeting people rather than places).5 To many Kenyans, the Paper was simply used to provide a theoretical justification for a de facto policy of targeted investment (in the most developed areas) and of ignoring the redistribution requirement altogether. 3.18 While an uneven geographic spread of economic development is inevitable, an uneven spread of the social benefits of economic growth (better services) is not acceptable. International experience supports the spatial concentration of economic activity, but the social benefits of the resultant growth and development must to the extent possible, be evenly distributed throughout the country. In this sense, development can be inclusive if the state redistributes national income through investments and services. While all areas will not have the potential to become centers of economic development, all Kenyans should be entitled to the same level of basic services and to equal opportunities, in order to lead a healthy productive life. 3.19 Devolution will not immediately address entrenched inequity across and within counties. Some counties will start at a relative disadvantage and it will take time to build up their capacity and ability to use resources well. The paradox is that counties that stand to benefit the most from devolution in theory, because they were neglected under the old dispensation, may be the least equipped in practice to make efficient and transparent use of their resources, and retain the skilled staff that is essential to making services work. This means that dramatic redistribution will not occur overnight: it will need to be phased in gradually. 4 GoK (1965). 5 World Bank (2009). 25 Chapter 3: Demographic and Geographic Diversity of Kenya’s Counties Figure 3-6: The uneven pattern of Kenya’s infrastructure development. Source: World Bank (2011a). 26 Devolution without disruption – Pathways to a successful new Kenya CHAPTER 3: KEY INSIGHTS / RECOMMENDATIONS. Equity is a core objective of the new Constitution and devolution is seen as the main vehicle to achieve equalization. But Kenya’s future counties are at dramatically different points (and possibly potential) in terms of overall development, economic infrastructure and access to social services. The architecture of devolved government will need to accommodate such diversity among Kenya’s future counties. Disadvantage in Kenya is strongly correlated with remoteness which is in turn related to historical patterns of marginalization. Diversity across counties is particularly relevant because of: (i) the large number (forty-seven) of these subnational units; and, (ii) the far reaching transfer of functions envisaged to entirely new levels of government. This diversity will pose specific challenges for: o function assignment, as a ‘one size fits all’ approach may not be realistic. o designing the intergovernmental fiscal equalization scheme, as: (i) a simple formula may not capture needs and costs accurately for all county governments; and, (ii) equalization must be balanced with fiscal constraints. o capacity building to avoid that some counties fail to deliver services or to put in place systems for sound management. *** Expectations that equalization can be achieved rapidly need to be managed to avoid disappointment and overloading of weak counties. Specific capacity building programs are needed for ‘lagging’ regions. Equalization policies should be focused first and foremost on equalizing opportunities to access basic social services. 27 PART II Equity between levels of government CHAP TE R FO U R Why it all starts with function assignment 4.1 Function assignment is the starting point for establishing a clear, accountable multi-level government system. The function assignment process allocates distinct and discrete functions and activities to separate levels of the government.¹ A lack of clarity on expenditure responsibilities of different levels of government, results in a lack of transparency and accountability, that in turn breeds confusion and conflict. A clear function assignment framework is also key to determining the funding requirements at each level of government, and determining how resources, including staff and structures should be deployed to the county level. This is difficult to achieve as the process itself is complex, and involves a multitude of stakeholders, but it is key that it results in a clean, clear list of all current functions (over time), listed against the level of government in charge of carrying them out. 4.2 Citizens should know which level of government to look to for service delivery. At the heart of devolution is the expectation that the government will be brought closer to the people, and become more responsive and accountable. For citizens to hold local governments accountable, they must know which level of government they can expect to receive different services from. Decisions on the distribution of function assignments should be published and disseminated, so that citizens can access them easily. By the same token, governments need to be clear about what roles they are expected to perform in the roll out of different functions, i.e. the budgeting and delivery of service. This is key for coordination between the two levels of government as they roll-out different components of the same function. 4.3 A fundamental principle of fiscal decentralisation is that funding should follow functions. Counties need sufficient funding to carry out the functions devolved to them, and to begin to address the most noticeable service gaps. In Kenya’s case, this means matching the revenues of county governments from all sources―the minimum 15 percent equitable share, own revenues and other transfers. But determining funding needs can only be done on the basis of clearly assigned roles and responsibilities for service delivery, and a roadmap to phase-out these functions over time. 4.4 Failure to clarify function assignment creates a risk that one level of government will have too little funding, and the other too much. Vertical fiscal imbalance arises when funding is not distributed in line with functions. A common failing at the start of decentralisation is to pay too much attention to funding and overlook the need to ensure assigned functions and funding match. In a number of cases, this has led to the centre giving away funding, but being left with functions that it can no longer afford to perform adequately. On the other hand, if subnational governments are given functions but insufficient funding, it is likely that service delivery will get worse rather than improve. International experience also suggests that when functions are not assigned clearly, they run the risk of not being performed ¹ Function assignment is referred to in the fiscal decentralization literature as “expenditure assignment�. 30 Devolution without disruption – Pathways to a successful new Kenya at all: if everyone is accountable―no one is accountable. In all these cases, where decentralisation is seen to under-perform, support for the system evaporates, and the center is often tempted to legislate towards recentralization (see Box 4-1). Box 4-1: Recentralization: The Case of Uganda. Recentralization can take multiple forms across countries and within individual countries. One extreme involves formal de jure retraction or alteration of robust/detailed decentralization policies and laws. In more typical cases, often where the decentralization framework is relatively general, recentralization can occur through new laws, regulations and decrees specifically intended to elaborate, modify or dilute the formal provisions of the original enabling legislation. In some countries, this may appear more like a slowing of what seemed to be a decisive reform than overt recentralization. The factors that motivate recentralization vary considerably in their nature and in their impact. Recentralization could simply be the result of a lack of political consensus to adopt or implement decentralization in the first place. Recentralization can also be motivated by awakening central agencies that did not pay much attention when the decentralization framework was being formulated, but later became concerned when they understood the implications for them. In some cases, recentralization is a reaction to the weak performance of subnational governments empowered under decentralization, and the emergence of a sense that these reforms had gone too far too quickly, a position that appeals to centralists or actors whose power was curtailed by decentralization reforms. Clearly this has been the case in Uganda. Uganda received considerable attention as a developing country that embraced decentralization with gusto. There was an emphasis on citizen engagement and capacity building as key drivers of economic development. By the late 1990’s a strong legal framework for decentralization was in pace, and local governments quickly became among the most empowered and best financed in Africa. The nature, pace and trajectory of reform was, however, too ambitious to take deep root, and there was probably too much emphasis on formal system development, and not enough on building local accountability and enhanced governance. After several years, central agencies grasped decentralization implications for their control over resources and some acted to protect their territory. Recentralization policies such as increasingly conditional grants and budget restrictions which undermined the incentives and ability of local governments to improve performance. Later, more politically driven policies such as extensive creation of new districts, recentralization of senior local government staff recruitment and establishment of regional tiers reinforced earlier steps to reduce local powers. Although Ugandan local governments have clearly made progress, they have also been weakened in recent years. There are no incentives evident for the government to step back from their over-reaction to problematic early performance of local governments burdened with initial unreasonable expectations. Source: USAID Uganda Desk Study (2010). 4.5 Function assignments need to be clear, but this is not as straightforward as it may seem. Even though some direction is given in the Fourth Schedule of the Constitution (on devolved government), there are complications associated with determining what functions fall under the prescribed definitions. Compounding the issue is the fact that functions can consist of a variety of activities, each of which could be assigned to different levels of government. It is important that all responsibilities be assigned to the appropriate level of government, using a detailed and transparent framework of allocation. Given that this will also be an on-going process of gradual transfer of functions, it will involve a continuous process of consultation and reallocation. 31 Chapter 4: Why it all Starts with Function Assignment 4.6 The first order of business for the newly established TA should be to initiate the function assignment process by creating a framework of assignment to guide the process. With Kenya’s elections slated to be held on March 4, 2013, the window of opportunity for a thorough and transparent function assignment process to start is fast closing; the added constraint being that Parliament will be dissolved 60 days before the elections. Given the short time frame, and in order to be effective, the TA should use the resources and work through institutions such as the CRA, Kenya Institute of for Public Policy Research and Analysis (KIPPRA) and the now dissolved Task Force on Devolved Government (TFDG), to start the intensive work of assigning functions, in conjunction with the central and line ministries.² It is likely that the function assignment process will not be completed before March 2013, and it is imperative that the TA prepare for this outcome by managing expectations, possibly through a phased transfer of functions. Options for a phased transfer of functions will be discussed later in this chapter. Designing a framework for function assignment 4.7 Kenya’s devolution follows international practice when it comes to the intergovernmental division of responsibilities. The functions assigned to the national level relate mainly to policy and standard setting, and the provision of public goods such as national security and macroeconomic policy. On the other hand, county level functions are mainly focused on policy implementation and local service delivery such as primary health care, urban service delivery, trade licensing and crop husbandry (see Table 4-1). There is one notable exception to international practice in the Kenyan case. Basic education service delivery is not devolved. The overwhelming practice in other countries is to devolve at least primary education. In Kenya only pre-primary has been devolved to counties. Table 4-1: Broad county, national and unassigned functions. County National Not assigned/unclear assignment Agriculture including Crop and Animal Tertiary/Research/Secondary and Social Protection and Welfare; Husbandry, Plant and Animal Disease Primary Schools/Sports Education/ Gender Policy and Programmes; HIV Control, Fisheries; County Health Special Education Institutions; Foreign Programmes; Provincial Hospitals; Services including County Health Affairs; Immigration and Citizenship; Sectoral Training Institutions (Health, Facilities and Pharmacies; Promotion International Waters; Protection of the Agriculture); Births, Deaths and of Primary Health Care; Veterinary Environment and Natural Resources; Marriages. Services (excluding policy); Cultural Disaster Management, National Activities; County Transport including Defence; Police Services; Courts; County Roads; Animal Control and National Economic Policy; Monetary Welfare; Trade Development; County Policy; National Statistics; Intellectual Planning including Statistics; Pre- Property Rights; Labour Standards; primary Education; Implementation of Consumer Protection; Transport and National Policies; County Public Works; Communications including National Disaster Management; Control of Trunk Roads; National Public Works; Drugs and Pornography; Coordination National Referral Health Facilities; with local communities. National Elections; Health Policy; Agricultural Policy; Housing Policy; Veterinary Policy; Energy Policy; Tourism Policy; County Capacity Building and Technical Assistance; Public Investment; National Gambling. Source: World Bank staff analysis. 2 In this instance central government refers to non-sectoral agencies including the Office of the President, Office of the Prime Minister, Ministries of Public Service, Justice, Planning, Finance, and Foreign Affairs. 32 Devolution without disruption – Pathways to a successful new Kenya 4.8 County governments can acquire functions in two different ways. Kenya’s constitution assigns both law-making powers and executive (implementation) functions to both national and county governments, as set out in the Fourth Schedule. Under the Sixth Schedule, the transfer of these functions must be completed by the end of the three-year transition period. The following discussion will refer to these as ‘assigned functions’ or ‘devolved functions’. The constitution also allows a function assigned to one level of government to be delegated to another, but the responsibility for its implementation still remains with the original level of government.3 Primary and secondary education are examples of functions that have been assigned to the national government. This means that all staff and programs associated with primary and secondary education will continue to be administered by the national government. However, this may not be efficient in practice. For example, in order to build a school, the national government would need to determine the school capacity, draw plans for its infrastructure and locate it in the area where it is needed most. This is information that counties may have more readily and accurately than the national government. For this reason, education can be a good candidate for transfer to the county governments. The Intergovernmental Relations Act (IRA) outlines a process of transfer of functions between levels of government, and ensures in section 27(2) that “any transfer or delegation of powers, functions or competencies shall be accompanied by adequate resources to ensure minimum service standards are achieved.� 4.9 The Fourth Schedule of Kenya’s Constitution assigns functions across the national and the counties in broad terms only. In order to understand what resources should be allocated to perform these functions, they need to be disaggregated, or ‘unbundled’ down to the detailed activities that comprise them, so they can then be assigned to the appropriate level of government. Take for example the function of water supply and protection. The national government is responsible for water protection including residual waters and dams, but counties are responsible for water and sanitation services. In Kenya today, the eight water boards that manage bulk reticulation infrastructure like pipelines, also undertake management of dams. Depending on how these different activities are characterised, mixing national and county functions in this case could be considered. It will be necessary to clarify what exact water services will be under the domain of county governments, and what will be under the national government. Much more work remains to be done in order to create a stable and unambiguous framework of function assignment. 4.10 Unassigned functions add a degree of complexity to the function assignment process, because there are multiple ways one can interpret their assignment. A number of functions are not mentioned in the Fourth Schedule. Under the Constitution, these unassigned or ‘residual’ functions are left to the national government (see Box 4-2). Even within concurrent areas, the terminology used in defining the functions that are assigned to counties suggests that the list is not exhaustive (see Box 4-3). A full analysis of functions is necessary to make a clear and final decision on what specific activities and spending responsibilities, the counties will be responsible for. 3 Article 187 Constitution of Kenya (2010). 33 Chapter 4: Why it all Starts with Function Assignment Box 4-2: What the Constitution says about Function Assignment. Concurrent and un-assigned functions. Article 186 (1) Except as otherwise provided by this Constitution, the functions and powers of the national government and the county governments, respectively, are as set out in the Fourth Schedule. (2) A function or power that is conferred on more than one level of government is a function or power within the concurrent jurisdiction of each of those levels of government. (3) A function or power not assigned by this Constitution or national legislation to a county is a function or power of the national government. (4) For greater certainty, Parliament may legislate for the Republic on any matter. **** The option to transfer or delegate additional functions. Article 187 (1) A function or power of government at one level may be transferred to a government at the other level by agreement between the governments if— (a) the function or power would be more effectively performed or exercised by the receiving government; and (b) the transfer of the function or power is not prohibited by the legislation under which it is to be performed or exercised. (2) If a function or power is transferred from a government at one level to a government at the other level— (a) arrangements shall be put in place to ensure that there sources necessary for the performance of the function or exercise of the power are transferred; and (b) constitutional responsibility for the performance of the function or exercise of the power shall remain with the government to which it is assigned by the Fourth Schedule. **** What happens if a national law conflicts with a county law? Article 191 (1) This Article applies to conflicts between national and county legislation in respect of matters falling within the concurrent jurisdiction of both levels of government. (2) National legislation prevails over county legislation if— (a) the national legislation applies uniformly throughout Kenya and any of the conditions specified in clause (3) is satisfied; or (b) the national legislation is aimed at preventing unreasonable action by a county that— (3) The following are the conditions referred to in clause (2) (a) the national legislation provides for a matter that cannot be regulated effectively by legislation enacted by the individual counties; (b) the national legislation provides for a matter that, to be dealt with effectively, requires uniformity across the nation, and the national legislation provides that uniformity by establishing— (i) norms and standards; or (ii) national policies; (c) the national legislation is necessary for— (i) the maintenance of national security; (ii) the maintenance of economic unity; (iii) the protection of the common market in respect of the mobility of goods, services, capital and labour; (iv) the promotion of economic activities across county boundaries; (v) the promotion of equal opportunity or equal access to government services; or (vi) the protection of the environment. Source: Constitution of Kenya (2010). 34 Devolution without disruption – Pathways to a successful new Kenya Box 4-2: What the Constitution says about Function Assignment (continued). (4) County legislation prevails over national legislation if neither of the circumstances contemplated in clause (2) apply. (5) In considering an apparent conflict between legislation of different levels of government, a court shall prefer a reasonable interpretation of the legislation that avoids a conflict to an alternative interpretation that results in conflict. (6) A decision by a court that a provision of legislation of one level of government prevails over a provision of legislation of another level of government does not invalidate the other provision, but the other provision is inoperative to the extent of the inconsistency. Source: Constitution of Kenya (2010). Box 4-3: Even within concurrent subject areas, not all functions have been assigned. Schedule 4 of the Constitution allocates health functions as follows: National Functions County Functions 28. Health policy. 2. County health services, including, in particular— 23. National referral health facilities. (a) County health facilities and pharmacies. 30. Veterinary policy. (b) Ambulance services. (c) Promotion of primary health care. (d) Licensing and control of undertakings that sell food to the public. (e) Veterinary services (excluding regulation of the profession). (f) Cemeteries, funeral parlours and crematoria. (g) Refuse removal, refuse dumps and solid waste disposal. This division of functions leaves some important health functions not assigned, including medical research, training of health personnel, and regulation of professional standards in the health sector. By virtue of Article 186(3), a function that is not assigned by the Constitution is a national function. Several issues need to be addressed to finalize function assignment in health: (1) Clarifying assignment of unspecified health functions. • Are provincial hospitals residual health functions or are they under county health services? • Are training institutions in health an education function (national) or part of county health services? • Is HIV considered a residual health function, or part of county promotion of primary health care? (2) Issues to be considered in deciding function assignments in health. • How will donor grants to rural health facilities (county function) be managed? • Under the new pull system of drug supply to government health facilities, where counties determine their drug requirements, who will manage drug procurement? • Who will manage the procurement of specialized health materials for public health programs like malaria and family planning? • How will Health Information Systems be managed? • How will environmental health be managed? • Are Provincial Health Service functions partially or fully county? • How will the funding for additional staff in rural health facilities employed under the Economic Stimulus Programme be costed? Source: World Bank staff analysis. 35 Chapter 4: Why it all Starts with Function Assignment Clarifying Function Assignments 4.11 The way to resolve these outstanding questions is through a well-structured, inclusive process of assigning functions according to a common set of principles. A process of clarifying function assignments has two dimensions: (i) the analytical process of unbundling functions to their constituent components (activities); and, (ii) the principles or criteria that guide decisions about where to allocate activities when the Constitution is not clear about which level of government is responsible. The consultative process through which these decisions are made is also important. Unbundling activities 4.12 Functions that have both policy-making and implementation components should be unbundled and assigned to the appropriate levels. Functions are made up of a variety of activities, some of which are policy oriented and others concerned with implementation. These activities, or “sub-functions�, should each be assigned to the appropriate level of government. Figure 4-1 below uses the example of the National Agriculture Extension Program, to illustrate how unbundling works. Agriculture (excluding Agricultural policy) is defined in the Constitution as a county function, but when the program is disaggregated to its component activities, it is obvious that some are policy-oriented, for example the identification of priorities for extension support and the research and development of learning methodologies. Many other functions have these characteristics, and so unbundling is a useful tool across the whole function assignment process. Figure 4-1: Unbundling ensures there is clarity about who funds what. National Agriculture Extension Program Components National (policy) County (implementation) Identify priorities for extension support X Research and development of farming techniques X Print resource materials ?? ?? Employ extension officers UNBUNDLING Train extension officers (run training colleges) ?? ?? Purchase seed and tools for demonstration farms X Provide buildings for farmer training X Extension officers travel X Source: World Bank staff analysis. Five key principles to guide the process 4.13 Once the components of each function are disaggregated through unbundling, a decision must be made about which level of government is responsible for each. Line ministries should be very involved in these decisions, since they understand best what is involved in carrying out a particular activity, but they should do so according to common principles. The following five principles are based on lessons from international experience of government-wide function assignment processes (summarised in Table 4-2). 36 Devolution without disruption – Pathways to a successful new Kenya Table 4-2: Potential function assignment criteria for the Transition Authority to consider. DEVOLVE RETAIN NATIONAL DEPENDS If speedy decisions or coordination Better economies of scale. Accountability based on how needed representative the level of Administrative capacity efficiency and government is effectiveness for larger regions Incidence of wide intense stakeholder Security considerations like national Where it does not make sense to communication and involvement defense and police delink functions that are closely Spill over effects that might lead to related sub-optimal investment and provision Cross-sectoral integration for services Involves technically demanding skills where sophisticated technology may require central management Source: Adapted from Ferrazzi and Rohdewohld (2009). 4.14 Principle 1: Subsidiarity dictates that a public service function should be assigned to the lowest level of government that is capable of delivering it. The principle of subsidiarity stems from the idea that decentralization benefits citizens by giving them greater influence over decisions affecting the services they receive. The Constitution supports the application of this principle. Article 174 includes the objective of giving the power of self-government to the people, and enhancing their participation in making decisions (clause c), and recognizing the right of communities to manage their own affairs (clause d). Article 176(b) states that county governments should decentralize their functions and service delivery, to the extent that it is efficient and practicable. 4.15 Principle 2: As far as possible, capital and recurrent expenditure responsibilities should be aligned. If one level of government is responsible for capital development (such as the construction of new health facilities) but another is responsible for recurrent functions (such as the hiring of health staff), a mismatch could result. When one level of government creates a liability that is ultimately borne by another, without prior coordination and agreement (and therefore adequate budgetary appropriation), facilities may end up without the staff, and recurrent funding to make them operational. In Kenya, the experience of the Constituency Development Fund (CDF)4 is a case in point. In some cases CDF has financed the construction of facilities, but line ministries have not provided budget for staff and operating costs, resulting in unused and essentially wasted resources. That said, handing over development spending responsibility to counties could exacerbate capacity constraints at the local level in some counties. If counties do not have the requisite capacity to run complicated construction projects, they are unlikely to spend development funds effectively. To address these limitations, responsibility for major and complex projects might be retained at the national level, under the Public Investment function, and funding for smaller projects provided as conditional grants, so that its application to capital investment can be assured. 4.16 Principle 3: Services that involve large economies of scale may be delivered more efficiently by a larger unit of government. An example currently being debated in the health sector concerns drug procurement. Drug procurement involves significant economies of scale, because purchasing on a large scale from overseas suppliers is many times cheaper than purchasing from commercial pharmacies. 4 A program of grants for minor works projects that is allocated equally to constituencies, and spent according to projects selected by a committee chaired by the member of Parliament for that constituency. 37 Chapter 4: Why it all Starts with Function Assignment Technical skills are also needed to ensure the drugs are of adequate quality. It may simply not make sense for counties to undertake forty-seven different procurement processes in such a highly technical and sensitive area. Instead, it would probably make economic sense for a central agency to continue to manage this. Counties can still own the function and need not relinquish control over the resources dedicated to drug procurement for their counties; they could simply be assigning the function of drug procurement to the national government, while retaining the responsibility and funding. 4.17 Principle 4: Services that provide benefits to more than one county (spillovers) are best assigned to the national level. This is because individual counties may be unwilling to pay for services whose benefits accrue partly to residents beyond their boundaries, resulting in under-provision. This may apply to the governance of water bodies including lakes and rivers, as well as the management of provincial level hospitals, two functions that remain to be assigned in detail. The principle of geographic spillovers is acknowledged implicitly in the Fourth Schedule of the Constitution, where the national government is given responsibility for national trunk roads, while county roads are devolved. Counties may also join together to form larger units―regional units if you will―to achieve these economies of scale. In a similar vein, some local public services not only provide benefits to those obtaining specific services, but also to society at large, and so involve issues of national interest. For instance in health, immunisation programs only prevent outbreaks if herd immunity is established through inoculation of a sufficiently large number of children. The Constitution anticipates this by largely assigning policy or standard setting functions, including housing, land planning, construction and maintenance of county roads, energy, health, agriculture, veterinary, and tourism to the national government. 4.18 Principle 5: Capacity should be built before functions are fully transferred to county governments. Capacity limitations may dictate in practice if subnational units of government are ready to assume expanded responsibilities. Counties should only receive functions that they have the capacity and infrastructure to deliver. The Constitution in Article 190(4) allows the national government to, if necessary, assume responsibility for functions that the county is unable to perform: therefore devolving too much too soon could result in re-centralization. Moreover, the Sixth Schedule of the Constitution provides for a gradual transfer of assigned functions as counties develop capacity, so that no county is responsible for a function that it does not have capacity to perform. Managing the process 4.19 The Task Force on Devolved Government (TFDG) recommended a process for clarifying function assignments. The Task Force was officially launched on November 8, 2010 by the Deputy Prime Minister and Minister for Local Government. The purpose of the TFDG was to help think through the implementation of the devolution process, and advise the government on policy and legal frameworks of devolving power, resources and responsibilities to the people of Kenya, for effective local development. In their final report released in 2011, the Task Force recommended a structure through which relevant line ministries would collaborate in sectors, to make recommendations on clarification of function assignment. These recommendations would flow to the Transition Authority (TA), which would then decide what to recommend to government. Since the TA leadership includes (in addition to appointed members) Principal Secretaries from the Office of the President, and the State Departments 38 Devolution without disruption – Pathways to a successful new Kenya of Devolution Affairs, Public Service, Finance, Planning and Justice, as well as the Attorney General as an ex-officio member, it will be able to decide whether higher level political approval is required to sign off the proposed assignments. 4.20 Many line ministries have started formulating function transfer strategies but they have had little guidance from central ministries. The TFDG in conjunction with KIPPRA, worked on developing a framework for implementing function assignment that will be able to help to inform the process. The purpose of the framework is to outline a conceptual basis for functional and competency assignment, based on theory and international experience, while placing the conceptual argument in the Kenyan context. The framework provides a platform on which line ministries can engage in the task of planning for devolution functions, from a common starting point. This framework is described in Box 4-4. Box 4-4: Framework for Function Assignment: Proposition by the Task Force on Devolved Government. To undertake the detailed work of function assignments within each sector, Function Assignment and Competency Teams (FACTs) and a Technical Working Group (TWG) should be established. A FACT should be established for each sector in which functions will be devolved. FACTs should align to sectors in the Medium Term Economic Framework (MTEF), and should fully cover the functions listed in Part 2 of the Fourth Schedule of the Constitution. The FACTs would then be requested to carry out analysis of the activities relevant to that sector and prepare a Position Paper and Strategy on Devolution in relation to the Ministries within their sector in consultation with the TWG, whose role it is to coordinate the FACTS and submit the final Position Papers to the Transition Authority. The roles of the FACTs will be to provide advice and recommendations in relation to concurrent functions as follows: ▪ un-bundle and define activities to be carried out at each level of government. ▪ define the roles and responsibilities of each level of government. ▪ define national norms and standards. ▪ identify the institutional arrangements for intergovernmental management. ▪ estimate the likely cost to counties of delivering the devolved functions. At the heart of the Functional and Competency Assignment process, is the preparation of detailed documents setting out how functions will be devolved. Some ministries and sectors have already begun or completed the preparation of a Sector Functional and Competency Assignment Policy Paper. Its primary objective is to provide the underlying rationale and justification for recommending which services currently being handled by the line ministry should be devolved; which can be delegated (under the provisions of Article 187 of the Constitution) to county governments; and which are to be retained by ministries as residual functions. The Policy Paper and Strategy will set out the current modes of service delivery, resources utilized, and existing service gaps. For each activity pertaining to the sector, the FACT would analyse and recommend how it could be devolved to county governments and include the rationale for their recommendation. To ensure comparability, each Position Paper would follow a common format and be based on a consistent set of assumptions, and terms of reference. The Policy Papers and Strategies will be assembled by the FACTs and presented to the Technical Working Group. There will be an opportunity for the public to react to the proposals for devolution of services and provide feedback. Ultimately, depending on which model is chosen, Cabinet or the Transition Authority will decide which activities will be transferred to county government. Where any changes are proposed, the FACTs will be advised along with requests for changes in the Policy Paper and Strategy. Source: Task Force on Devolved Government (2011). 39 Chapter 4: Why it all Starts with Function Assignment Sectoral implications 4.21 Devolution will succeed or fail through service delivery. This means that the sectoral implications of devolution are among the most important details to get right. This is precisely why it is so important to develop a process that sector ministries own and actively participate in, and to get started on it as soon as possible. It seems likely that the drafters of the Constitution did not look to existing service delivery arrangements, and so they left some extremely important questions unanswered, in the way the Fourth Schedule, under which county functions are assigned, addresses county responsibilities. 4.22 International best practice shows that successful devolution requires a framework outlining a well- defined and consultative process for formulating and defining functions. This applies, for example, on how to unbundle functions, how to create new policies around staffing, financing and capacity building, and how best to pace implementation, among other issues. 4.23 In the health sector, clarity on what constitutes county health services will be key to unbundling the sector. The Constitution assigns national referral health facilities and health policy functions to the national government, while assigning county health services to the counties. Whether functions such as provincial health management, provincial hospitals, HIV programs, drug supply, public health programs, health information systems and environmental health are county health services need to be decided by the TA. Programmes such as the National Aids Control Programme and the Family Planning and Maternal and Child Health programme which involve a mix of policy, implementation and specialized materials, might be managed concurrently. 4.24 Water functions are going to be one of the most complicated to assign going forward. The water sector in Kenya has already undergone significant reform to streamline water supply and water service provision by: (i) creation of corporate entities with professional management and at arm’s length from politics; (ii) separation of assets development and holding (water service boards) from retail services (water service providers); (iii) establishment of a national regulator to oversee the sector; and (iv) introduction of a uniform national tariff application mechanism and ring-fencing, to ensure that revenues can be reinvested in maintenance and expansion of services. Restructuring the sector once again to conform to constitutional provisions potentially puts these reforms at risk. The water boards that manage bulk assets span county boundaries; this function might be better implemented by the national government, or managed collaboratively by the relevant counties. 4.25 Public Investment as a national function overlaps with county capital investment functions. County functions in roads, energy and even ferries and harbours need to be distinguished from public investment functions, with significant spill over effects. Considerations such as the ability of the county to properly invest in large-scale infrastructure, might influence how the national government should contribute in its capacity as public investor. Guiding the function assignment process successfully 4.26 The TA should guide the function assignment clarification process, by enshrining the appropriate principles in criteria provided to the teams that are developing the detailed proposals. The transition process necessitates close communication, coordination and collaboration between the national 40 Devolution without disruption – Pathways to a successful new Kenya ministries and eventually with county governments. Ensuring the activities that constitute the full implementation of a function are carried out requires frequent interaction between the national and county governments. Beyond the transfer of devolved functions, strategic thinking is needed about how to make best use of the delegation provisions that allow additional national functions to be performed by county governments, which will allow the full benefits of devolution to be realised. This process will take time and there should be room for adjustment as devolution is implemented. The sooner that line ministries begin the process of collaborating amongst each other, the better. The TA will play a crucial role in making this happen. 4.27 The transfer of assigned functions will take place over time. The Sixth Schedule of the Constitution calls for the transfer to happen ‘no more than three years from the date of the first election of county assemblies’. Instead of having all county functions transferred on day one of the operations of county governments, the Constitution recognizes the need for a gradual transfer to ensure that services are not disrupted. At the same time, it puts the onus on the national government to ensure that counties are ready, within the three-year period, to run these functions. 4.28 The TA has been established under the Transition to Devolved Government (TTDG) Act, to facilitate and coordinate the transition. To do this effectively, the TA needs to be properly resourced to carry out its functions. Funds need to be provided immediately for them to set up internal operations, including staff and infrastructure. Staff will be needed at the national level as well as in the counties, where tasks such as taking stock of assets and liabilities will be intensive. The Transition Authority will also need power and authority to compel national ministries to follow its directives with regards to function assignment. Although the TTDG Act section 8(1) indicates that the TA will have “all powers necessary for the proper performance of its functions�, this does not necessarily give it the teeth it needs. A better source of sufficient authority may be the provision in the Transition to Devolved Government Act, allowing the Authority to make regulations. 4.29 In practice, the TAs effectiveness as a coordinating body will depend on its credibility and relationships with central and line ministries. Implementing devolution is a change management process, and meaningful outcomes are more likely to be achieved through co-operation and fostering ministry ownership, than by heavy-handed application of law. It will be challenging for the TA to quickly earn the respect of line ministries, particularly if it is non ex-office members are new to government, and do not have established networks in the bureaucracy. Getting active involvement from the Permanent Secretaries themselves (rather than junior representatives) will be crucial for the TAs effectiveness. 4.30 The TTDG Act mandates a seemingly simple but practically complex process of handing over functions to counties. It prescribes that each county should apply to the TA for the transfer of functions. This allows the Transition Authority to evaluate whether or not the county has the capacity to run the function before it is transferred. There are two advantages of this “pull� approach, where the counties apply for each functions, rather than the national government deciding when to “push� them to the county level. First, it ensures a thorough evaluation of county capacities, and second, it serves the double purpose of allowing the county to understand its starting point, and define its goals. It is likely 41 Chapter 4: Why it all Starts with Function Assignment that the very requirement to apply will serve as a natural test of basic level of capacity. However, there are drawbacks. Depending on how the process is managed, it could get very complicated. It may prove costly and time-consuming to conduct many different capacity assessments―time and money that could be better spent operationalizing counties. To avoid these potential complications, it is suggested that functions be bundled together,and that there should be a clear pathway for counties to follow― beginning with the most localized and straightforward functions, and progressing to those involving more complex systems or technical capacity. Table 4-3 includes some criteria for readiness, such as having PFM systems, sectoral plans and HR systems in place and other function specific infrastructure matches with proposed timings. 4.31 Line ministries as well as Central Government should play a key role in the process. Since line ministries have already been running the functions to be devolved, they will bear critical information on costs and challenges to implementation. Indeed, the most striking thing about the devolution in Kenya so far, is the complete absence of any sustained coordination between line ministries and central agencies working on preparation for devolution. The TA should recognize that line agencies are key to devolution, and that involving them in a collegial way to secure their involvement, is the best way to ensure its success. At the same time, the Transition Authority will also provide an important buffer against the tendencies national ministries have to resist devolution―in part to retain control over resources, but also out of genuine concern that the counties may not be able to deliver the functions to standard. 4.32 Proposals clarifying function assignments and identifying readiness criteria should be part of the planning process mandated in the TTDG Act. The Act provides that the Commission on Implementation of the Constitution can require each ministry to submit its plans for implementing the Constitution. The TA is responsible for issuing guidelines about what the implementation plans should contain. At a minimum, the guidelines should include a requirement for the ministries to develop proposals for clarifying function assignments in their sector. Some ministries have already begun this process, but not all the function assignment issues in each sector have been addressed. The proposed process developed by theTFDG and refined by KIPPRA would be a useful way for ministries to approach this task. As ministries consider how functions will operate under a devolved system, they will also need to reflect on their own future role and the intergovernmental coordination mechanisms for their sector. Implementation plans should provide a comprehensive description of how the sector will operate, and the steps required to put the necessary legislative, policy and institutional architecture, to make this effective. Box 4-5 suggests a number of topics that could be included in implementation plans. Asymmetric transfer of functions 4.33 Since some counties will be ready before others, the constitution allows for asymmetric transfer of functions. While some counties could be ready from day one, in othersthe national government must run county functions on the county’s behalf, until the county is prepared to take them up. The “pull� approach of counties applying to the Transition Authority for their functions virtually guarantees asymmetry. 42 Devolution without disruption – Pathways to a successful new Kenya Box 4-5: Possible components of ministry implementation plans. ▪ Proposed revision of sector legal framework. ▪ Ensure counties adequately empowered through amendment of national laws to carry out regulatory functions. ▪ Establish intergovernmental architecture (e.g., National Council of both levels of government to discuss and agree policy priorities). ▪ Proposed assignment of functions (unbundled to activity level). ▪ How sector planning will work and how it will intersect with integrated plans at county level. ▪ Proposed sequencing (and bundling) of functions for phased transfer. ▪ Proposed criteria for county readiness to receive each bundle of functions. ▪ Process for deciding on national standards for the sector (and, if appropriate, expenditure norms) including involvement of county governments. ▪ Common sector performance measurement and monitoring framework. ▪ Plans for capacity gap analysis. ▪ Approaches to capacity building to address capacity gaps. ▪ How donor funding will be managed in the sector to ensure it is aligned with county plans and budgets. Source: World Bank staff analysis. 4.34 A fully asymmetric transfer process would be overly complex. In theory, a process whereby each of the forty-seven counties could apply for a unique set of functions in a different order would result in a highly complex (to manage) and in-transparent (to monitor) architecture. Transferring groups of functions together in phases according to their complexity, could mitigate this risk. Such phasing-in could be approached in three ways. One approach is to transfer functions sector by sector, for example, by first focusing on health functions, then agriculture and so on. Another way would be across several sectors by types of activities. A recommended approach would be to transfer functions as bundles, using a phased approach as described in Table 4-3. During the first phase immediately after county formation, the TA would transfer functions that the county must take charge of on day one. According to the TTDG Act, these functions must be identified 30 days before the next election. Next, counties could apply for functions involving basic service delivery, provided they have basic human resource and PFM systems in place, and a plan or strategy for the relevant sector has been approved. The final phase would involve more specialised functions that have more complex technical requirements. Those functions would be transferred once the county demonstrates the systems or technical skills required. Table 4-3: A phased approach to transferring functions to counties. PHASE Timing of transfer Readiness criteria Types of functions Examples PHASE 1 Immediately after • None/ Automatic Functions not • Local government functions county formation currently being • County assemblies and executives performed by any • Establishment of basic systems (PFM, national agency HR, Procurement, etc.) PHASE 2 As soon as basic • Basic county Basic operation of • Operation of rural and district health county systems are HR and PFM service delivery facilities, ambulance services operational systems programs • Provision of agriculture extension • Sectoral plans services PHASE 3 Only when specific • Specific system More complex • Purchasing and distribution of readiness criteria or capacity programs and pharmaceuticals (only after stock are met requirements supply chain control and ordering system in place) management Source: World Bank staff analysis. 43 Chapter 4: Why it all Starts with Function Assignment Refining function assignments going forward 4.35 Even with a framework in place, sectoral assignments are likely to be readjusted many times over the medium term as Kenya works toward a function assignment mix that suits its circumstances. It is very common for function assignments to be dynamic, evolving over time. There is no right way to assign functions; sometimes these changes result from a process of fine-tuning, or they may simply reflect the perspectives of different political parties in power. It is to be expected that function assignment will not be static. Responsibilities change once it becomes apparent that another level of government might be better suited to undertake them. What will ensure a smooth function assignment over the years is a stable, collaborative and coordinated cooperation between the county and national government. 4.36 The constitution allows functions to be transferred between levels of government or delegated to decentralized entities such as urban areas and cities. Article 187 specifies that a function can be transferred if it can be more effectively performed at another level, and that the necessary resources to perform that function must also be transferred to the receiving level of government. This allows a fair degree of flexibility into the system, retaining accountability and promoting coordination over time between the two levels of government. Part 3 of the Intergovernmental Relations Act (IRA) outlines how this will be operationalized, but subsequent legislation in regulations is needed to flesh the details out more specifically. However, once the funding of line ministries is reduced in order to free up the fiscal space needed to pay county transfers, it may be in the interests of line ministries to transfer some of their functions (and thus cost responsibilities) to county governments. The procedures for delegation of additional functions should ensure that funding considerations are adequately accounted for, and that the annual calculation of the vertical share takes account of any major adjustments in function assignment. 4.37 Functions can be further delegated to the sub-county entities such as urban areas and cities. The transfer should be effected by written agreement between the two government levels, and include key details such as performance standards, a resourcing framework, a capacity building framework, and a dispute resolution framework. The delegation of functions to urban areas and cities through boards needs to be clear too. 4.38 Function assignments can become very complex―the more detailed they are. The more there is joint provision of a function between two levels of government, the more detailed it has to be. In these circumstances, it is important that both the public being served, and the respective government officials, understand what the split is. If they do not, the likely result is gaps in services, or duplication. There should be an authoritative document that everyone can refer to, that solves any disputes about what each level of government is expected to do. This is important if function assignments are changing on a regular basis. 4.39 It is therefore important to build transparency within the function transfer process. This implies that consultations involve the key stakeholders in the process, and that the framework, discussions and outcomes are communicated throughout the process. Citizen participation over the period of constitutional implementation, and the legislation of new bills has been good, and it should ideally be maintained through the function assignment process. 44 Devolution without disruption – Pathways to a successful new Kenya 4.40 The government should publicize function assignments once they are finalized, and continue to update them. The TA is required in its parent law to gazette final function assignments and responsibilities, as well as progress reports, including changes in assignments over the three-year period of its existence. However, these bulletins are not typically read by the average Kenyan, as they are not easily accessible. To be effective, function assignment responsibilities should also be disseminated further, using channels such as civic education campaigns, making information available online, using civil society outlets, as well as booklets at county offices. In Kenya, much use is made of signs and posters at government offices, that detail its office functions and these can be maintained for function assignment allocations. CHAPTER 4: KEY INSIGHTS / RECOMMENDATIONS. A clear assignment of functions and responsibilities between levels of government forms the basis of the future intergovernmental fiscal transfers system, and is a sine qua non condition for accountability. Schedule Four of the Constitution provides high level guidance, but a much more granular approach is needed that unbundles each function into its constituent activity. The Transition Authority to be set up has the mandate to carry out function assignment – in close cooperation with, but also independence from central and line ministries- but it will take some time before it is operational. Managing an asymmetric functions’ transfer process to forty-seven counties would be overly complex, unless some uniform framework for ‘bulk transfers’ is agreed upon. *** The process of clarifying function assignments should begin now. The first order of business for the newly established Transition Authority should be to initiate the function assignment process, by creating a framework of assignment to guide the process. Line Ministries and Central Government should be key stakeholders involved in the Function Assignment Process. Line ministries understand the functions best, and are well placed to advise on assignment. However Line Ministries should also be challenged because they will naturally seek to hold on to functions. Central government is best placed to advise on matters such as staffing, at the county levels. There should be a roadmap to streamline and simplify the initial asymmetric transfer of functions. One way to do this would be through a three phase approach, with functions transferred in ‘lumps’ at each stage, based upon clearly defined readiness criteria. A process for publishing the function assignments, and regularly updating them should be agreed upon. The objective should be that citizens have access to clear, simple, up-to-date information, on what the respective responsibilities of national and county governments are―recognizing that for some time, the functions may differ from county to county. These should also accurately identify the urban functions which different counties have delegated to urban boards under their control. 45 CHAP TE R FIV E Determining how much counties need 5.1 Assigning functions clearly across levels of government is so crucial because this assignment will need to drive the intergovernmental division of resources. At minimum, before other objectives are considered, counties and the national government should be given enough resources to carry-out their mandated functions. This process is described as “costing�, but in reality it is more about working out a fair split of available resources between levels of government- because there is never enough funding available to satisfy all possible needs-taking into account the assignment of functions. This process is likely to be even more complicated, during the first three years of Kenya’s county government, because some counties will have the responsibility for some functions, which in other counties still remain with national government. 5.2 In practice, intergovernmental transfer systems are often developed without much thought to aggregate subnational needs. This approach is understandable, because resource allocation is highly political, but it is also risky. There are two risks that arise if there is a mismatch between service delivery responsibilities, and available funding. The first risk if counties get less than they need, adjustments need to be made on the expenditure side to fit the funding, resulting in below-target service delivery, or outright interruption. The second risk, if funding to counties exceeds their needs, there will be wasteful consumption at the local level and fiscal stress at the centre (possibly resulting in under- provision of public goods). 5.3 By prescribing a minimum transfer to counties, Kenya’s Constitution has not pre-empted the need for aggregate costing. This is because counties are likely to need significantly more funding than the minimum 15 percent of national revenues that has been guaranteed to them. This chapter discusses how county needs can be quantified, and in the following chapter, different instruments available to meet them will be considered. 5.4 Time is quickly running out and an important decisions needs to be made. In 2013/14, Kenya will finance its county governments, for the first time, for a full year. The 2013/14 budget process has already begun, but many issues involved in financing county governments have not yet been decided. First, it is key to estimate the current cost of devolved service delivery. Then only, once remaining available fiscal space has been determined, can a discussion begin on what additional objectives need to be pursued. Determining county needs 5.5 Function assignments form the basis for determining county needs. Many international approaches do not cost county needs, in order to determine a reasonable vertical share; instead it is the outcome of a politically negotiated process. In some cases these ad hoc approaches have proven to be problematic because they are not transparent, subject to political manipulation, and make it easier to deny the link between expenditure responsibilities and revenue resources.1 The best way to ensure a closer match between functional assignments and total subnational revenues, is to use a more analytical approach. 46 Devolution without disruption – Pathways to a successful new Kenya 5.6 Historical spending by the national government on future county functions is a good starting point, but there will be additional needs. The Ministry of Finance (MoF) has decided to use historical spending as the initial approach to costing county functions. Devolved functions are already being performed by the national government at various levels: provincial, district, and division. It is therefore reasonable to assume that counties, adjusting for scale and efficiency, will need at least as much as is being allocated to these functions cost now. However there are some major gaps in historical spending to take into consideration: (i) urban service delivery that is currently funded through a combination of the Local Authorities Transfer Fund (LATF) and own revenues of Local Authorities (LAs); (ii) donor funding especially in health where there is extensive off-budget spending; and, (iii) new county start up and administration costs. In addition, historical costs reflect service gaps that have resulted from long standing neglect of certain areas of Kenya, and therefore do not capture the resources needed to provide a more equitable level of services. Therefore, historical spending is not the end point of the costing methodology, but only a start towards a steady progression to more equitable sharing of resources around the country. Simulating aggregate county expenditure needs 5.7 To begin to estimate aggregate county needs, budget data for 2010/11 budget data was uesd to make some initial estimates of how much devolved functions are likely to cost. It rests on a number of assumptions about how functions might be assigned. We have made separate calculations along three scenarios: (i) a radical scenario, whereby all functions that could possibly be financed and administered by counties (as per the Fourth Schedule of the Constitution) are assumed to be devolved; (ii) a moderate scenario where more functions are undertaken concurrently between national and county governments; and, (iii) a conservative scenario where more functions are assumed to remain with the national government. 5.8 The three scenarios reflect different possible responses to questions about function assignment. As such, they help to test the fiscal impact of different models of function assignment. For the purpose of illustration, consider how functions currently being performed by district administrations are assigned in three scenarios. (1) In the radical scenario, the entire District Administration function, staff and costs are assumed to be fully absorbed by county governments, with county officials administering former district administration tasks, such as conflict resolution functions performed at the village chief level. (2) In the moderate scenario, it is assumed that District Administrations are retained as a national agency with fewer district officers performing more limited functions, such as those interventions affecting national security. To simulate this function assignment decision, 75 percent of the district administration budget is allocated to the counties, while the rest is retained by the national government. (3) In the conservative scenario, more of the existing functions of the District Administration continue to run parallel to the county government structures, while the existing functions that are assigned to counties―liquor licensing and control of drugs and pornography―are assumed to involve a transfer of half the existing costs to county government. In this conservative scenario, 50 percent of the current costs of running district administration (including staff) are assumed to be devolved to counties. Table 5-1 illustrates the fiscal impact of these different scenarios on the estimated costs to counties of running all or some of the district administration functions. 1 Bahl and Wallace (2004). 47 Chapter 5: Determining how much Counties need Table 5-1: Devolving District Administration – three scenarios. Vote 01: Provincial Administration and Internal Security Sub Vote 011: Field Administration Services Head 006: District Administration Total Costs (R+D) KES 5,491,964,376 Radical Scenario 100% devolved 5,491,964,376 Moderate Scenario 75% devolved 4,118,973,282 Conservative Scenario 50% devolved 2,745,982,188 Source: World Bank staff analysis and estimates of Recurrent and Development Expenditures 2010/11. Note: R = recurrent, D = development. 5.9 The outcome of the simulation is critically sensitive to assumptions about assignment of big ticket items. A few of them will make up the largest costs that might be shifted to county governments (see Box 5-2). This simulation was designed at a time when public debate assumed many functions would be fully devolved to counties. The quantification of county needs will vary considerably, if all or any of these assumptions are changed. 5.10 Many large functions, including Education, will not be devolved at all, or will only have a small percentage of devolved functions. This makes sense for some of the more intuitively national functions such as State House, Defence, Immigration, Cabinet Office, National Assembly, State Law Office and Foreign Affairs. Within Education, the only devolved function in this simulation is Early Childhood Development Education which under the Fourth Schedule can be categorised as pre-primary education ―a county function. Some other functions which are retained national or are residual include those in Gender Children and Social Development and Environment and Mineral Resources. 5.11 Table 5-2 shows the amount budgeted for each of the devolved functions under each ministry for the 2010/11 year. It also shows the breakdown between recurrent and capital, and the overall percentage of the total each set of functions accounts for. Kenya’s budget is organised in ministry units, and each vote represents a single ministry. The figures sum the amounts in both the recurrent and development budget. In the table, “Vote� is equivalent to a Ministry. Implications of the simulated costing analysis 5.12 In a moderate scenario, total estimated devolved cost is KES 169 billion in the simulation year 2010/11. Legislation passed since August 2011 forecasts a costing based on assumptions similar to the moderate scenario. Issues surrounding certain ‘big ticket’ items, such as who manages the provision of urban services, who runs provincial hospitals, and the fate of projects funded by the Constituency Development Fund (CDF), appear to have been resolved. The Urban Areas and Cities Act (UACA) of 2011 provides that the cost of urban service management will be borne by the counties through city boards or county governments, which is an assumption made in the moderate costing analysis. Counties are also assumed to be responsible for the costs of running provincial hospitals, according to the Budget Policy Statement of April 2012 (BPS), aligning with a view that provincial hospital functions are county functions. The Ministry of Finance proposes that community-based projects funded by the CDF be managed by the county governments, rather than national government Members of Parliament (MPs), which is also an assumption made in the moderate scenario. The 2012/13 budget estimates also provide an indicative coding of the budget lines that are assumed to be devolved. 48 Devolution without disruption – Pathways to a successful new Kenya Box 5-2: How the big ticket items were allocated differently in the three scenarios. Provincial Hospitals (FY2010/11 cost KES 4.5 billion): The allocation to Provincial hospitals amounts to 18 percent of the KES 24.1 billion budgeted for the Ministry of Medical Services (MoMS). In both the radical and moderate scenario, Provincial hospitals are assumed to serve as a high level county health function, and ‘belong’ to the county in which they are located. As a result, costs are assumed to go to the county level. In the conservative scenario, Provincial Hospitals are retained at the national level, assuming they are upgraded to referral hospitals, or are treated as unassigned, residual function, so costs remain at the national level. Small capital works (projects currently funded out of the CDF) (FY2010/11 cost KES 14.4 billion): The allocation for CDF funded small capital works amounts to 77 percent of the KES 18. 6 billion budgeted for the Ministry of Planning, National Development and Vision 2030. In the radical and moderate scenarios, CDF projects are assumed to be taken over by counties, and costs are assigned accordingly, in full. However, in the conservative scenario, Parliament retains control over CDF projects to be funded through a separate mechanism (such as a conditional grant): costs remain with the national government. Construction of new Roads (FY2010/11 cost KES 30.4 billion): The allocation for the construction of new roads amounts to 68 percent of the KES 44.8 billion budgeted for the Ministry of Roads―this does not include funding from RMLF which is a recurrent maintenance cost rather than construction. In all three scenarios, funding is allocated according to the Fourth Schedule: national highways (national function), and rural and urban roads (county functions). Currently approximately 50 percent of the funding budgeted for road construction is going to national highways, with the rest going to rural and urban roads. The allocation of cost is therefore applied at 50 percent to national and 50 percent to county in all three scenarios. Despite the fact that the rural and urban road networks make up 90 percent of the coverage road network,2 cost associated with their construction are much lower than for national highways. Road Maintenance (FY2010/11 cost KES 11.32 billion): The allocation to road maintenance through the Road Maintenance Levy Fund is considered external revenue to the Roads budget, and therefore is not included in Roads overall budget. In all three scenarios, road maintenance allocates the portion going to county roads including funds to the Kenya Rural Roads Authority and Kenya Urban Roads Authority, which is 47 percent of the overall RMLF to county governments. Urban Services (FY2010/11 cost KES 12.3 billion): The allocation for urban services through the Local Authorities Transfer Fund (5 percent of national income tax) is over 2 times the Local Government vote budget of KES 5.2 billion. LATF is also treated as external revenue to the Local Government budget in the same way RMLF is. In the radical and moderate scenario, urban service provision and management costs are assumed to be devolved―which is in line with legislation under the UACA Act 2011. In a conservative scenario, urban management was assumed to be a national function, though now it is confirmed that it is a county function. These four items alone amount to approximately 11 percent of the total KES 639.5 billion 2010/11 budget (both recurrent and development), and 43 percent of the KES 168.6 billion total estimated devolved government costs (excluding county start-up and administrative costs) in a moderate scenario. Source: World Bank staff analysis. 2 Road Project Project Appraisal Document, World Bank (2011). 49 Table 5-2: Funding for devolved functions under the three scenarios using simulation year of 2010/11*. Devolved Budget (KES billions) % of recurrent/development vote % total vote devolved % devolved vote to total devolved devolved 2010/11 budget Development� Vote “Recurrent/ Conservative Vote (Ministry) Total budget for Radical Moderate Conservative Radical Moderate Moderate Conservative 2010/11 (KES Billion) Radical Radical Moderate Conservative R 1.36 1.23 1.02 65.4 59.3 49.1 Roads (Note 2) 44.85 70.8 70.5 70.0 17.42 18.75 27.75 D 30.40 30.38 30.35 71.1 71.0 71.0 13 Road Maintenance Levy Fund R 11.32 11.32 11.32 50.0 50.0 50.0 22.64 n/a n/a n/a 6.21 6.71 10.01 (RMLF) Projects (Note 2a) D - - - - - - R 9.38 9.21 7.95 43.0 42.3 36.5 Medical Services (Note 3) 19.63 45.1 44.4 39.7 5.97 6.36 8.47 D 1.51 1.51 1.62 65.5 65.5 70.4 11 R 4.27 4.27 - 19.6 19.6 0.0 Provincial Hospitals 4.49 18.6 18.6 0.0 2.46 2.66 0.00 D 0.21 0.21 - 9.1 9.1 0.0 Public Health and Sanitation R 7.51 6.96 6.20 78.8 73.0 65.0 49 17.38 65.3 55.7 46.9 6.23 5.74 7.22 (Note 4) D 3.84 2.72 1.96 48.9 34.7 24.9 R 2.80 2.58 2.12 74.0 68.3 56.0 20 Water and Irrigation (Note 5) 23.88 84.2 76.9 32.8 11.03 10.89 6.93 D 17.31 15.79 5.72 86.1 78.5 28.4 R 3.18 3.17 2.95 89.5 89.2 82.9 19 Livestock Development 6.51 91.8 88.3 81.4 3.28 3.41 4.69 D 2.79 2.58 2.35 94.6 87.3 79.5 R 4.79 3.71 3.28 59.7 46.3 40.9 10 Agriculture (Note 6) 16.40 46.7 39.2 31.7 4.20 3.81 4.60 D 2.86 2.71 1.92 34.2 32.3 23.0 Regional Development R 0.78 0.78 0.78 95.5 95.5 95.5 9 6.19 77.9 77.9 77.9 2.65 2.86 4.27 Authorities (Note 7) D 4.04 4.04 4.04 75.2 75.2 75.2 R 0.95 0.95 0.94 100.0 100.0 99.4 56 Fisheries Development (Note 8) 4.17 100.0 100.0 99.9 2.29 2.48 3.69 D 3.23 3.23 3.23 100.0 100.0 100.0 R 0.70 0.70 0.70 33.7 33.7 33.7 30 Energy 19.83 17.1 17.1 17.1 1.86 2.01 3.00 D 2.69 2.69 2.69 15.2 15.2 15.2 R 0.90 0.79 0.72 61.9 54.3 49.2 59 Public Works 6.50 52.6 50.9 49.8 1.88 1.96 2.86 D 2.52 2.52 2.52 49.9 49.9 49.9 Provincial Administration and R 6.55 4.31 2.77 15.7 10.3 6.6 1 45.63 16.3 10.2 6.6 4.09 2.76 2.65 Internal Security (Note 9) D 0.91 0.35 0.23 23.1 8.9 5.7 R 1.60 1.28 0.99 29.8 23.9 18.4 42 Youth Affairs and Sports 9.17 37.0 32.3 27.9 1.86 1.76 2.26 D 1.79 1.68 1.57 47.3 44.3 41.4 R 1.96 1.81 1.71 49.9 46.1 43.4 55 Forestry and Wildlife 5.92 52.8 46.2 40.7 1.72 1.62 2.13 D 1.16 0.92 0.70 58.7 46.5 35.4 Development of Northern Kenya R 0.26 0.21 0.13 100.0 80.0 50.0 58 3.36 100.0 80.0 50.0 1.84 1.59 1.49 and Other Arid Lands D 3.10 2.48 1.55 100.0 80.0 50.0 R 0.27 0.13 0.03 51.6 24.7 6.2 44 Housing 3.00 73.4 62.0 49.4 1.21 1.10 1.31 D 1.93 1.73 1.45 78.0 69.9 58.5 Source: World Bank staff estimates based on Estimates of Expenditure (Recurrent and Development) 2010/11 Government of Kenya. *See note 1 in Box 5-3. Chapter 5: Determining how much Counties need 50 51 Cont. Table 5-2: Funding for devolved functions under the three scenarios using simulation year of 2010/11* (continued). Devolved Budget (KES billions) % of recurrent/development vote % total vote devolved % devolved vote to total devolved devolved 2010/11 budget Development� Vote “Recurrent/ Conservative Vote (Ministry) Total budget for Radical Moderate Moderate Conservative Radical Radical Moderate Conservative 2010/11 (KES Billion) Radical Moderate Conservative R 1.17 1.17 1.17 3.8 3.8 3.8 7 Finance (Note 10) 52.07 2.3 2.3 2.3 0.64 0.70 1.04 D 0.00 0.00 0.00 0.0 0.0 0.0 R 1.36 1.18 0.96 80.5 69.6 56.6 41 National Heritage and Culture 2.09 76.9 67.3 55.7 0.88 0.84 1.03 D 0.25 0.23 0.21 61.7 57.6 52.0 Cooperative Development and R 0.92 0.92 0.92 94.3 94.3 94.3 22 1.15 90.0 90.0 90.0 0.57 0.61 0.92 Marketing D 0.11 0.11 0.11 65.3 65.3 65.3 Nairobi Metropolitan R 0.32 0.24 0.16 100.0 75.0 50.0 57 1.48 100.0 75.0 50.0 0.81 0.66 0.65 Development D 1.16 0.87 0.58 100.0 75.0 50.0 Planning, National Development R 0.67 0.35 0.35 32.3 16.8 16.8 86.1 81.0 3.6 0.94 0.44 0.59 and Vision 2030 (Note 11) D 1.04 0.40 0.32 92.9 89.0 2.0 6 18.64 Constituency Development Fund R - - - - - - 77.0 77.0 0.0 7.87 8.51 0.00 Projects (Note 1a) Devolution without disruption – Pathways to a successful new Kenya D 14.35 14.35 - 100.0 100.0 0.0 R 0.90 0.75 0.65 100.0 83.6 72.6 12 Local Government (Note 12) 5.26 100.0 81.8 12.5 2.88 2.55 0.58 D 4.36 3.55 - 100.0 81.4 0.0 Urban Services funded through R 12.30 12.30 - 100 100 - 12.30 n/a n/a n/a 6.75 7.29 0.00 LATF (Note 12) D - - - - - - R 0.39 0.39 0.39 0.3 0.3 0.3 31 Education (Note 13) 135.89 0.3 0.3 0.3 0.23 0.25 0.37 D 0.03 0.03 0.03 0.4 0.4 0.4 R 0.10 0.10 0.10 7.8 7.8 7.8 14 Transport 5.44 7.7 7.7 7.7 0.23 0.25 0.37 D 0.32 0.32 0.32 7.6 7.6 7.6 R 0.49 0.42 0.22 33.6 28.8 15.2 16 Trade 1.99 35.6 30.3 12.0 0.39 0.36 0.21 D 0.21 0.18 0.02 41.1 34.6 3.1 R 0.07 0.07 0.07 6.2 6.2 6.2 15 Labour 1.56 13.0 13.0 13.0 0.11 0.12 0.18 D 0.13 0.13 0.13 31.7 31.7 31.7 R 0.74 0.42 0.19 44.4 25.3 11.6 36 Lands 3.95 18.8 10.7 4.9 0.41 0.25 0.17 D - - - 0.0 0.0 0.0 R - - - 0.0 0.0 0.0 35 Special Programme 7.40 14.7 7.3 2.3 0.60 0.32 0.15 D 1.09 0.54 0.17 32.4 16.0 5.0 R 0.22 0.22 0.13 14.9 14.9 8.8 32 Information and Communications 6.82 4.1 4.1 2.3 0.15 0.17 0.14 D 0.06 0.06 0.03 1.1 1.1 0.5 R 0.05 0.05 0.05 4.3 4.3 4.3 3 Public Service 1.58 6.8 6.8 6.8 0.06 0.06 0.10 D 0.05 0.05 0.05 16.0 16.0 16.0 R 0.08 0.08 0.08 5.5 5.5 5.5 60 Industrialization 4.28 1.8 1.8 1.8 0.04 0.05 0.07 D - - - 0.0 0.0 0.0 R 0.18 0.10 0.05 13.0 7.3 3.6 46 Tourism 2.38 8.8 5.1 2.7 0.12 0.07 0.06 D 0.03 0.02 0.01 3.1 2.1 1.4 Office of the Vice-President and R 0.19 0.05 0.05 1.8 0.5 0.5 5 11.95 2.4 0.4 0.4 0.15 0.03 0.04 Ministry of Home Affairs D 0.10 - - 5.3 0.0 0.0 Total Estimated Cost to County Governments R+D 182.35 168.64 113.06 Source: World Bank staff estimates based on estimates of Expenditure (Recurrent and Development) 2010/11, Government of Kenya. *See note 1 in Box 5-3. Chapter 5: Determining how much Counties need Box 5-3: Reference notes for table 5-2. Note 1 Devolved functions based on existing levels of services currently provided by national government and funded either through national budget or devolved funds. Does not include new costs of establishing county governments, running county assemblies and executives, or expansion of services to address infrastructure gaps. Only that part of urban service costs met from the LATF are included; current expenditure on urban services is approximately double this amount as around half is funded from Local Authority own revenues that are not captured in the national budget. Note 2 Roads includes development funding for the construction of new roads excluding national trunk roads. Note 2a RMLF is funded mainly through a levy on petroleum products and is used to manage the road network in all parts of the county. This funding is channelled through the Kenya National Highways Authority, Kenya Urban Road Authority and the Kenya Rural Roads Authority. 47 percent of this funding will go to county roads maintained by KURA and KERRA, while the rest will be retained at the national level. RMLF funding in the budget estimates is classified as a recurrent AIA . Note 3 Medical services covers costs of 2 national referral hospitals, specialist hospitals, 8 provincial hospitals and 280 district hospitals. It also covers the operating costs of the drug procurement agency. For the radical and moderate scenarios, it is assumed that provincial hospitals are a devolved function. Note 4 Public Health and Sanitation covers the operating costs of rural health centers and dispensaries. Promotive health interventions like family planning and child health and environmental health services are also covered here. Note 5 Water and Irrigation covers 10 water service boards and one water management authority. In a radical and moderate scenario, the water service boards are assumed to be devolved. Note 6 At 12, the highest number of parastatals can be found in Agriculture, some of which in the radical scenario are assumed devolved such as Horticulture Crops Development Authority and some of which are assumed national, such as the Kenya Agricultural Research Institute. It also covers extension services and 2 agricultural colleges. Note 7 Regional Development Authorities mainly covers 6 river and lake basin management authorities. Traversing several counties, all scenarios assume county management of the river and lake basins possibly through an intergovernmental arrangement. Note 8 Fisheries have been assigned to the county governments, except with regards to deep sea fisheries, which under the conservative scenario have been retained as a national function. Note 9 Provincial and district administrations provide overall coordination of government functions including security in Kenya's 280+ districts. Vote covers costs of police as well as district administration staff. Budget for district administrations covers the devolved functions of liquor licensing and control of drugs. Note 10 Finance covers district audit and treasury functions, which in all scenarios are assumed to be devolved. Note 11 Planning covers district development services and national programmes such as the national economic strategy and the Poverty Eradication Commission. Note 11a Projects covered by the Constituency Development Fund are put forward and managed by the local communities. The fund is divided amongst constituencies which have been increased from 210 to 290 according to the Constitution. Since CDF manage projects at the community level, the radical and moderate scenarios assume that they are devolved―a view shared by the Commission on Revenue Allocation and the Ministry of Finance. Note 12 Local Government covers the management of the 175 local authorities abolished in the Urban Areas and Cities Act. Urban Services have been devolved to the counties by the Constitution and further by the UACA 2011 as reflected in the radical and moderate scenario. Note 13 Primary and secondary education, university and research institutions are retained as national costs. Education costs at the county level are early childhood education, as well as village polytechnics. 52 Devolution without disruption – Pathways to a successful new Kenya 5.13 Roads contribute the highest costs to county governments. At 25.46 percent, a quarter of the total devolved budget goes to road function expenditures including construction and operational costs (18.75 percent), and road maintenance costs (6.71 percent). According to the 2012/13 budget estimates, the Kenya Rural Roads Authority (KERRA) and the Kenya Urban Roads Authority (KURA), will still be run at the national level, but the actual spending on rural and urban roads has been devolved to the counties. 5.14 After roads, health, water, agriculture and urban functions, as well as CDF projects are the six highest cost contributors to devolved government.3 Health functions including those of Medical Services and Public Health take up 14.76 percent of current spending on devolved functions. Water services and Regional Development Authorities take up a total of 13.75 percent of the devolved budget; and Agriculture and Livestock development take up a 7.22 percent share of the devolved budget. Urban functions (7.29 percent) and CDF projects (8.51 percent) represent a total share of 15.8 percent of the devolved budget. These few sectors combined take up a whopping 77 percent of the total devolved budget in 2010/11. 5.15 The 2012/13 Budget classifies Provincial hospitals as county function but, these regional institutions would be a significant burden on the individual counties where they are located. Provincial hospital costs were included only in the radical and moderate costing scenarios. If counties are expected to meet the cost of provincial hospitals from their equitable share, the cost burden of running each of them will be disproportionately larger on the counties where provincial hospitals are located. There are two ways to mitigate this consequence: (i) assigning provincial hospitals as a national function; and, (ii) provide conditional transfers to counties in which these hospitals are located. In the latter case, if provincial hospitals were funded through a conditional grant, the funding that currently finances them could not be counted towards the calculation of the equitable share. Converting the simulation into a real estimate of aggregate county needs 5.16 Simulating the current cost of devolved functions is only a starting point because it: (i) relies on bold assumptions; (ii) does not fully capture future county needs; and, (iii) ignores the timing of the actual transfer of functions. At the time of developing this analysis, the MoF was undertaking its own costing using information from ministries. Substantial assumptions made in the analysis need to be reconsidered and rationalized, in particular with the detailed coding of devolved functions in the 2012/13 budget. Since our initial assessment was undertaken without consultation with Ministries, specific issues remain to be discussed, and decisions taken before the costing can be finalized. A case in point is with health functions. The function assignment decision with respect to provincial hospitals has a substantial bearing on the overall number (both for aggregate devolved functions, and for the costing of the equitable share), since these costs account for 2-4 percent of the total. 5.17 Development spending is difficult to cost, since needs are not constant. S pending varies from year to year, so it makes sense to determine if the figures derived from a single year are anomalies with respect to the trend. In Kenya, the trend over the past six years (see Figure 5-1) indicates that the percentage allocation to development spending increased at a relatively constant rate, with the exception of spending in 2009/10 which saw a spike in spending due to the roll out of the Economic Stimulus Package projects in Medical Services, for the construction of health centers in every constituency, employment 3 Note that we have not factored in county start up and administration costs in this simulation. 53 Chapter 5: Determining how much Counties need of contract nurses (20 per constituency) and the purchase of motorcycles and bicycles for community health workers. There was also the rehabilitation and expansion of irrigable land under the Regional Development Authorities in the aftermath of the Horn of Africa Drought of in 2007/8. Agriculture in the same year took a slight dip to accommodate an increase in Livestock and Development due to the hit it took over the drought. In 2010/11 development expenditures picked up to the original growth rate following the previous year’s shift in spending―to pick up from 2008/9. In 2010/11, Medical Services was adjusted back to the previous range of 2009/10, as did Agriculture. Regional Development Authorities development spending goes down though not as pronounced as Medical Services. Livestock and Development sees a slightly stronger growth trajectory in 2009/10. Figure 5-1: Except for a spike in 2009/10, development spending has seen relatively constant growth from 2005/6 to 2010/11. 6 Economic Stimulus Estimates of Development Expenditure Package 5 4 (Kshs billions) 3 2 1 0 2005/06 2006/07 2007/08 2008/09 `2009/10 `2010/11 Regional Dev. Authorities Livestock and Dev. Health Medical Services Agriculture Source: World Bank staff analysis. 5.18 Functions funded by donors, especially in Kenya’s health sector, must be added to fully account for county needs. A large percentage of donor expenditures are not captured in the official budget. These ‘off-budget’ expenditures can be quite substantial in both recurrent and capital commitments. The first step would be to establish what expenditures are contributing, to the cost of delivering devolved services. 5.19 Another gap is the cost of urban services not covered by the LATF. Now that urban services have become the responsibility of county governments, it will be important to include them in the costing. Two aspects of urban service costs need to be factored into the costing. First, it is important to note that LATF does not only cover urban services, but also a slew of other costs including the running of schools and health facilities in some local council areas. These costs should be separated out, and added to the health and education sectors (where education is retained as the domain of national government). Second, a very substantial proportion of urban costs are not met from LATF, but from county own revenues. In developing a comprehensive estimate of county needs, it is important that these costs are not forgotten, just because they are not financed from the national budget. 5.20 The cost to start up county administrations should also be factored in. These include building of offices and the installation of county specific infrastructure, for example financial management systems and human resource management systems. There will also be substantial additional costs of new positions, including over 2,000 county assembly members, staff of county assemblies, governors and 54 Devolution without disruption – Pathways to a successful new Kenya deputy governors and their offices, nearly 500 county executive members, 250 county Public Service Board members, and 500 or more senior staff to lead county administrations. 5.21 The Constitution foreshadows that in the early transition years, functions will be phased in asymmetrically rather than in one fell swoop. This being the case, the cost of devolved government in the first transition year could be less, as the national government continues to run selected county functions and channels resources to county capacity building. However, funding for many of the additional costs (like those of county politicians, executives and staff) is likely to be needed from day one. 5.22 Finally, and most importantly, an approach must be developed to cost the expansion of inequitably distributed service networks. Historical approaches to costing merely capture the existing spending, and do not account for inequity in service delivery. Since eliminating this inequity is one of the main objectives of the new Constitution, future approaches to costing should seek a more equitable basis. Counties which have been traditionally neglected should have the benefit of having increased numbers of facilities, and therefore, increased operating costs including staff. Costing county needs therefore will have to take into account an expansion of service delivery, particularly in lagging counties. At the same time, the answers will need to be mapped to a reasonably affordable level of resourcing. Finding the fiscal space to move beyond historical spending 5.23 For 2012/13, Treasury and the CRA have significantly different estimates for the total cost of devolved government―KES 160 billion and KES 255 billion respectively. The Budget Policy Statement (BPS) for the 2012/13 budget indicates that Treasury has costed total devolved functions (including projects under CDF, LATF and RMLF) as amounting to KES 154 billion and added KES 6 billion to cover county administrative costs. Figures in the 2012/13 budget estimates, show a lower amount for total county needs of KES 149 billion,4 inclusive of CDF, LATF and RMLF which are clearly maintained as separate amounts. The CRA on the other hand recommends a county equitable share of KES 203 billion equal to 33 percent of revenues in 2010/11, plus additional conditional grants in form of LATF, CDF and RMLF (KES 52 billion in the budget). The CRA equitable share figure of KES 203 billion is based on the following: (i) an initial estimate from Treasury of KES 148 billion as the total cost of devolved functions; (ii) county administrative costs of KES 36.3 billion; and, (iii) a 10 percent contingency cost of KES 18.4 billion. The challenge going forward now is how to reconcile these two recommendations. 5.24 Finding the fiscal space for the additional KES 106 billion recommended by the CRA ontop of the funding currently allocated to devolved functions will be difficult. The additional amounts recommended by the CRA include both a more expansive costing of existing functions (to compensate for the likelihood of under costing by ministries), and cost estimates relating to the new overhead costs of county government. The definition of fiscal space is: “room in a government´s budget that allows it to provide resources for a desired purpose, without jeopardizing the sustainability of its financial position, or the stability of the economy�. It can be created in a number of ways: (i) reprioritizing expenditure; (ii) boosting efficiency; (iii) raising additional revenue; and, (iv) increasing borrowing. 4 This figure also includes expenditures covered by donor funding. 55 Chapter 5: Determining how much Counties need 5.25 The most promising way to create additional fiscal space may be by limiting growth in the national budget over the medium-term, and allowing growth in revenue to finance an increasing county share. Given wage pressures, it seems highly unlikely that reprioritization by cutting national recurrent budget is feasible in the short term. Efficiency gains could be realized from the consolidation of ministries, but these savings might prove insufficient. Increasing borrowing to finance recurrent expenditure will risk an increase in the structural budget deficit. The more feasible alternative would be, to find fiscal space from increasing revenues rather than cutting costs, although this necessitates taking a longer term view. With a constant rate of GDP growth of 5 percent, an estimated KES 75 billion in national revenue can be raised each year, adding significant fiscal space. Still, this would be insufficient to immediately, simultaneously and fully fund: (i) a continuation of service delivery at existing levels in leading regions; (ii) an expansion of services and investments in underserved counties; (iii) the cost of setting up the new devolved institutions; and, (iv) the adjustment implied in civil service transition. Therefore, while the CRA recommendation makes sense from a normative point of view, it may only be consistent with fiscal balance over the medium term, and tough tradeoffs will need to be considered in the short run. Implications for design of intergovernmental financing arrangements 5.26 Once total county needs are agreed upon, the next question is how they will be financed, and with what mix of instruments. Revenue sharing arrangements should ensure that counties have sufficient resources to meet the costs of delivering functions from all sources of revenue, and that these resources continue to be sufficient as county needs change over time. However, as is highlighted above, determining how much counties should receive in total cannot be decided in isolation from the total resource envelope, or the macro-economic implications of different financing options. 5.27 A secondary question concerns the size of the equitable share and other (conditional) transfers. Programs to be retained, but funded conditionally, would not form part of the equitable share going to counties. The 15 percent (minimum) equitable share of revenues is an unconditional transfer of resources to the counties and no conditions can be applied to it. It follows that any funding going to the counties that are subject to conditions, must be in addition to the equitable share transfer. For example, if the constituency development funds continue to be earmarked for constituency-focused projects, this funding must necessarily be conditional and over and above the equitable share to counties. This final point emphasizes the need to consider intergovernmental financing arrangements in totality―as a whole system―rather than focusing only on its constituent parts, like the equitable share. 56 Devolution without disruption – Pathways to a successful new Kenya CHAPTER 5: KEY INSIGHTS / RECOMMENDATIONS. By prescribing a minimum guaranteed transfer to counties of at least 15 percent of national revenue, the Constitution has not pre-empted the requirement of aggregate costing of county needs. Implementing devolved government necessarily involves additional costs―both to run county governments, and to expand services networks, so that access is more equitable. However, decisions about quantifying these needs cannot be taken in isolation from the macro-economic considerations involved in creating the “fiscal space� to finance them. Functions that could possibly be devolved would require significantly more than the 15 percent floor, if they were to be devolved in one go, when counties come into existence. There is a large gap between the Treasury estimates of total county needs in the 2012/13 budget, and those of the CRA. A key question is how to reconcile these two figures. Costing aggregate county needs is not the same thing as costing the equitable share. The size of the equitable share should be determined as part of the overall design of transfers. Functions financed through earmarked funds. *** Kenya should view the historical costing approach as a place to start determining the vertical share, rather than a final methodology. Costing should evolve over time, aiming to address the entrenched historical marginalization reflected in the historical costing. A complete and transparent function assignment process must be undertaken first before costing can be finalized. In the immediate a decision should be made on assignment of a number of key ‘big ticket’ functions, which will greatly influence the estimation of total county needs (including provincial hospitals, and construction of new roads). Over time, a more detailed methodology for measuring county needs should developed. It should include the realization of equalization objectives, and incorporation of costs not covered by a historical expenditure-based costing The calculation of the equitable share should be made in conjunction with the determination of the conditional grants. In practice, the equitable share may have to be decided, by subtracting conditional grants from the total county vertical share, if the decision is taken to retain the big earmarked funds―CDF, LATF and RMLF. 57 CHAP TE R SIX Solving the intergovernmental division of revenue jigsaw 6.1 There will be five options for resourcing counties, and it is important to stress that the equitable share is only one instrument among others to meet total county needs. Yet, much of the public debate around the financing of county functions has focused on the equitable share or “15 percent� and―to a lesser extent―on county own resources. This truncated view is problematic for a number of reasons. First, the recommendations of Commission on Revenue Allocation (CRA) on the equitable share (both in terms of its total size and of its horizontal split) need to take into account the scope and design of other sources of financing (additional transfers and own revenues). Second, the constitutionally guaranteed nature of the equitable share means that the ‘space’ for conditional transfers will be greatly affected by the size of the equitable share, potentially above and beyond the 15 percent floor. Lastly, government transfers may lead to distorted outcomes―including for equalization―if they do not factor-in own and external sources of revenues available to counties. 6.2 A key objective of the intergovernmental revenue sharing architecture should be to preserve flexibility around the constraints of the equitable share transfer. Because the equitable share transfer is unconditional, it is critical in giving future counties substantial autonomy (see Box 6-1). But it also imposes significant constraints. Because it is untied, the central government will need to rely on additional conditional instruments to correct spillovers (programs run by one county that serve others), to secure funding for critical strategic programs at the local level, or to reward performance. Also, given the relatively crude formula through which the equitable share will be allocated, and the wide range of county circumstances and needs, additional flexibility will also be required to target area- specific interventions or residual needs. 6.3 Given these constraints, there is a case―at least at the start―for limiting the equitable share transfer at or close to 15 percent of national revenue. At the very least, increases of the equitable share transfer above and beyond the constitutionally mandated 15 percent should be decided, once it becomes clear how much other (particularly conditional) financing streams and instruments will absorb, and how much fiscal space remains available. This does not mean total transfers to counties should be limited to 15 percent―far from it―but rather that the size of the equitable share should be determined as part of a process of deciding the total size of all transfers for counties. How will counties be resourced? 6.4 Counties will be funded through five different types of instruments. While the Constitution of Kenya guarantees counties―collectively―an unconditional transfer equivalent to 15 percent of national revenue (last audited), and makes provisions for some counties (or communities) to receive conditional equalization grants, this does not mean that other transfers cannot be considered. Alternative options include (i) increasing the equitable share above and beyond the 15 percent floor; and, (ii) transfers from the national equitable share, either on a conditional or unconditional basis (formula driven or not). In addition, counties will raise own revenues, and may receive direct funding from donors. Some counties may also receive an equalization fund payment (see Figure 6-1). 58 Devolution without disruption – Pathways to a successful new Kenya Box 6-1: Why the equitable share will be unconditional. Conditions cannot be placed on the equitable share granted to counties for three reasons: 1. The revenue that is to be divided, first between the national and county levels of government and, secondly, among the counties, is not national revenue. As the national government had no entitlement to the county share, it has no authority to place direct conditions on the way counties dispose of it. 2. The Constitution intends counties to have a degree of autonomy. This is reflected in: (i) the list of “Objectives of devolution� in Article 174 which includes “accountable exercise of power� (para (a)), “powers of self-governance� (para (c)), and “the right of communities to manage their own affairs� (para (d)); (ii) the substantial institutions set up to govern counties (Chapter 11 generally); and, (iii) the tight constraints on national intervention in county governments (Articles 190 and 192). A substantial portion of the revenue of many counties is likely to come from the equitable share. If this were conditional, the autonomy of counties would be fundamentally undermined. 3. Article 202 expressly allows additional allocations from the national government to counties to be conditional. This demonstrates that the question of conditionality was considered when the Article 202 arrangements were decided, and demands the interpretation that the equitable shares themselves should not be conditional. Conditions can also not be imposed indirectly, for instance, by first dividing national revenue by sector and then allocating the county share within each sector─ by what are called “block grants� in South Africa. Constitutionally, such block grants are conditional grants. Similarly, the equitable share cannot be channeled through national departments. For instance, an approach under which each national department is obliged to allocate an equitable share of its budget to the counties is not constitutional. Finally, the Constitution does not allow different treatment of the “minimum� 15 percent equitable share and any greater amount to which counties are entitled under the constitutional process of equitable sharing. If in applying the criteria in Article 203(1), it emerges that the counties require 45 percent of the revenue raised nationally to fulfill their functions, the Constitution requires that sum to be transferred to counties without attached conditions. Under Article 202(2), “additional allocations� from the national government, whether conditional or unconditional, may supplement a county’s revenue (made up of “own revenue� (i.e., revenue raised by county taxes etc) and a county’s equitable share). Money for additional grants need not be distributed equally to all counties. It might be used to ensure that particular counties deal with cross-border issues (i.e. where a county provides a service that is used by inhabitants of a neighboring county). Rather than expect every county to provide every service (such as perhaps specialized disaster management equipment or abattoirs), there may be situations where counties could share resources. Conditional grants might encourage this. Grants may also be used to persuade counties to do things the way that the national government wishes them to, or to assist counties with infrastructural backlogs. Source: Legal Opinion by Christina Murray. 6.5 Aggregate county funding should be matched to needs as part of a managed process. Estimates suggest that financing the full range of devolved functions―even under a conservative scenario― would require well over 15 percent of national revenues. However, this may not be the case on day one―depending on the pace at which functions are actually transferred (following an asymmetric model); whether major dedicated sector specific funding instruments are converted into conditional grants; and, whether the salaries of staff initially seconded to county governments will be paid out of the national or the county share. Once total needs have been agreed upon, the next step should be to estimate the need for transfers, taking into account own source revenues. Finally, the government should decide the mix of financing instruments, namely unconditional and conditional (capital and recurrent) as part of the budgetary process. 59 Chapter 6: Solving the Intergovernmental Division of Revenue Jigsaw 6.6 The higher the share of county Figure 6-1: Flows of revenues from different sources needs that will be financed through for county governments the equitable share, the lesser the scope for targeted conditional transfers (and vice versa). Figure 6-2 shows four different models for financing county needs. In option 1, total county needs are addressed via the equitable share on a purely unconditional basis (with each county’s share determined via the formula decided by the Senate). In option 2, the equitable share is paid at the minimum level of Source: World Bank (2011b). 15 percent and remaining needs are met through conditional Figure 6-2: Four basic options for �nancing counties. grants. Option 3 is a variation on option 2, where the balance OPTION 1 OPTION 2 above 15 percent is met through a combination of conditional grants and unconditional transfers Conditional grants (although via a different allocation Equitable 15% 15% formula). Finally, in option 4, the share equitable share is set at above 15 Equitable share percent but leaving some room for conditional instruments to fill the gap. OPTION 3 OPTION 4 6.7 Conditional financing will be entirely borne out of the national Conditional Conditional government’s share. A number grants grants of national programs, which are >15% Un-conditional grants currently delivered via earmarked 15% 15% funding streams, are related to Equitable Equitable share share functions that are to be devolved (see Box 6-2). Going forward, the national government will have to Source: World Bank staff analysis. make a choice: (i) shift the cost over to counties (as part of the equitable share transfer), but relinquish control over the allocation of funds; or, (ii) secure the funding (via conditional grants) but absorb the cost. The choice will not be trivial given both the strategic and economic importance of these programs (and hence the national interest in their continuation) and the amounts at stake. Three of these programs alone, the RMLF, the CDF and the LATF amount to KES 52 billion in the 2012/13 budget. If the amount allocated to these funds is added to the base minimum equitable share, the total county share comes to over 23 percent of national revenue. 60 Devolution without disruption – Pathways to a successful new Kenya Box 6-2: Existing earmarked programs that would become conditional grants if continued. Although a 2011 report by the Parliamentary Budget Office identifies a large number of devolved funds in Kenya, there are three that attract very substantial amounts of funding: Local Authorities Transfer Fund (LATF) is established under the Local Authorities Transfer Fund Act. The Act directs that 5 percent of income tax (corporate and personal) should be directed into the Fund, which is statutorily separate from the Consolidated Fund. The fund is administered for the benefit of Local Authorities and allocated according to a formula that favors urban population. The annual contribution to the Fund for 2012/13 is estimated in the Budget as KES 21.5 billion. Road Maintenance Levy Fund (RMLF) is established under the Road Maintenance Levy Fund Act and channels an additional excise on the sale of petroleum products (and some other minor revenues including agricultural cess) into a Fund that is separate from the Consolidated Fund. It is administered by the Kenya Roads Board, which divides the proceeds of between three statutory bodies that look after Kenya’s road maintenance―the Kenya National Highways Authority, the Kenya Rural Roads Board and the Kenya Urban Roads Board. The value of the fund in the 2012/13 year has decreased since 2010/11 when our simulation was done, because the fund revenue is sensitive to fuel prices, which have dropped. The 2012/13 budget forecasts the value of the Fund’s share for rural and urban roads as KES 8.5 billion. Both RMLF and LATF can be identified in the budget each year. They are shown under the recurrent budgets of the Ministry of Roads and Ministry of Local Government respectively, but they are categorized as appropriations-in-aid (AIAs). This means that both expenditure and the corresponding revenue from the funds themselves are shown, effectively cancelling each other out. Constituency Development Fund (CDF) is established under the Constituency Development Fund Act. The purpose of the fund is to finance small grant projects selected and managed by committees within constituencies. To a varying degree in different constituencies, MPs may influence the project selection. The Fund is capitalized annually by an amount equivalent to 2.5 percent of ordinary revenue, but unlike RMLF and LATF, this amount is not legally hypothecated into the fund. Instead, provision is made for CDF in the annual budget each year, under the Development Budget of the Ministry of State for Planning. The amount budgeted for CDF in 2012/13 is KES 21.8 billion. Should they remain earmarked to specific uses, these programs would have to be financed through conditional grants under the new dispensation, and their costs would not be included in the calculation of the county equitable share. Source: World Bank staff analysis. 6.8 The government will need to decide on the future of these programs before deciding the amount of the equitable share. They could be abolished in their current, earmarked/conditional form (by repealing the Acts which provide for their establishment) and the funds shifted to counties. Counties could then decide whether to maintain the programs and fund them out of the equitable share, or not. In this case, the equitable share would need to be more than 15 percent. On the other hand, if the Acts remain in place, and continue to specify the aggregate amount, the formula for distribution and the purposes for which funds can be used, the amount available to fund the equitable share will be less. This may limit the scope to use conditional grants in order to meet the different objectives outlined above, including financing regional services that one county provides to the benefit of other county residents, and safeguarding national priorities like capital infrastructure. 61 Chapter 6: Solving the Intergovernmental Division of Revenue Jigsaw 6.9 While much public debate has focused on the optimal rate of the equitable share, it matters only in conjunction to the base because what really matters is the level of funding. A larger base may mean that the percentage rate needs to be lower. A small base may mean that a higher percentage is needed. Yet there is a lack of clarity as to what exactly the base includes and what year the formula applies to. 6.10 There are some good reasons for maintaining conditional grants. They include; a) compensating counties for programms that service more than one county, b) compensating counties for unique cost burdens, and c) ensuring that national priorities are met (see Box 6-3). Box 6-3: Reasons why conditional grants are desirable in certain circumstances. Compensate counties for programs they provide that serve other counties, to ensure there is no under-provision • The 2012/13 budget indicates that provincial hospitals are assumed to be a devolved function. Kenya has eight provincial hospitals. Almost KES 8 billion is allocated to this function in the 2012/13 budget, an average of around KES 1 billion per county. This is a cost that the other 39 counties do not have to bear, yet their citizens will access the services provided by those 8 provincial hospitals. • In order to ensure that the eight counties do not reduce funding to provincial hospitals (and so lead to under- provision of services), conditional funding of these services should be considered. Compensate counties for unique cost burdens • Mombasa and Lamu are the counties that will host international port facilities. Under the Constitution it seems likely that the may be responsible for providing port health services (including quarantine public health inspection of foreign vessels and their cargo and passengers). The CRA-recommended formula for the equitable share does not take account of unique costs that only a few counties bear. Arguably, these two counties should receive conditional grants to operate these services. • Many countries give special grants or make special cost allowances in calculating transfers, for the cost of providing national capital services. Ensure that national policy priorities are financed • Water supply in many parts of Kenya is crucially dependent on major trunk water infrastructure projects managed by the eight regional water boards. KES 20 billion is allocated to capital water infrastructure projects in the 2012/13 budget. If this funding for this capital infrastructure is divided among the forty-seven counties, and these counties are left to decide how much they each want to contribute to these regional infrastructure projects, the likely result is that capital funding will cease. In some cases projects are likely to be half completed, so this may result in wasting the initial investment as well. • Kenya is committed to meeting the Millennium Development Goals (MDGs) in relation to reduction of malaria and TB and reducing child mortality. Specialised programs providing access to drugs, vaccines and bed nets are the main vehicle for achieving these policy goals. It may be appropriate to finance parts of these programs through conditional grants, to ensure that counties continue to spend an adequate amount on vaccines and other program inputs so that the MDG targets can be achieved. Source: World Bank staff analysis. 62 Devolution without disruption – Pathways to a successful new Kenya Estimating the 15 percent 6.11 Estimating the value of the minimum equitable share is the starting point for thinking about the broader intergovernmental transfers ‘package’. Kenya’s counties enjoy a constitutional guarantee that at least 15 percent of national revenue will be handed over unconditionally, but the definition of the base has not always been clearly agreed upon (see Box 6-4). First, the definition of the universe of national revenues was initially the subject of some debate. Second, although the Constitution is clear that the formula will be applied to the last audited set of accounts, the implied lag in base years does not appear to be always well understood in policy circles (and much less the media or the public who understand the 15 percent as applying to the current year’s budget). 6.12 The ultimate definition of the base may not matter so much, although some choices appear more rational. To the extent that the CRA has the ability to recommend any percentage for the unconditional equitable share above the ceiling of 15 percent, it has the flexibility to determine what combination of base and rate corresponds to the targeted level of unconditional transfer. The CRA act (see Box 6-4), currently the only legal definition of the base, appears sensible to the extent that it includes the main sources of revenue (tax and non-tax revenues), but excludes donor funding, appropriations in aid, domestic borrowing and ‘revenues’ linked to dedicated funding schemes (such as RLMF and LATF). Because the equitable share is guaranteed in terms of a percentage (and not an amount), including highly volatile elements in the base (such as domestic borrowing or donor funding) that would be detrimental to the predictability of transfers to counties. Moreover, it would make little sense to re-channel to counties on an unconditional basis, funds that are collected (internally or externally) with specific purposes in mind, such as donor funds or earmarked funds such as RMLF or LATF. 6.13 The percentage will be applied to past―not current!―revenues. This is because the Constitution (Article 203(3)) stipulates that the minimum percentage “shall be calculated on the basis of the most recent audited accounts of revenues�. Audited accounts are typically available in Kenya with a two year lag. In other words the unconditional share for year “t� will amount to 15 percent (or whatever other share is agreed) of all revenues collected (and audited) in year “t-2�. As a result the amount of the unconditional transfer will almost certainly be lower than 15 percent (or whatever other percentage is chosen) of the current year’s revenues. For the 2012/13 budget year, the Budget Policy Statement proposed allocating KES 160 billion for devolved functions, which amounts to 26 percent of revenues in 2010/11 (the last audited), but 16 percent of the 2012/13 budget (see Figure 6-3). Figure 6-3: The chosen rate will apply to a lagged base. Revenue sharing for 2012/13 based on audited accounts from the year before last 2010/11 2011/12 2012/13 2012/34 Division of Revenue Bill Year to which Base year presented to Parliament revenue sharing applies Source: World Bank staff analysis. 63 Chapter 6: Solving the Intergovernmental Division of Revenue Jigsaw Box 6-4: Calculating the basis for equitable sharing: 15 percent of what? The Constitutional provisions on how to calculate the revenue base for the equitable share, referred to as ‘national revenue’, have caused some confusion because it is not entirely clear which revenues should be included. In particular: - Article 202(1) refers to “revenue raised nationally� (the phrasing here implies a consideration for revenues raised in the whole nation rather than by one level of government). - Article 203 refers to “all revenue collected by the national government� (this might be interpreted to include donor funds, funds from domestic borrowing and AIAs). - Article 206 describes the money that should go into the consolidated fund as “all money raised or received by or on behalf of the national government� (again this might be interpreted to include donor funds, funds from domestic borrowing and AIAs). During 2011 this ambiguity was exacerbated when two alternative definitions of national revenue were put forward in the Commission on Revenue Allocation (CRA) Act and the Intergovernmental Fiscal Relations Bill. The latter has now been dropped and the definition included in the CRA Act (Section 2(1)) is now accepted, namely that: - “’revenue’ means all taxes imposed by the national government under Article 209 of the Constitution and any other revenue (including investment income) that may be authorized by an Act of Parliament, but excludes revenues referred to under Articles 209(4) and 206(1)(a)(b) of the Constitution.� The revenues excluded from the CRA Act definition referred to in Article 209(4) relate to fees and charges of national and county governments for services and those in Article 206(1) relate to: (a) money excluded from the Consolidated Fund by an Act of Parliament and payable into another public fund (e.g. RMLF and LATF); and, (b) money retained by the State organ that received it for the purpose of defraying its expenses (e.g. AIAs). For the 2009/10 audited accounts, this results in a total shareable pool of national revenue of KES 507 billion, of which the constitutional bare minimum county equitable share of 15 percent is KES 76 billion for 2011/12 budget year (see table below for detailed calculation). Using the same approach for the 2010/11 audited accounts, total national revenue is KES 608 billion, giving a minimum 15 percent county equitable share of KES 91 billion for 2012/13 budget year. See Figure 6-3 for an explanation of why there is a two-year gap between the year of the audited accounts and the budget year. Value (KES billion) of actual receipts in Included in CRA Act Definition of Revenue items 2009/10 as per audited accounts (GoK, 2011) ‘National Revenue’? 1. Donor Funds 21 No 2. Domestic Borrowing 134 No 3. Appropriations in Aid (AIA) 41 No 4. Non-tax revenue 18 Yes 5. Tax Revenue 489 Yes 6. RMLF & PDL Fund 25 No 7. LATF 10 No 8. Local Authorities Revenue 17 No TOTAL KES 755 billion KES 507 billion Source: Constitution of Kenya (2010), CRA Act (2011), GoK (2011), GoK (2012) and MoF (2012). 6.14 The CRA has recently made recommendations on a sharing formula and recommended counties should receive KES 203 billion for 2012/13. If the CRA’s recommendations relate just to the equitable share, this will substantially constrain the scope for conditional grants. The risk is that allocating the bulk of the resources to be transferred (to meet aggregate county needs) through the equitable share, will be excessively constraining in terms of: (i) the ability of the centre to guarantee that specific programs (such as urban services) or types of expenditures (such as capital investment) will be maintained; and, (ii) the criteria used to allocate resources across counties. 64 Devolution without disruption – Pathways to a successful new Kenya CHAPTER 6: KEY INSIGHTS / RECOMMENDATIONS. Although public debate has focused almost exclusively on the equitable share transfers to counties, there is a broad array of additional instruments for resourcing devolved functions including county own-source revenues, other unconditional and conditional transfers and external sources. Addressing aggregate county needs should take into account all of these sources to maximize efficiency and avoid fiscal stress, at the center and potential mis-matches between revenues and needs at the local level. The size of the equitable share transfer―which is unconditional―will determine the fiscal space left for conditional instruments. The base applicable to the equitable share transfer is subject to debate and confusion, and therefore requires clarification and consensus. *** Once estimates of aggregate county needs will have been worked out, resourcing them should consider all possible sources of revenue available to them, and the full range of instruments that the center can mobilize depending on the objectives sought. The tradeoff between unconditional and conditional funding deserves particular attention, as the scope for one will constrain the use of the other. The GoK should decide the fate of existing earmarked programs which, if maintained, will constitute conditional grants to be resourced above and beyond the equitable share. When thinking about the level of the equitable share (if above 15 percent) it may make sense to define a nominal target to reach, via a combination of rate and base (to be defined clearly). 65 PART III Equity across Kenya’s diverse counties CHAP TE R SE V E N Funding from within: County own revenue powers and capacity 7.1 County own revenues are of crucial importance to the success and nature of the devolution process for at least two reasons. Firstly, in line with the principle the funding must follow function, estimates of intergovernmental resourcing will be incomplete if they do not take into account counties’ own revenues, which may result in a mismatch between functions and funding. Secondly, devolution is not simply about how much funding counties will receive: the type of funding for county governments will also be crucially important. While heavy reliance on intergovernmental transfers may be advantageous for overall macroeconomic stability, there are strong arguments for providing counties with additional revenue sources, in particular, because this can help to promote a stronger accountability relationship between county governments and their citizens, whose tax money they would be collecting and spending. 7.2 This chapter examines the likely impact of the constitutionally granted revenue raising powers on the counties, and considers what options may exist for expanding county own revenue powers. The constitution grants counties very limited own revenue raising powers, while also granting the national government the power to pass legislation granting additional revenue raising powers to county governments. Without additional legislation, the new counties will be highly fiscally dependent on the national government. If a consensus emerges that counties are too dependent on national government, then this chapter also offers a framework together with some suggestions, on possible options to consider in expanding counties’ own revenue raising powers. The crucial relationship between own source revenues and subnational autonomy was clearly understood by those leading the centralization of government in Kenya, with the removal of the Graduated Personal Tax from municipalities in 1973. The question is, will those leading Kenya’s devolution recognize and correct this imbalance? Or will counties be heavily reliant on intergovernmental transfers, lacking genuine fiscal autonomy? Legal framework for own revenue powers of county governments 7.3 The vast majority of revenue collection and expenditure in Kenya is undertaken by the central government, with total local authority revenues (including transfers from central government) accounting for 3.9 percent of general government spending in 2009/10. General government spending in 2009/10 amounted to around KES 754 billion, of which KES 725 billion was spent by central government (96.1 percent). There is therefore very limited financial autonomy for Kenya’s local authorities, which are heavily reliant on central government for funding.1 7. 4 Local authorities in Kenya currently have access to two major revenue sources: “shared� revenues collected by central government, a portion of which is transferred to local authorities, and; own source revenues administered by the local authorities themselves. Shared revenues with central government, compriseof a share of personal income tax through Local Authorities Transfer Fund (LATF) 1 There are 175 local authorities in Kenya, comprising Nairobi City Council, 45 Municipal Councils, 67 County Councils (predominantly rural), and 62 Town Councils. 68 Devolution without disruption – Pathways to a successful new Kenya and―until the 2010/11 financial year―a share of the fuel levy through the Road Maintenance Levy Funds (RMLF). Own source revenues include general revenues and fees that are levied and collected locally (levied by most local authorities), and special service-linked fees and charges, such as for water and nature reserves (levied by only a few local authorities). The legal framework for these revenue sources is provided by a set of national laws (summarised in Table 7-1) together with by-laws developed by the local authorities themselves. Table 7-1: Legal basis for Local Authorities’ revenue collection. Revenue Source Legal Instrument Shared revenues Local Authorities Transfer Fund (LATF) LATF Act (No. 8 of 1998), LATF Regulations (1999), administrative circulars from Ministry of Finance & Ministry of Local Government *Road Maintenance Levy Fund (RMLF) Road Maintenance Levy Fund Act, Kenya Roads Act (2007) Own source revenues Property taxes collected in the form of land Local Government Act (Cap. 265), 1998 Revised, Rating Act(Cap. rates 267), 1986, and Valuation of Rating Act(Cap. 266), 1956 Single Business Permit (SBP) Local Government Act Cap. 265 (1998 Revised) Note: *RMLF ceased to flow to Local Authorities in FY2010/11 under the Kenya Roads Act (2007). Source: World Bank staff analysis. 7.5 The LATF Act provides for 5 percent of national income tax to be distributed to local authorities, in order to improve service delivery, financial management and accountability, and eliminate outstanding debts. It has been the primary source of funding for all 175 local authorities since its introduction in 1999/2000, to replace the Local Authority Service Charge (LASC). Funds are distributed according to a formula in a predictable and transparent manner, with release conditions including: submission of an annual budget, financial statements, debt reduction plan, revenue enhancement plan, and a participatory service delivery action plan. It is these LATF conditions which have allowed systematic collection of data on local authority revenues by the Ministry of Local Government. Once funds are released, local authorities are free to spend them like general “block� grants, in accordance with their annual budgets.2 7.6 The Road Maintenance Act was amended in 1997/98 so that a share of the RMLF was allocated to a local authority account to finance local authority roads. The RMLF Advisory Committee (chaired by the Ministry of Local Government) advises the Ministry of Finance (MoF) on the allocation of funds in the local authority account. Funds are largely spent through the Ministry of Local Government, on behalf of local authorities rather than being transferred directly to the authorities. However, following the implementation of the Kenya Roads Act (2007), since 2010/11 RMLF funding no longer flows via local authorities―it is now administered directly by the Kenya Urban Roads Authority (KURA), which is responsible for urban and municipal roads. 7.7 In practice, local authorities have no authority over either the tax rate or the tax base of these ‘shared revenues’, and they are effectively intergovernmental transfers. Since the RMLF flows were removed entirely from local authorities, they have become even less important. This section will focus primarily on the role of own source revenues and possible lessons these provide for building a revenue base for the new county governments. 2 World Bank (2002). 69 Chapter 7: Funding from within: County own revenue powers and capacity 7.8 The legal basis for the collection of local authorities’ own source revenues is set out in a number of laws, which give local authorities the right to raise revenue only with approval from the Minister for Local Government. This means that the local authority tax base is, to a large extent, influenced by central government’s discretionary decisions as to whether to grant or withhold fiscal authority―see Table 7-6 and associated discussion below for a typology of own revenues measured by extent of local autonomy.3 The main sources by amount are listed in Table 7-2. In order of importance, the major sources of own source revenues are: property rates, “other� revenues, the Single Business Permit (SBP), vehicle parking, market fees, cess receipts, game park fees, house rents, contributions in lieu of rates, plot rents and water and sewerage fees. Table 7-2: Main revenue sources for Local Authorities, 2009/10. Percent Average Percent Revenue Sources 2009/10 annual growth Contribution since 2004/05 LATF 10,398,595,830 35.3 22.3 RMLF 1,474,885,113 5.0 29.9 Other Government 304,591,496 1.0 n/a Sub-Total Revenues from Central Government 12,178,072,439 41.3 23.2 Property Rates 3,552,484,115 12.1 14.7 Others 3,164,628,970 10.7 33.3 SBP 2,900,685,813 9.8 11.7 Vehicle Parking 2,336,586,263 7.9 19.4 Markets Fees 1,195,455,613 4.1 11.6 Cess Receipts (agricultural output tax) 1,161,467,378 3.9 19.4 Game Park Fees 1,135,002,537 3.9 15.2 House Rents 614,291,972 2.1 61.6 CILOR 583,767,298 2.0 18.0 Plot Rents 370,801,549 1.3 20.1 Water & Sewerage Fees 264,606,555 0.9 -6.0 Sub-Total Local Own Revenues 17,279,778,063 58.7 15.3 Total Central & Local Revenues 29,457,850,502 100.0 18.1 Source: LATF Reports, 2004/05 and 2009/10. 7.9 Property taxes―known as “rates� in Kenya―are the most important source of local authorities’ own source revenues. Despite this, Kenya is under collecting property tax, judging by international standards. As Table 7-3 illustrates, property taxes only accounted for around 12.1 percent of total local government finance, 0.75 percent of total government taxes, and 0.14 percent of GDP in 2009/10. This comparess unfavourably with other developing countries that typically collect up to 40 percent of total local finance, 2 percent of total government tax, and 0.5 percent of GDP in property taxes.4 This suggests that there is significant room for improved revenue collection for counties through reforms to the collection of property rates in Kenya. Understanding how this might be possible requires taking a closer look at how rates presently work in practice. 3 World Bank (2008). 4 Kelly (2000). 70 Devolution without disruption – Pathways to a successful new Kenya Table 7-3: Comparative under collection of property taxes in Kenya, 2009/10. International percent Developing countries percent Kenya (2009/10) percent Percent of local government finance 40 – 80 Max. 40 12.1 Percent of total government taxes 2–4 2 0.75 Percent of GDP 0.5 – 3.0 0.5 0.14 Sources: Kelly (2000), LATF Report 2009/10, QEBR 2nd Quarter 2011/12. 7.10 Property rates are not uniformly important for all local authorities: there is important variation in collections across different types of authorities. Rates are a much more important revenue source for municipalities, accounting for 25 percent of their own source revenues in 2009/10, compared to around 6–7 percent for town and city councils. This is because the valuation rolls capture the larger, more valuable urban tax base for properties, within the gazetted area of local authorities. 7.11 Rates can be legally applied as taxes on either land or land improvements, although in practice, they are applied only to land. Most local authorities exclude ‘freehold’ land (agricultural land less than 12 acres), and most private land in area rating rolls, and public land (central government or council trust land) that is yet to be ‘registered’ is also excluded.5 While the Rating Act allows a large degree of flexibility in the assessment approach, in practice, all local authorities are using either area rating or valuation rating. With the exception of the valuation rolls (which most local authorities rely on the Ministry of Local Government to maintain), local authorities are responsible for all aspects of property tax administration. While there is an adequate legal framework for property taxation in Kenya, the main barrier to better rates collections appears to lie in ineffective administration.6 7.12 It is hard to determine what makes up the “other� category which comprises of 18 percent of local revenues, and 11 percent of total local authority resources. Mboga estimates that in 2006/07 it comprised of advertising fees and planning approvals, income from commercial activities, public health and wastage charges, administration charges, fines and interest, and grants (see Table 7-4).7 However, “miscellaneous� still accounts for the majority of other revenues (and 7.5 percent of all local revenues), suggesting that data collection on local authority revenues could be improved. Table 7-4: Composition of ‘other’ revenues, 2006/07. Description KES millions % contribution to total local revenues Sub-Total Local Own Revenues 10,847.9 100 Others 2,156.4 20 Miscellaneous 818.8 6 Advertising fees and planning approvals 402.4 4 Income from commercial activities 330.5 3 Public health and wastage charges 214.4 2 Administration charges, fines and interest 175.3 2 EU Poverty Reduction Fund Grant 138.5 1 Other Grants 76.5 1 Source: Mboga (2009). 5 Kelly (2002). 6 Ibid. 7 Mboga (2009). 71 Chapter 7: Funding from within: County own revenue powers and capacity 7.13 The Single Business Permit (SBP)―the third most important own source revenue for local authorities ―was introduced in all local authorities in 1999, to address the proliferation of business licenses in Kenya. Previously, the license fee schedule was determined for each municipality through negotiation with central government, but fees set for different businesses were not linked to their relative ability to pay. There were multiple and time-consuming preconditions which had to be met in order to obtain a licence, with businesses often being required to hold multiple licenses. The Local Government Act (Cap. 265) was therefore amended to state that: “A single business permit means a permit issued in respect of a class of business activities in lieu of the separate licenses which would otherwise require to be issued in respect of each activity. (Cap 265, Section 2 revised)�. The SBP aims to simplify the process of business licencing to be more transparent, introduce an element of local accountability (with each local authority choosing appropriate rates from a ‘pre-approved’ fee schedule) and improve revenue generation.8 Introduced through an amendment to the Local Government Act, it has been incrementally adopted nationwide, and is intended as a requirement for all businesses within a local authority’s jurisdiction. SBPs are issued for a one-year period, and must be renewed each year. 7.14 Game park fees only accounted for 3.9 percent of aggregate local authority resources in 2009/10, but this masks the fact that they are heavily concentrated in a few local authorities and are therefore very important in these areas. This explains why they accounted for 30 percent of County Council own source revenues in 2009/10―their single largest source of revenue. In fact, game park fees are heavily concentrated in only four County Councils: Isiolo (which will be in Isiolo County, with revenue from Buffalo Springs and Shaba National Reserves), Samburu (Samburu county, with revenue from Samburu National Reserve), Narok and Transmara (both in Narok County, with revenue from Masai Mara National Reserve). 7.15 Water and sewerage fees were Figure 7-1: Water and sewerage fees, once major Local Authority once the most important source revenues, have fallen dramatically. 2,000 25 of subnational revenue for local 1,800 authorities, but have plummeted 1,600 20 in recent years. In 2001/02 water 1,400 and sewerage fees were the most 1,200 15 KES millions Percent important single source of local 1,000 800 10 authority revenues (i.e. excluding 600 transfers from central government 400 5 such as LATF), accounting for 200 23.5 percent of the total (KES 1.9 0 0 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09 2009/10 billion). As Figure 7-1 illustrates, Total Reveue Collected (KES) they then underwent a dramatic Percent of Subnational Revenues (Excl. Revenues from Central Government) decline, falling to 1.5 percent Source: LATF Annual Reports, 2004/05 – 2009/10. of local authority revenues in 2009/10 (KES 0.3 billion). This decline was a direct result of the implementation of the Water Act (2002), which restructured the water sector, creating Water Service Providers as autonomous limited liability corporations, with water and sewerage revenues ring fenced from local authorities. Clearly there will be a strong temptation for cash strapped counties to seek to control this revenue stream. See discussion below. 8 FIAS/World Bank (2011). 72 Devolution without disruption – Pathways to a successful new Kenya 7.16 Local Authorities are not the sole subnational revenue collectors in Kenya: deconcentrated units of central government within districts also collect fees and charges, many of which are likely to accrue to the new county governments, once devolved functions are transferred. Collecting comprehensive data on the extent of these revenues is extremely difficult because a large, and largely unknown, proportion is not captured in the Audited Appropriation Accounts compiled by KENAO. In particular, user fee income from hospitals, health centres and dispensaries is not recorded―they are “off budget� revenues.9 However, estimates have been undertaken for the size of the collections made in the health sector, and these suggest that such revenues could be an extremely important source of funding for county governments (see Table 7-1). In fact, estimated user fees in 2008/09 were larger than any other single source of local authority revenues (for example, KES 4.15 billion compared to KES 3.22 billion for property rates for example). Even if some or all of the revenues from ‘hospitals’ remain with the national government, health user fees from lower level health facilities will be a significant source of revenue for county governments. Table 7-5: Off budget estimates of user fees collected at public health institutions (KES billions). 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09 Hospitals … … … 1.57 2.35 2.35 Others 1.00 1.10 1.47 1.42 1.57 1.80 Total estimated user fees 1.00 1.10 1.47 2.99 3.92 4.15 Total health spending 18.04 22.54 26.97 38.05 42.15 84.58 Percent financed by user fees 5.5 4.9 5.5 7.9 9.3 4.9 Note: “…� indicates data not available. Source: Ministry of State for Planning, National Development and Vision 2030 (2010; p. 81). Estimating capacity from current local authority collections 7.17 Kenya’s devolution will lead to an increased financial reliance of subnational government on the national government. Presently, local authorities, on average, finance around 59 percent of their spending from own source revenues, with LATF transfers accounting for the remaining 41 percent. Under devolution, assuming that only the constitutional bare minimum of 15.5 percent of national revenue is transferred to the counties,and that counties continue to collect the same revenues that local authorities do, own source revenues will finance 17 percent of expenditures while transfers will finance 83 percent of expenditures. If the equitable share is set above 15.5 percent, or if the national government provides additional transfers over and above the equitable share, this financial reliance will increase further. By comparison, subnational governments in China (57 percent), Colombia (40 percent) and Kosovo (21 percent) all receive a greater share of their resources from own source revenues. By East African standards, Kenya’s counties will be relatively financially independent however: both Tanzania (6.7 percent) and Uganda (4.3 percent) are very heavily reliant on transfers from their national governments, to finance subnational spending (see Figure 7-2). 7.18 However, the bigger story is that total subnational resources (both own source revenues and transfers) as a percentage of total government spending―even after devolution―may be lower than a number of other unitary states (Bolivia, China, Colombia, Indonesia, Kosovo, Tanzania and Uganda). This is illustrated in Figure 7-3, which shows that subnational revenue as a percentage of total general 9 Ministry of State for Planning, National Development and Vision 2030 (2010). 73 Chapter 7: Funding from within: County own revenue powers and capacity government spending―assuming Figure 7-2: Kenya will become increasingly reliant on intergovernmental the equitable share is set at 15.5 transfers to �nance subnational spending under devolution. 100 percent of national revenue― 90 will rise after devolution from 3.9 80 Percent total Subnational Revenue percent to 13.4 percent of general 70 60 government spending. This will still 50 be relatively low, with Tanzania 40 (17.6 percent), and Uganda (25.6 30 20 percent) raising more revenue 10 at local level as a percentage of 0 general government spending. (2009/10) (2009/10) (2005) (2007) (2006) (2006) (2008) (2006/07) (2008/09) 15.5% Kenya Simultion Bolivia China Colombia Indonesia Kosovo Tanzania Uganda Kenya-Existing Of course, this share could still risefurther in Kenya, if conditional Own Revenue Transfers from Central Government Source: IMF (2009) and World Bank staff calculations. grants are provided for over and above the equitable share, or if the equitable share is set above the Figure 7-3: Subnational resources as a percentageof total general government spending - international comparison. constitutional bare minimum of 90 15 percent. Kenya’s local revenues Percent of general government spending 80 expressed as a share of GDP (4.1 70 percent) would also be relatively 60 50 low compared to other countries 40 (see Figure 7-4) although quite 30 close to Tanzania (4.4 percent) and 20 Uganda (4.2 percent). 10 0 Kenya - Kenya - Bolivia China Colombia Indonesia Kosovo Tanzania Uganda 7.19 This aggregate level view masks Existing 15.5% (2005) (2007) (2006) (2006) (2008) (2006/07) (2008/09) (2009/10) Simultion important differences between (2009/10) counties. Based on existing data Own Revenue Transfers from Central Government Source: IMF (2009) and World Bank staff analysis. on local authority revenues and LATF transfers, most counties will be heavily dependent on Figure 7-4: Subnational resources as a percentage of GDP - international comparison. intergovernmental transfers. As 16 Figure 7-5 illustrates, 28 of the 14 forty-seven counties currently 12 receive over 50percent of their Percent of GDP 10 resources from transfers (LATF and 8 RMLF), with Nairobi and a group of 6 counties with County Councils that 4 are heavily reliant on game park 2 0 fees (Samburu, Isiolo and Narok) all Kenya - Kenya - Bolivia China Colombia Indonesia Kosovo Tanzania Uganda at the upper end of the spectrum, Existing (2009/10) 15.5% (2005) Simultion (2007) (2006) (2006) (2008) (2006/07) in terms of fiscal independence (2009/10) Own Revenue Transfers from Central Government with 70 percent or more of their Source: IMF (2009) and World Bank staff analysis. funding coming from own source revenues. 74 75 Percent Percent 0 100 10 20 30 40 50 60 70 80 90 0 10 20 30 40 50 60 70 80 90 100 Tharaka Nyamira Turkana Garissa Mandera Turkana Wajir Wajir Nyamira Pokot Tana River Mandera Lamu Makueni Pokot Homa Bay Homa Bay Siaya Marakwet Vihiga Marsabit Nandi Vihiga Tana River Garissa Kisii Bomet Kakamega Kwale Bomet Makueni Bungoma Nandi Devolution without disruption – Pathways to a successful new Kenya Migori Trans Nzoia Trans Nzoia Baringo Marakwet Transfers Samburu Tharaka Kirinyaga Baringo Transfers Bungoma Kitui Isiolo Meru Kitui Kwale Siaya Kisumu Nyandarua Lamu Source: LATF report (2009/10). Own Revenue Own Revenue Meru Busia Kili Marsabit Busia Nyandarua Migori Kili Murang'a Nakuru Source: LAFT report (2009/10) and World Bank staff analysis. Taita Taveta Kirinyaga Laikipia receive less than half of its resources from transfers. Murang'a Kericho Kericho Kisumu Kajiado Embu Mombasa Kisii Nyeri Nyeri Uasin Gishu Kakamega Embu Kajiado Taita Taveta Figure 7-6: Vertical imbalance--applying the CRA formula to 2009/10 data, only Nairobi would Nakuru Figure 7-5: 28 of the 47 counties relied on transfers for over half of their resources in 2009/10. Machakos Uasin Gishu Kiambu Kiambu Laikipia Narok Nairobi Mombasa Samburu Machakos Isiolo Nairobi Narok Chapter 7: Funding from within: County own revenue powers and capacity 7.20 A simulation of post devolution transfers suggests that there will be a substantial ‘vertical imbalance’ under devolution―i.e. counties will be even more dependent on fiscal transfers than they are at present. As Figure 7-6 illustrates, if we apply the proposed CRA formula to 2009/10 data, only Nairobi would receive less than half of their resources from transfers of the equitable share (assuming that the total equitable share is set at 15.5 percent of national revenue and that they do not receive conditional transfers). Without efforts to improve county tax collection or to expand county tax bases compared to practices under the local authorities, the counties will be overwhelmingly reliant on transfers from central government, to finance their budgets. Getting the legal framework right for own revenues 7.21 Kenya’s devolution model involves a large disparity between the expenditures and revenues assigned to county governments. While a significant portion of expenditures are assigned to county governments, under the Article 209(3) of the constitution, counties are only formally assigned revenue raising responsibility for property rates, entertainment taxes, and any other tax that a county is authorized to impose by an Act of Parliament. Schedule 4 also assigns county governments the responsibility for trade development and regulation, including markets and trade licenses (excluding regulation of professions), which it seems reasonable to interpret as providing a constitutional basis for Single Business Permit collection to transition from local authorities to county governments. 7.22 Presently, these constitutional bases for county tax collection have not been converted into legal instruments, to enable counties to collect revenues from day one after the elections. Local authorities, currently responsible for revenue collection, are set to be abolished under the County Governments Act, ass.134(1) (Repeal of Cap.265) provides for the repeal of The Local Government Act: “(1) The Local Government Act is repealed upon the final announcement of all the results of the first elections held under the Constitution�. The Urban Areas and Cities Act (UACA) provides that the transitional implications of abolishing local authorities should be covered in other laws10, however, the legal instrument envisioned in the UACA has yet to be drafted,and until it is,there will be no legal instrument for the continued existence of local authorities, under county control during the transition period. 7.23 The most important legal vacuum is the lack of legal authority for county governments to collect county revenues on day one after the elections. The Constitution assigns the counties the power to collect property taxes, entertainment taxes, as well as imposing charges for the services they provide (Art. 209), but they will require legal authority to do so. It will be extremely important to ensure that there is a legal instrument in place to allow counties to continue collecting the revenues and charges that local authorities are currently collecting; one day after the elections, and before county assemblies can pass their own legislation. Since the rights and obligations of individual taxpayers are involved, it is important to adopt a “belt and braces� approach that is not vulnerable to legal challenge. A challenge to county tax collecting powers would undermine the authority and effectiveness of county governments. 10 Staff are to be “deployed as may be provided by law� (UACA s.57), contracts will “continue in force and … vested in a body established by law� (UACA s.58), and assets and liabilities will be “dealt with as provided by law� (UACA s.55). 76 Devolution without disruption – Pathways to a successful new Kenya 7.24 A new legal instrument is required to allow counties to collect revenues immediately after the elections, through local authorities’ existing staff and systems. The County Governments Act assigns all responsibility for drafting legislation to replace the Local Government Act to the Transition Authority (TA), which should therefore take the lead in coordinating the drafting for a legal instrument to address this. For example, a “County Governments Taxation and Business Regulation Act� could provide clear authorization for county governments to collect taxes and fees constitutionally assigned to them. The Act could explicitly contain a ‘sunset clause’ providing that it is only in force in any given county until such time as the county assembly passes its own taxation laws and excludes the operation of the national law. 7.25 If the very limited time available does not permit a comprehensive Act, a series of amendments to existing Acts could address this. In order to allow counties to collect revenues in place of the existing authorities, the following amendments could be introduced: • The Rating Act (Cap. 267) empowers “rating authorities� to collect rates. “Rating Authorities� are defined as municipal councils, county councils and town council. This Act should be amended to empower county governments to collect rates. This amendment could also clarify: (i) Whether a uniform rate should be applied across the country, or should county governments be able to choose their own rate? (ii) If rates are not to be uniform, should there be any limits (maximum percentage)? (iii) How the existing requirement for the Minster to approve the rate should be revised under counties? (iv) Whether there should continue to be different basis for rates (site value rate and area rate)? • Single business permits are authorized under the Local Government Act, which is being repealed. There should be a short Act to allow SBPs to continue being issued by county governments. The applicable fees are fixed with reference to whether a municipal council, a town council or a county council imposes the tax. This will need to be changed. • The Entertainments Tax Act (Cap. 479) empowers district commissioners to collect entertainments tax. This should be amended to authorize county governments to collect the tax. The law could also provide that district authorities hand over past tax records to county governments. • A hotel accommodation tax is provided for in the Hotel Accommodation Act (Cap 478) that could also be assigned to county governments. This is currently a national government tax. Rates should be revised at the same time. • Other fees and charges imposed by local councils include parking fees and market fees, and are generally imposed under council by-laws. Section 56 of the UACA says council by-laws are deemed to have been “given, issued or made by the boards established under this Act…� (emphasis added). However, these revenue sources belong to county governments. Under Section 20(1)(m) boards can collect rates, taxes, levies, duties, fees and surcharges on fees “as delegated by the county government�. However, it is unclear whether these by-laws can continue to empower boards to collect fees even if the county government has not delegated that power. It is also unclear whether these fees and charges can be collected prior to the establishment of the boards. A legal opinion could help to clarify this. Also, as noted in Chapter 14, there are only going to be three urban boards in Kenya, and all remaining local authorities will become town committees. The UACA does not empower town committees to collect revenues. County governments will need a legal basis to continue imposing these charges from the time they take office, until they have the capacity to pass their own by-laws. 77 Chapter 7: Funding from within: County own revenue powers and capacity • If time permits, the amendments may provide an opportunity to address problems with existing local authority revenues. This will be discussed in detail, but one example would be for the local authorities to complain that the enforcement procedures for non-collection are cumbersome, and this is their main reason for low collection rates. 7.26 A secondary measure would involve developing a model county taxation and business regulation law, to guide counties in developing their own taxation and business regulation laws (county legislative drafting capacity is likely to be very limited in all, but a few counties). However, such a template would need intergovernmental consultation, and would therefore take time to develop. Until then, a legal instrument is needed to allow counties to collect revenues. Both measures―the interim national legislation and template county taxation and business regulation law―would also help to promote standardization of county revenue collection going forward. Addressing challenges in existing local own revenues 7.27 A new interim national law to govern county own revenues would provide an opportunity to revise existing taxes. In particular, property rates―the most important source of own revenue for local authorities―could be re-examined. Revenue base information is generally neither up to date nor complete. In Kenya, for example, the fiscal cadastre and valuation rolls include only between 20 and 70 percent of the total taxable land.11 Another problem is the lack of an adequate system for monitoring and recording land transfers. Records are not computerized: property records are kept manually and maintained in an ad hoc manner.12 Local authorities typically do not have the capacity to systematically maintain and coordinate their fiscal cadastre information. With the exception of Nairobi, Mombasa, Nakuru and Kisumu, they depend on the Ministry of Lands rating department to create and update their valuation rolls. All fiscal cadastre information is maintained manually.13 Local authorities also cite the costly and difficult legal process they must follow in order to prosecute land owners who default on their rates. 7.28 Overall, the key to improving property tax revenue in Kenya is to have improved administration. The weakest component of administration is collection, which results from: (i) lack of taxpayer confidence or understanding in how the tax is levied, collected, enforced and used; (ii) lack of legal and administrative collection and enforcement mechanisms; and, (iii) lack of political will, which has been the primary obstacle to implementing property tax reform.14 Box 7-1 summarises the recommendations of the Task Force on Devolved Government (TFDG) regarding how best to reform property rates in Kenya. Box 7-1: Recommendations of the Task Force on Devolved Government on property rates reform. ▪ The Rating Act and the Valuation for Rating Act need to be harmonized so as to provide adequate systems for the administration of property taxes. ▪ Rates taxes are currently based on unimproved site values, which deny local authorities substantial amounts of revenues. The harmonized legislation should therefore provide for rating of properties, to be based on both the improved and unimproved site values. ▪ Government ministries, departments and agencies are currently exempted from paying property taxes, and instead pay contributions in lieu of rates (CILOR) on an ad hoc basis. There should be no exemptions, and the ineffective CILOR payment system should be discontinued. Source: Republic of Kenya (2011). 11 Bird and Slack (2006). 12 Ibid. 13 Kelly (2002). 14 Ibid. 78 Devolution without disruption – Pathways to a successful new Kenya 7.29 The Single Business Permit (SBP) is another existing local authority revenue that could be revisited under a revised legal and administrative framework. While a good idea in principle, the efficacy of the SBP hinged upon disciplined implementation. However, the introduction of the SBP saw the burden on formal businesses increase. This was because: (i) local officials tended to (arbitrarily) apply the higher “bands� of the SBP to maximise short-term collection; (ii) many of the licenses it was designed to replace were never in fact repealed (most notably advertising licenses); and, (iii) fees remained high. The SBP is now cited by businesses as one of the main deterrents to business activity at the local level. Businesses that are formal (to a great or less extent) still face higher administrative and financial costs under the SBP, and they are subject to other charges and fees that should have been consolidated into the SBP. LAs in Kenya also rely on production, transport, and market cesses as output taxes collected at the time of sale or transit. These cesses affect informal businesses as well. Businesses also complain about coercive enforcement and rent-seeking by local officials, which foster taxpayer resistance and encourage evasion.15 7.30 The problems with the early implementation of the SBP leads to a series of reforms aimed at streamlining the regulatory environment for business activity. This has included reducing the total number of licenses, and simplifying many of the remaining ones. However, these reforms have had only limited impact in feeding through to a simplified regulatory environment for businesses in Kenya, as is evidenced in Box 7-2. Since business licencing will be a county functionunder the devolution process, the problems encountered with the SBP will only be addressed through leadership at county level. While the extent to which national government can set parameters for the SBP is unclear, it seems likely that it will be at the discretion of individual county governments on how to improve the implementation of the SBP. One thing that may help in this regard is introducing a degree of horizontal competition among counties, by publishing information regarding county Figure 7-7: How long does it take businesses to get things done? level business environments― 140 500 building on the type of 450 120 information generated by 400 100 350 the World Bank’s subnational 300 Time (Days) 80 Time (Days) doing business survey―and 250 using intergovernmental 60 200 coordination fora to share 40 150 experience on good practices 100 20 applied in counties which seem 50 to score more highly on these 0 0 o ret rissa i olo i li sum u l a ba asa o bi rok y eri hika rage measures. The Subnational El d G a I s K Ki Ma om b Na i r N a N T Av e M Doing Business Survey is Starting a Business [LHS] Dealing with Construction permits [LHS] Registering property [LHS] Enforcing contracts [RHS] already revealing interesting Source: Kenya subnational doing Business Survey (2010}. differences across municipal areas in Kenya (see Figure 7-7). 15 World Bank (2009). 79 Chapter 7: Funding from within: County own revenue powers and capacity Box7- 2: Regulatory environment for businesses―has the SBP worked? According to a spring 2010 survey of a representative sample of formal businesses in Nairobi, Mombasa, and four other towns in Kenya, the average formal sector business (that is, one registered for taxes with the Kenya Revenue Authority) had three licenses and/or permits, most of which had to be renewed annually. The annual financial cost of obtaining or renewing such licenses and permits averaged about US$ 1,000, plus an average of 30 work-days’ time for the entrepreneur or his/her staff. A particular irritation for businesses are “advertising licenses,� which are imposed not only for billboards, but also for signs placed outside buildings, including small signs above or next to street entrances. These were recommended for elimination, but are still in force. In Nairobi, businesses complain that the SBP is enforced unfairly and is subject to corruption. The categories for sector and size (and the different fees associated with them), are widely claimed to be subject to negotiation. According to the survey, the transport sector and restaurants have the most problems in this regard. Source: FIAS/World Bank (2011). County fees and charges 7.31 The County Governments Act (CGA) gives county governments the responsibility for setting tariffs.16 The water sector provides a particularly important case study in some of the implications this clause may have for county revenues (although it will be an extremely important issue in other sectors such as health too). The Kenyan experience prior to the 2002 Water Act suggests that public utilities like water providers should be operated at arms-length from political interference, and that they should as far as possible be subject to market discipline. As shown above, prior to the 2002 Water Act, water and sewerage fees were the primary revenue source for local authorities, with limited reinvestment of these revenues back into the water sector. The Act comprehensively restructured the water sector, establishing water service providers as limited liability corporations with ring-fencing of their revenues for reinvestment in the water sector. This ring-fencing is largely seen within the water sector as being successful, in that it provides a reliable stream of finance for the expansion of water services in Kenya. 7.32 In Kenya’s water sector, there are two distinct categories of entities. Most counties have their own water service providers, but these only manage the smaller reticulation networks. The eight water service boards manage the bulk assets like dams and large pipelines that provide services which cut across future county boundaries. The draft Water Bill currently proposes that all property, assets, rights, liabilities, obligations, agreements and other arrangements which exist regarding the operation of the eight water services boards shall be transferred to water works development boards (s.131(1)), and that county or cross-county water service providers will take on water service provision (s.132(1)). 7.33 Some provisions in the PFM Act may adversely affect the independence of these county and cross- county water companies. The PFM Act allows county governments that wish to do so, to bring county corporations fully within the county Public Financial Management system, by declaring them to be “county government entities� under Section 5 of the PFM Act, with the approval of the county assembly, and subject to criteria as defined by regulations. Once declared to be a county government entity, a water corporation would operate fully under the control of the county treasury, including having revenues and expenditures fully incorporated into the county budget process, and their bank 16 County Governments Act, article 120. 80 Devolution without disruption – Pathways to a successful new Kenya accounts managed by the county treasury. The PFM Act also emphasizes that in cases of conflict with other legislation on inter alia, “raising of revenue and making of expenditures�, it prevails. This may create a conflict with some of the provisions of the Water Bill, which provides that funds collected by county level water service providers be used exclusively for cost recovery and asset development. Without revenue ring-fencing, water service providers will not be attractive to private investment,and national or county governments will have to finance capital expansion of water networks from within their own limited resources,which may limit the long term growth of the sector. 7.34 Obviously counties will have many differing expenditure priorities of which water and sanitation service expansion may not be the highest. Clearly, though an explicit discussion needs to be held between policy-makers in the water sector―currently preparing a legal framework with national-level regulation of tariffs and ring-fencing of revenues for reinvestment in the sector―and those charged with the implementation of the County Governments Act (allowing counties to set tariffs) and the PFM Act (allowing county corporations to be declared county government entities) to resolve any perceived or actual tensions between the sectoral and cross-cutting policy and legal frameworks on devolution. This coordination challenge is likely to be one that faces other sectors, as they develop policies and associated legislation to respond to the constitution and devolution. The results will have a significant impact on the amount of revenue available to counties through fees and charges, and how much discretion they have to spend it. The case for additional county revenue sources 7.35 While the primary challenge facing new counties is to improve administration and collection of existing subnational revenues, this would still leave the counties extremely reliant on national government transfers. Addressing this heavy dependencewould require granting of additional own source revenues to counties through national government legislation. If such county revenues are carefully selected, they may increase total countrywide revenue collection (i.e. they would not be offset by reductions in national revenues) by reaching a greater share of GDP. This is because national taxes typically have relatively high entry thresholds, while these taxes that might be raised by counties could reach smaller firms, individuals and owners of immovable property currently not subject to national taxation, counties would have a comparative advantage compared to national government. 7.36 Central governments are often reluctant to transfer responsibility for more lucrative taxes such as income taxation, partly motivated by fears of loss of macroeconomic control. It may also harm infrastructure spending as subnational government expenditures tend to be dominated by recurrent spending. Decentralising taxing power is also likely to increase inequality among counties, by favouring those with greater taxing capacity, such as Nairobi and Mombasa.16 However, there are a number of strong counter-arguments as to why Kenya’s counties may benefit from being granted additional own revenue sources. 7.37 Why might additional revenue raising powers―over and above those already collected by local authorities―be a good idea? First, greater reliance on own source revenues may also increase county governments’ fiscal responsibility, since increasing spending carries a direct political cost as taxation also needs to be raised to finance it. In contrast, heavy reliance on intergovernmental transfers creates incentives to increase spending, since the subnational government is “spending other people’s money�.17 16 Bahl and Bird (2008). 17 Ibid. 81 Chapter 7: Funding from within: County own revenue powers and capacity 7.38 Second, it may improve the accountability of county governments to their citizens. Local residents are likely to hold county governments more accountable, if they are directly financing a greater proportion of county expenditures, rather than indirectly through intergovernmental transfers, since they care more about how their own tax money is spent than other peoples’. This requires local taxes to be both visible to local voters and large enough to impose a noticeable burden on taxpayers. The relatively minor taxes and fees levied to date by local authorities in Kenya, do not meet this criteria―county revenue raising powers would need to go beyond this. Related to this accountability argument is that it may be preferable in a more heterogeneous country such as Kenya (e.g. with very wide variations in population density and geography between counties) to allow for the greater diversity in service provision that own revenues (and the associated “downward� accountability) allow, compared to the standardisation (and “upwards� accountability) that can come through heavy reliance on transfers from national government, which often have conditions attached.18 7.39 A growing body of academic research supports this “accountability argument�, making the point that taxation is not purely instrumental (i.e. a means to funding public services) but it is also a central part of the ever-evolving contract between citizens and the state (see Box 7-3). This link between taxation and state building suggests that building a strong tax base for county governments is not simply about providing more and better devolved public services. It implies that there is a qualitative distinction between money received in the form of transfers from the national government, and money raised directly by the county, and that the latter will be extremely important for state-building at county level. Over reliance on national government to finance county governments could undermine the establishment of strong accountability relationships between taxpayers and county governments. Box 7-3: Tax systems and state building. Tax systems are instrumental to building effective states because taxation is a core manifestation of the social contract between citizens and the state. How taxes are raised (and spent) shapes government legitimacy by promoting accountability of governments to tax-paying citizens, and by stimulating effective state administration and good Public Financial Management. Sound and fair domestic taxation systems promote good governance because: ▪ raising taxes efficiently requires political effort to secure taxpayer consent; ▪ raising taxes effectively requires the development of a competent bureaucracy; and, ▪ raising taxes equitably requires political concern for the fair and equal treatment of citizens by the state. So taxation is at the centre of good governance and state-building. The perceived fairness of the tax system is crucial to building an effective state based on citizens’ consent. Willingness to pay taxes is a good indicator of the legitimacy of the state. The tax system is also central to state-building, because building compliance is a central incentive for the state to engage with society. The ‘state-building’ approach to taxation requires the construction of a tax policy and its administration, which can strengthen the legitimacy of the state in the eyes of its citizens through five core characteristics―political inclusion; accountability and transparency; perceived fairness; effectiveness; and political commitment to shared prosperity. Source: FIAS/World Bank (2009, p. 3). 18 Bahl and Bird (2005). 82 Devolution without disruption – Pathways to a successful new Kenya 7.40 However, this concern must be balanced with the fact that it is far from automatic that increased fiscal autonomy for counties will result in improved efficiency and responsiveness of county government. It will be essential to ensure that staff in county governments have adequate capacity and integrity to manage increased autonomy. During Tanzania’s decentralisation process, this was highlighted as a major concern: although subnational taxation only accounted for around 6 percent of total national revenue at the time, the multiple revenue sources and their “unsatisfactory� structure and implementation meant that they had a negative economic and social impact, far beyond that suggested by this figure. Deterioration of public services and corruption further reinforced tax resistance, in turn requiring more expensive and aggressive approaches to collection by subnational governments. As Fjeldstadt and Semboja caution:19 “Sustained development in local governments cannot grow from an institutional framework which encourages coercion and extra-legal tax enforcement. Furthermore, attempts to raise additional revenues from poorly designed taxes may aggravate the negative effects of the tax system on the economy and the society in general. It is unrealistic to expect that the present staff in many councils has adequate capacity and the required integrity to manage increased fiscal autonomy. In fact, there is a real danger that, in the absence of substantial simplification and restructuring of the current revenue system combined with capacity building and improved integrity, increased autonomy may cause greater mismanagement and corruption in local authorities.� There are many parallels between the situation reported in Tanzania by Fjeldstadt and Semboja and the current challenges faced in Kenya. Addressing this challenge will involve not only ensuring that revenue administration capacity in local authorities is preserved, but also that it is significantly enhanced in the transition to county governments. Design of subnational tax regime 7.41 While there may be a case in principle for legislating nationally for additional county revenues, this also requires identification of an appropriate tax. A good local tax is typically seen as one that is easy to administer locally, with an immobile base (i.e. imposed mainly on local residents), does not raise problems of harmonisation or competition horizontally among subnational governments, or vertically between subnational governments and national governments, and allows control over the rate to maximize inter-county competition.20 These strict tests are usually only met by property taxes, supported by taxes on vehicles, and user charges and fees. However, as noted above there are powerful arguments in favour of considering additional revenue sources for county governments. 7.42 We can therefore consider refining the list of criteria of what typically constitutes a good local tax. Bahl and Bird (2008) suggest seven criteria: i. The tax base should be relatively immobile, to allow local authorities some leeway in varying rates without losing most of their tax base. ii. The tax yield should be adequate to meet local needs, and sufficiently buoyant over time (that is, it should expand at least as fast as expenditures). iii. The tax yield should be relatively stable and predictable over time. iv. It should not be possible to export much, if any, of the tax burden to non residents. v. The tax base should be visible, to ensure accountability. vi. The tax should be perceived to be reasonably fair by taxpayers. vii. The tax should be relatively easy to administer efficiently and effectively. The cost of efficient administration should be a ‘‘reasonable’’ proportion of revenue collections. 19 Fjeldstadt and Semboja (2000). 20 Bahl and Bird (2008). 83 Chapter 7: Funding from within: County own revenue powers and capacity Table 7-6: Classification of subnational taxes by degree of central vs. local control. Local Autonomy in SNG sets tax rate and base Greatest access to own-source revenues. These revenue policy usually include fees and charges. SNG sets tax rate only Necessary and sufficient condition for “own- revenue�. Piggybacking and tax base harmonization permitted. SNG determines the tax base Refers largely to local authority to grant exemptions that erode the local base. Limited autonomy SNG sets tax rate, but within centrally In this case the centre typically specifies a high/low permissible ranges tax range or caps the top rate. Tax sharing, whereby central/local Can result when a local authority collects the tax revenue split can change only with and remits it to the centre. consent of SNG No local autonomy Revenue sharing, with share determined 100 percent control by centre; this is a source of unilaterally by central authority misspecificationof central vs. Local revenue. Central government sets rate and base May accompany political decentralization. of “SNG revenue� Source: Taliercio (2005, p. 109). 7.43 But how do we define a truly “local� own source tax? As Taliercio (2005) observes, many taxes that appear local are in fact national and vice versa, and a clear typology is needed in order to distinguish between these based primarily on the degree of autonomy the subnational government has over revenue policy (as opposed to revenue administration)―see Table 7-6. There is therefore a spectrum of revenues, with genuine subnational “own� taxes at one end (subnational government sets rate and base), and entirely national revenues at the other (no subnational control of base and rate of tax). 7.44 Taxes that could potentially be considered for counties to enable them to raise significant additional own revenues include: Personal Income Tax (PIT) surcharges, taxes on the use of motor vehicles, and payroll taxes.21 “Piggybacking� of county taxes on existing national taxes is a particularly credible option, as it would allow counties to set tax rates, while the costs of administration would solely be borne by the national government. 7.45 It also seems highly likely that forthcoming legislation in Kenya may provide for the intergovernmental sharing of royaltiesfrom the exploitation of natural resources. Article 66(2) of the constitution requires parliament, while regulating the use of land, to enact legislation that ensures that investment in property benefits local communities. The Task Force on Devolved Government interpreted this as providing a basis for the sharing of royalties paid by investors exploiting minerals and other natural resources.22 There are ad hoc arrangements in place in some communities―for example in the Maasai Mara, Narok County Council receives fees which are equivalent to royalties by the exploiters of the National Reserve―while this is not applied in other National Parks. The Task Force recommends that royalty taxes levied on minerals be applied to all other natural resources, including game reserves, forestry and marine witha tax to be shared between the national government and the affected county government in a proportion of 70:30. The rationale for this share is to benefit the local communities, who would be bearing the highest impact of the exploitation of the natural resources. A clause was originally included in the County Governments Financial Management Bill to this effect, although this was deleted when it was merged into the harmonized PFM Bill. 21 Bahl and Bird (2005). 22 Republic of Kenya (2011). 84 Devolution without disruption – Pathways to a successful new Kenya 7.46 The Third Schedule of the draft Geology, Minerals and Mining Bill, 2012, also includes a royalties sharing clause, proposing 75 percent for national government, 15 percent for county government and 10 percent for the community.23 It also proposes that, where more than one county or community are involved in sharing of royalties, the royalties will be shared accordingly. Although the Bill does not cover oil and gas, this proposed approach to sharing of royalties would set an important precedent,which may come to be regarded as a minimum floor for oil and gas legislation. The recent oil find in Turkana as well as an upsurge in offshore oil and gas exploration, adds to the importance of such a legal provision, as it has potentially huge implications for counties discovering significant reserves of oil or gas. 7.47 Whatever the decision on whether or not to grant counties additional revenue raising powers, there is a strong case for limiting the total number of taxes that counties can impose. This could be done by ensuring that national laws governing county revenues (over and above those granted to counties in the Constitution), define a “closed list� of allowable revenues available to counties, from which they can chose, in order to avoid nuisance taxes or highly distortionary taxes, which can adversely affect equity and economic growth. As noted above, multiple local taxes have been identified as a major problem, for example in Tanzania24. This must be balanced by the need to ensure the inclusion of sufficient revenue instruments on the “closed list�, with sufficient county government discretion, to generate stable and buoyant local revenues (e.g., property taxation, appropriate local excises and business taxes).25 In 2003, Tanzania introduced amendments to the Local Government Finances Act of 1982 to switch to a closed list approach, while Indonesia in 2009 replaced Law No. 34 (2000) with Law No. 28 (2009) to reintroduce a closed list approach, including local access to the rural and urban property tax and to the property transfer tax (see Box 7-4).26 Box 7-4: Indonesia’s experience with “open� and “closed� lists of subnational revenues. Indonesia had traditionally followed a closed list‘ approach, including Law No. 18 (1997) which was passed just prior to the 2001 Decentralization Reforms. However, Law No. 18 did not include the property tax as a LG revenue option. Following the Big Bang decentralization reforms, the Government introduced Law No. 34 (2000), which shifted from a closed‘ list to an open‘ list approach, to provide increased discretion to LGs in support of decentralization. However, under this open list approach, LGs were still not allowed access to property taxes as these tax already existed as central government shared taxes. As will be explained in Section 3, the open‘ list approach led to a proliferation of nuisance and highly distortionary taxes―as LGs creatively sought ways to improve their local revenues. In reaction to the problems resulting from Law No. 34 (2000), the Government shifted back to a closed‘ list approach, but this time devolved the rural and urban property tax (PajakBumidanBangunan/PBB) and the property transfer tax (Bea Perolehan Hakatas Tanah danBangunan/BPHTB), giving LG‘s access for the first time to potential revenues from property taxation. Source: Kelly (2011), p. 14. County tax collection 7.48 The constitution is silent regarding whois responsible for tax collection at county level. The Final Report of the TFDG identifies four possible options for county tax administration: 22 Republic of Kenya (2011). 23 Draft Geology, Mining and Minerals Bill accessed from CIC website on 9th October 2012: http://cickenya.org/sites/default/files/Geology_Minerals_and_Mining_Bill_2012.pdf 24 Fjeldstadt and Semboja (2000). 25 Taliercio (2005). 26 Kelly (2011). 85 Chapter 7: Funding from within: County own revenue powers and capacity 1. County governments can set up tax collection departments or divisions. 2. County governments can contract out the task to private organizations. 3. A new County Tax Revenue Authority could be empowered through national legislation to collect tax revenues for all counties. 4. The Kenya Revenue Authority (KRA) could be empowered through national legislation to collect all tax revenues for counties.27 7.49 The Task Force Report argues that reliance on options 1 (County collection) and 2 (contracting out) by local authorities, has not led to improvements and cost-effectiveness. It also argues that option 3 (County Tax Revenue Authority) would come with initial capital expenditure and disadvantages of economies of scale, compared to option 4 (using KRA). The Task Force went on to argue that Kenyan experience has shown that KRA involvement in the collection of land rents, and the ceding of the power to collect water charges and fees to the Nairobi Water and Sewerage Company, has improved collection efficiency. The Task Force therefore recommended that proposed legislation on county governments’financial management should vest the power for the collection of county taxes in the KRA, proposing that the KRA Act could also be amended to make it a shared institution that serves the two levels of government. This would, it is argued, have the added advantages of enforcing uniform standards and generating economies of scale, thereby lowering tax administration costs and enhancing tax revenues. 7.50 The PFM Act sets out a basic framework for county tax collection. It empowers the County Executive member for finance (CEMF) to designate persons to be responsible for collecting, receiving and accounting for county government revenue, and those receivers are responsible to the CEMF. Receivers of revenue can in turn authorise other public officers employed by the county government to be collectors of revenue, responsible for collecting and remitting revenues to the receiver of revenue. Receivers of revenue must provide quarterly statements to the County Treasury with a copy to the National Treasury and the CRA. Importantly, the Act also permits the CEMF to authorise the Kenya Revenue Authority (KRA), or appoint a collection agent to be a collector of county government revenue in accordance with the PFM Act regulations. The Act therefore leaves options 1, 2 and 4 from the Task Force report open to county governments. 7.51 International experience on contracting out of local revenue collection to private organisations (Task Force option 2) has been mixed. Also known as ‘tax farming’, this practice has been tried by local governments in both Tanzania and Uganda, as well as in Kenya to some extent. Box 7-5 summarises research on the experience with contracting out of local revenue collection in Tanzania and Uganda. In short, the evidence is at best mixed, and Uganda’s case suggests that inattention to the incentives created by the detailed design of contracting out mechanisms has lead to substantial loss of revenues. This suggests that Kenya’s counties should be cautious about adopting option 2, and if they decide to do so they should pay careful attention to detailed design of contracting out arrangements and the incentives they create. 27 Republic of Kenya (2011). 86 Devolution without disruption – Pathways to a successful new Kenya Box 7-5: Regional experience in outsourcing local revenue collection. Tanzania A recent study on outsourcing of subnational revenue collection in Tanzania experience in seven local government authorities (Fjeldstadt et al., 2009). The study covered two City Councils, two Municipal Councils, and three rural district councils. Mwanza City Council pioneered the outsourcing of revenue collection as early as 1996, and since 2000, an increasing number of local authorities have adopted private tendering systems for the collection of different types of taxes. These include property taxes, market fees, forestry levies (until 2005), cess on certain agricultural products, bus stand and parking fees. Revenue collection is outsourced to a range of different types of agents within and across the councils studied. These include: a private consultancy firm specialised in providing tax advice to the private sector, a firm principally engaged in operating private schools, market associations, co-operatives, and a fish dealer organisation. Outsourcing of revenue collection is generally based on an open tender process, referring to guidelines provided by the Local Government Reform Programme (June 2003) and the Public Procurement Act No 21 of 2004. Revenues from taxes where collection is outsourced as a share of the total revenue collection from the council’s own sources differ across councils, ranging from 10 percent to over 60 percent. The study found that outsourcing often yields more revenue, compared to council based collection from similar sources. However, the data from these case studies suggest only marginal differences over time, in revenue collection from sources collected by private agents and those collected by councils. Uganda Uganda introduced privatised tax collection relatively early in its decentralisation process―including market fees, and fees at fish landing sites―with a view to improve both the yield and efficiency of tax collection. Iversen et al. (2006) review Uganda’s experience with private tax collection, drawing on evidence from six rural councils. The authors conclude that the call for an increased role for private collection of taxes at local level fails to examine how such systems actually work. They find that rather than enhancing local revenue, the private tax collection system in Uganda transfers money from ordinary and often poor rural tax payers to local elites (tenderers, local bureaucrats and politicians). This is because elites collude to ensure that the agreed tender value―the price the private tenderer pays for the right to collect, which is based on the estimated revenue yield of the market prepared by the district―is far below the revenue actually collected. Even after allowing for the official 20 percent ‘permitted margin’ the lost revenue across the six markets studied (i.e. additional revenue councils would have collected if the revenue potential of each market had been properly assessed) ranged from 25 percent to 74 percent of total revenue collected in each market. The authors conclude that this phenomenon arises from the conflict of interest between the official objective―local revenue enhancement―and the private incentives of local elites. They propose that the market tax yield assessment function should be taken out of the hands of district administrations, and conducted independently. They suggest that the disparity between taxes collected and local service delivery under this system will lead to a deterioration in state-citizen relations. Source: Fjeldstadt et al. (2009). Iversen et al. (2006). 87 Chapter 7: Funding from within: County own revenue powers and capacity CHAPTER 7: KEY INSIGHTS / RECOMMENDATIONS. Estimating county own-source revenue potential is important because: (i) own revenues should be taken into account when designing the intergovernmental resourcing framework; and, (ii) they will be fundamentally important for building local accountability between county governments and their citizens. The constitution grants counties very limited own revenue raising powers, but allows the national government to grant them additional sources of revenue via legislation. Without such additional sources of revenue, counties will remain highly dependent fiscally on national government. Kenya’s future counties will remain highly dependent on transfers from the center, although this masks large variations among them, and the treatment of natural wealth may soon become an issue that needs to be addressed. Presently, the constitutional base for county tax collection has not been converted into legal instruments to enable counties to collect revenues from day one after the elections. Property taxes are the main source of revenue for local authorities (and future counties) but currently, they are under collected. Presently, fees and charges are important sources of subnational revenueand will be important revenue sources for county governments. The emerging legal framework may contain some conflicts though, especially in the water sector, with regard to who sets tariffs, and how much discretion counties have in spending revenues. *** As a matter of urgency legal instruments ought to be put in place to allow counties to continue collecting revenues and charges currently being collected by local authorities through local authorities’ staff and structures until taxation laws are passed by county assemblies. A new interim national law to govern county own revenues could provide an opportunity to revise existing taxes, particularly property rates (to apply to both land and improvements and to ensure that public land generates adequate fiscal revenue) and SBPs. Sectoral fees and charges need special attention. Currently, there seems to be tensions between the overall cross-cutting legal framework for devolution (e.g. County Governments Act and PFM Act) which promote county discretion and sectoral draft legislation (e.g. in water) that seeks to limit county discretion. Dialogue is required to resolve these apparent tensions. Tax administration at the local level ought to be strengthened, in order to increase taxpayer confidence, clarify legal, administrative and enforcement mechanisms, and incentives for politicians. Additional local fiscal revenues should be sought among a range of possible options, including PIT surcharges, taxes on the use of motor vehicles, payroll taxes, etc. Particular consideration should be given to the sharing of taxation related to natural resources as their potential may increase vastly in years to come, and this issue often constitutes a source of conflict. 88 CHAP TE R E IG HT Allocating the county share equitably but carefully 8.1 Transfers will have a major role to play in ensuring that county needs are met, and that enduring gaps between them are filled. This is because Kenya’s future counties will start from very different positions in terms of funding needs and the capacity to meet them out of own resources. Determining the optimal amount that counties will need collectively is challenging. Equally daunting is the task of further allocating these resources horizontally across forty-seven counties. The complexity comes mainly because the goal of equalization will need to be pursued, in a tight fiscal environment, without disrupting existing services or undermining growth and efficiency. De facto, the parameters of the horizontal formula for allocating the equitable share will also impact the vertical split of revenues, as more redistributive formulas will require a greater aggregate transfer to counties. This has two major implications: (i) existing imbalances may only be tackled over time; and, (ii) equalization policies should target people who will benefit from improved services, rather than trying to achieve equality between different geographic locations. 8.2 Unpacking individual county needs will be challenging. There are many ways in which needs can be understood. At minimum, counties will require resources to continue to finance existing levels of services once they are devolved. But the Kenyan devolution project is fundamentally about equalization, and a much more ambitious definition of needs could also support resourcing counties to allow them to meet benchmarks of service levels and quality, and to mitigate access or outcome inequalities among them. In practice, policy-makers will need to walk a fine line between delivering on equalization, while ensuring that fiscal equilibrium can be maintained. The most realistic way to achieve this is likely to be by situating equalization objectives within a medium-term context. 8.3 Defining equity, as a basis for equalization, is very complex. The difficulty derives from the fact that people have various understandings of what is equitable, and because broad principles eventually need to be translated into a workable transfer formula. While formulas can capture many different variables, there is always a trade-off involved. The more finely tuned a formula is, the more complicated it becomes. This is particularly important for Kenya, because the Senate will decide the formula for horizontal sharing, and members of the Senate are likely to want a simple formula that they and their voters can readily understand. However conversely, since most counties will be very dependent on transfers to finance their service delivery responsibilities, a formula that does not respond accurately to some counties’ needs will have a major impact on their capacity to deliver (see Box 8-1). 8.4 The Commission on Revenue Allocation (CRA) has proposed a formula for allocating the equitable share, which will need to be fine-tuned over time and complemented with other transfers. The relatively―purposefully―crude nature of the formula, given the wide dispersion of needs across forty- seven counties, means that other instruments will be needed. In turn,in order to preserve enough fiscal space for additional targeted transfers, a case can be made to limit the size of the equitable share transfer at―or close to―its constitutionally defined minimum of 15 percent. 90 Devolution without disruption – Pathways to a successful new Kenya Box 8-1: A note on defining “equalization�. Kenya’s Constitution provides two broad means to address the historical and current inequity between counties. First, it provides for an equitable share to be allocated among counties in accordance with the criteria in Article 203(1). Second, it provides for an Equilization Fund that is aimed at addressing backlogs of infrastructure for basic service delivery. In keeping with international use we apply the term “equilization� to both these mechanisms. This chapter concerns the allocation of the county share of revenue―the horizontal division―among the counties. It is assumed that the Constitution intends that the equitable share should be allocated in such a way as to ensure counties have a similar capacity to deliver public services. In the international literature on intergovernmental financing, this is described as “equilization". Source: World Bank staff analysis. 8.5 The horizontal sharing formula, as proposed, does not resolve the contradiction between two intertwined objectives: to ensure adequate funding for service delivery within each county (as opposed to the aggregate vertical level), and promote geographic redistribution. A major implication is that the redistributive objectives of the transfers scheme may need to be pursued gradually and over time, and may require additional transfer instruments. A powerful constitutional mandate: Equalization across Kenya’s forty-seven counties 8.6 Equalization objectives have received a strong backing in the Constitution. This is not surprising, given Kenya’s historical legacy of important interpersonal and―even more strongly felt―interregional inequality. But equity is not as straightforward a concept as it seems. In working out the ways to operationalize these goals, it is essential that a fair, efficient and transparent system can be worked out. The constitutional mandate for equalization 8.7 The Constitution of Kenya emphasizes equity as a core objective of the decentralized public financial architecture. Article 201 (Chapter Twelve) underscores that “the public financial system shall promote an equitable society�. The division of revenue is to promote this, both in terms of the vertical split (“revenue raised nationally shall be shared equitably among national and county governments�), and when it comes to allocating revenue across the forty-seven counties (“expenditure shall promote the equitable development of the country, by making special provision for marginalised groups and areas�). These broad considerations are further elaborated in Article 203, which lays out the criteria to be taken into account when determining the county equitable shares, including: - “the need to ensure that county governments are able to perform the functions allocated to them� (d) - “the fiscal capacity and efficiency of county governments� (e) - “developmental and other needs of counties� (f) - “economic disparities within and among counties and the need to remedy them� (g) - “the need for affirmative action in respect of disadvantaged areas and groups� (h) - “the need for economic optimisation of each county, and to provide incentives for each county to optimise its capacity to raise revenues� (i) - “the desirability of stable and predictable allocations of revenue� (j) 91 Chapter 8: Allocating the county share equitably but carefully 8.8 Equity is central but undefined in the Constitution and efficiency is also underscored. While the Constitution flags equity as a central consideration to guide the revenue sharing arrangements, its definition remains wide-open and therefore the basis on which to design redistributive policies also remains to be decided. Moreover, efficiency is also underscored (ensuring minimum service delivery, preserving fiscal discipline, and guaranteeing predictable allocations) and in practice, the twin objectives of promoting equity with efficiency could come into conflict. Defining and measuring equity 8.9 What is an equitable allocation of resources? Unpacking the equity requirement of the Constitution implies giving substance to the notions of “development and other needs of counties� as well as of “economic disparities�. The goal of fiscal equalization, as typically expressed in academic literature, is to enable each subnational unit to provide comparable levels of public services, at comparable tax rates. This traditional definition falls short of a more ambitious equalization agenda that would take into account existing gaps in infrastructure coverage, natural wealth, economic activity or sheer (average) household income variations (although those would be reflected in measures of fiscal capacity). But it would already constitute a very ambitious agenda in Kenya, given the magnitude of existing gaps. 8.10 What is a feasible allocation of resources? Ultimately, the feasible extent of equalization will be determined by available resources (to finance equalization objectives over and above existing obligations) and the speed at which service delivery can be scaled up in under-served counties and streamlined in (relatively) over-served areas. Given the: (i) limited fiscal space for equalization initially; (ii) absorptive bottlenecks in lagging counties; (iii) the need to keep the equitable share allocation formula as simple as possible; and, (iv) the desirability of maintaining space and flexibility to address region or sector specific needs, via other transfer instruments, it makes sense to focus the equitable share transfers to fulfilling a limited set of objectives. Further equalization could be pursued through additional instruments and introduced over time. The CRA proposal: An important step toward redistribution 8.11 The CRA has made recommendations to Parliament for dividing the equitable share across counties. The proposed formula takes into account the major driver of needs differentials across counties (population), while introducing strongly redistributive elements. Given its relative simplicity, the formula will not be able to account for the wide diversity of county needs and capabilities, creating risks (of under/over provision) at each extreme of the spectrum. Mitigating these risks will be a major policy challenge going forward. The “textbook� approach to fiscal equalization 8.12 The textbook objective of an equitable sharing formula is to fill the ‘fiscal gap’ between expenditure needs and revenue (potential) for each subnational government. Following is a simple depiction of how a fiscal gap formula works (see Box 8-2). The first step is to define subnational expenditure needs (this is the first element in the first equation), given the service delivery functions assigned to them and the cost of delivering these functions. The second step is to assess the extent to which subnational governments could meet (part of) these needs out of their own revenues (this is the second component 92 Devolution without disruption – Pathways to a successful new Kenya of the first equation). The job of a formula is to share the available resources fairly between individual counties, based on their relative needs: resources available to subnational government units are allocated among them, according to the relative size of each county’s fiscal gap (this is the second equation). Box 8-2: The fiscal gap model. How a Fiscal Gap formula works Equation 1: the �scal gap EXPENDITURE NEEDS FISCAL CAPACITY FISCAL GAP minus Revenue a county can raise using equals Amount a county needs to deliver Shortfall (gap) its own taxing powers, to meet their mandated functions those costs Equation 2: the county share Fiscal gap of County X X County equitable share Allocation to County X Fiscal gap of all Counties in Kenya Source: World Bank staff analysis. 8.13 In practice, most countries have adapted this theoretical model to suit their particular circumstances. Some countries (including Australia, Denmark, China and Ethiopia) have opted for the full fiscal gap approach, but others have chosen to emphasize either on the revenue or the expenditure side. For instance, countries such as Canada―where differences in service delivery levels are modest across provinces―equalize across revenue capacity only.1 By contrast, other countries including South Africa equalize across needs only due to the paucity of local own-source revenues. While the theoretical model is fairly straightforward, in practice there are a number of practical complications (see Box 8-3). Box 8-3: The many complexities of estimating fiscal gaps. Capturing revenue potential. Fiscal capacity is not only a function of per-capita income at the county level, but must be evaluated against the specific structure/composition of own-source county revenues. Fine-tuning needs. Needs for services are not fully proportional to population. Devolved services are not meant to be available to all residents equally. Instead they typically target specific groups in society (such as school age children, the disabled and elderly). Therefore the target ‘population’ (a subset of the overall county population) must be identified, to adequately gauge county needs. Factoring-in cost disability differentials. Meeting existing needs at county level requires an understanding of cost differentials of service provision across areas. These are typically related to geographic factors (urbanization, population density, terrain, etc.) and socio-demographic characteristics (overall poverty level, family structures, etc.). Source: World Bank staff analysis. 1 Although the recognition of significant differences in cost of delivery later gave rise to special purpose grants. 93 Chapter 8: Allocating the county share equitably but carefully 8.14 The proposed CRA formula adopts a different approach, with good reasons. A complex fiscal gap formula may not be possible or desirable for Kenya in the near future. There are inherent methodological and practical obstacles to adequately capture revenue capacity. In addition, complex formulas can undermine the transparency of the allocation, and therefore the extent to which all citizens can understand the public finance architecture and have faith in its objectivity. This is particularly important when geographic distributional decisions overlap with political divisions. Over time, it may become judicious to add a fiscal capacity equalization component into the formula. Estimates of fiscal capacity could be built into it, possibly relying on a single macro measure of capacity (such as county-level GDP), although such macro indicators are imperfect (they do not reflect the actual mix of taxes available to subnational governments, and they are typically not available in a timely manner at the subnational level). Alternatively, if the CRA wishes to design a formula that accounts for own-revenues, the best option in the short term would perhaps be to include actual revenue collection data in the formula, but with a discount to minimize perverse incentives on revenue collection effort. 8.15 It captures needs at a very high level. The proposed CRA formula does not identify target populations for specific decentralized services, set out benchmarks for service delivery levels, or factor-in an estimation of unit costs of service provision (an example of such attempts is provided in Box 8-4). Instead, it proxies them at very high level through its population, poverty and land-area criteria. The risk in this approach is that it may result in significant discrepancies between actual needs and resources granted in a significant number of counties, and therefore an inefficient allocation of scarce resources for service delivery. Box 8-4: Example of a needs based approach. The South Africa ─ general purpose ─ equitable share transfers to provinces includes the following: ▪ A basic share based on the province’s share of the total population (weight 14 percent). ▪ An education share based on the size of the school age population and the average number of learners enrolled in public schools over the previous three years (weight 51 percent). ▪ A health share based on the proportion of the province’s population without access to medical aid (weight 26 percent). ▪ An institutional component divided equally among provinces (weight 5 percent). ▪ A poverty component based on poverty incidence (weight 3 percent). ▪ An output component based on province-level GDP (weight 1 percent). Note: It is important to keep in mind that the equitable share transfers to provinces are complemented by conditional grants, which account for 15-20 percent of total funding to provinces. The equitable share transfers to municipalities are considerably more complex, and a review of the Local Government Equitable Share formula is currently underway. Source: Boadway and Shah (2007). The parameters of the CRA formula make it highly redistributive 8.16 The proposed formula revolves around five simple components. The main components of the formula―each with attached weights―are: (i) population; (ii) poverty; (iii) equal shares; (iv) land area, and; (v) fiscal responsibility. They closely mirror the Constitutional emphasis on: (i) matching resources with service delivery needs (Article 203(d)―via the population and land area criteria); (ii) redistribution (Article 203(g&h)―via the poverty criterion); and, (iii) incentives for efficient management (Article 203(e)―via the fiscal discipline component) (see Table 8-1). 94 Devolution without disruption – Pathways to a successful new Kenya Table 8-1: Components’ source, weights and objectives in CRA's formula. Components Population Poverty Equal share Land area Fiscal responsibility Year / Source Census (2009) Year / Source TBD (allocated equally initially Weight 45% 20% 25% 8% 2% Objective Resource counties Promote Provide each Factor in the Provide incentives to deliver services redistribution county with higher cost for prudent fiscal equally on a per in favour of resources to cover of delivering management capita basis historically lagging the fixed costs of services in areas running county remote, sparsely administrations populated areas irrespective of population or size Source: World Bank based on CRA report to Parliament, August 2002. 8.17 Unlike many equalization formulas used elsewhere, the CRA proposal is not grounded in a detailed estimation of individual county needs. By contrast to more sophisticated examples, the CRA formula has the merit of being highly transparent, and therefore easily understandable by all. Given Kenya’s history of marginalization and poor transparency in the allocation of funds across regions, this is in―and―of itself a major achievement. At the same time, the potential wedge created between allocations and needs (funding and function) means that large gaps are likely to occur, on day one, at both ends of the county distribution (ranked according to existing spending commitments), for two mutually reinforcing reasons: (i) because the formula is in fact highly redistributive; and, (ii) because existing spending was not historically allocated, in Kenya,based on consideration of needs (as proxied by population). 8.18 The simulated county transfers, using this formula, display a strong equalization bias. This is not obvious at first because, on an absolute basis, the better-off urban counties (such as Nairobi and Kakamega) will be receiving the lion’s share of the funds, while poorer ones (like Tharaka Nithi and Lamu) will find themselves at the tail end of the distribution (see Figure 8-1). This is because population Figure 8-1: Total equitable share allocations to counties. 5.0 (Assuming 15 percent equitable share, i.e. KES 91.2 billion for 2012/13) 4.5 4.0 3.5 3.0 2.5 KES billions 2.0 1.5 1.0 0.5 0.0 Nairobi Kakamega Bungoma Kisii Turkana Nakuru Kiambu Kitui Meru Machakos Makueni Wajir Mandera Homa Bay Mombasa Trans Nzoia Garissa Marsabit Kwale Kisumu Uasin Gishu Murang'a Bomet Nandi Siaya Busia Baringo Kericho Kajiado Nyeri Migori Tana River West Pokot Nyamira Nyandarua Vihiga Embu Laikipia Kirinyaga Taita Taveta Elgeyo-Marakwet Samburu Isiolo Tharaka Nithi Lamu Narok Kili Source: World Bank staff calculations. 95 Chapter 8: Allocating the county share equitably but carefully largely drives the allocation. However, the picture is almost entirely reversed once one looks at allocations on a per head basis―Isiolo, Lamu, Marsabit and Tana River are the big winners (see Figure 8-2). Once population is “netted-out� the poorer and smaller counties receive disproportionately big allocations, on account of the poverty criterion and the fixed equal share, as well as the fiscal discipline component, because it is to be distributed initially on an equal basis. The figures which follow assume that the equitable share is set at 15 percent for the fiscal year 2012/13 but the fundamental insights remain the same, irrespective of the share that is ultimately agreed upon. 8.19 One element of the Constitution not reflected in the formula is the need to factor in the fiscal capacity of counties. This omission could lead to distorted outcomes given the wide variation in own-revenue raising capacity across counties. This capacity can be estimated by looking at the revenues, historically collected by local authorities in each county (given that counties will inherit the same tax base and instruments). By factoring in own revenues (using 2009/10 data), the picture changes, with major shifts for the counties which collected most own revenues such as Nairobi, Mombasa, Narok, and Nakuru (see Figure 8-3). 8.20 Because the proposed CRA formula is highly redistributive, managing the transition will be a tall order. Should most of the funding to future counties be channelled through the equitable share― above and beyond 15 percent―there is likely to be a significant wedge between the actual service delivery costs that counties face, and the allocation they receive. This is because population, the main driver of needs, is only a part of the formula, and because the geographic distribution of services has been historically unequal across Kenya. While equalization will need to be pursued, the transition will have to be managed carefully to avoid service delivery interruptions in leading areas, and to mitigate the risk of serious absorption bottlenecks in lagging counties. Figure 8-2: Per head equitable share allocations to counties. 8,000 (assuming 15 percent equitable share, i.e. KES 91.2 billion for 2012/13) 7,000 6,000 5,000 KES per head 4,000 3,000 2,000 1,000 0 Elgeyo-Marakwet Isiolo Lamu Marsabit Tana River Samburu Wajir Turkana Busia Laikipia Garissa Kwale West Pokot Baringo Migori Kitui Embu Makueni Vihiga Bomet Nyamira Nyandarua Kirinyaga Nandi Trans Nzoia Machakos Kajiado Nyeri Mandera Siaya Kericho Homa Bay Uasin Gishu Mombasa Murang'a Kakamega Kisumu Kisii Bungoma Meru Nakuru Kiambu Nairobi Narok Taita Tharaka Kili Source: World Bank staff calculations. 96 Devolution without disruption – Pathways to a successful new Kenya Figure 8-3: Total county revenues including own sources. 12 Total equitable share 2012/13 (assuming 15 percent equitable share, i.e. KES 91.2 billion) and total own revenue 2009/10 10 8 KES billions 6 4 2 0 Nairobi Nakuru Kiambu Kakamega Bungoma Mombasa Machakos Kisii Kitui Turkana Meru Makueni Wajir Kisumu Uasin Gishu Mandera Murang'a Trans Nzoia Kericho Homa Bay Kwale Marsabit Garissa Bomet Siaya Nyeri Nandi Busia Kajiado Migori Baringo Nyandarua Embu Laikipia Nyamira West Pokot Tana River Vihiga Kirinyaga Taita Taveta Samburu Elgeyo-Marakwet Isiolo Tharaka Nithi Lamu Narok Kili County equitable share transfer County own revenue Source: World Bank staff calculations. Managing the transition: Equalizing over time to avoid fiscal stress 8.21 15 percent is only a constitutional bare minimum and transfers will be needed above and beyond that amount although not necessarily via the equitable share. In the aggregate, more than the 15 percent equivalent would be needed to continue funding for existing services, assuming a broad definition of the functions in the Fourth Schedule. As explained in Chapter 6, there are several different possible combinations of transfer instruments. The size of the equitable share transfer will be the main parameter. If the equitable share transfer amount is meant to cover the totality of county needs, because its formula is so redistributive, there will be funding gaps in many historically “over-funded� counties. Bridging these gaps in order to preserve service delivery at current levels, will create a large additional fiscal burden that Kenya may not be able to finance out of the national budget without major and painful adjustments. 8.22 Increasing the amount of the county equitable share transfer would not resolve the twin challenge of service failure and absorption bottlenecks. It would fix one problem (under-resourcing of leading areas), while further exacerbating the other (weak capacity in lagging counties to manage increased funding) and fiscal stress at the national level. Winners and losers: the risk for existing and future service delivery 8.23 Because it promotes substantial redistribution of resources across counties, the CRA formula also creates risks. There are two problems. In those areas that have been historically privileged, inherited service delivery obligations will require substantial funding, above and beyond even what a strictly population-based formula would provide. One way to overcome this problem would be to increase the amount of the equitable share above historical funding levels for devolved services, but potentially to unaffordable levels. By contrast, counties in areas that were previously under-served will receive substantial extra cash, relative to what they would need to simply carry over existing service delivery 97 Chapter 8: Allocating the county share equitably but carefully levels. In these counties, the challenge will be to manage these resources well, with limited capacity, and given inherent difficulties in rapidly scaling up supply of (and possibly also demand for) services. 8.24 Modelling inherited costs of service delivery at individual county level illustrates these risks vividly. The report “MP’s Budget Watch: Financial Year 2012/13� provides a breakdown of the estimated cost of service delivery on functions to be devolved, as well as spending via earmarked funds such as Constituency Development Fund (CDF), Road Maintenance Levy Fund (RMLF) and Local Authorities Fund (LATF), based on historical spending patterns. In the analysis below, these historical spending figures are shown (with CDF, LATF and RMLF netted out) with the allocations that would result from applying the CRA formula. In the two simulations different possible total amounts for the county equitable share are used: simulation one (Figure 8-4) assumes that the total equitable share transfer is KES 97.5 billion, reflecting the total amount tagged to ‘devolved functions’ (code 98) in the 2012/13 budget (net of CDF, LATF and RMLF), while simulation two (Figure 8-5) assumes that the full KES 203 billion recommended by the CRA can effectively be accommodated.2 8.25 There are significant challenges associated with both scenarios. The first simulation, reflecting a conservative fiscal scenario, shows that 31 of forty-seven counties would simply not be receiving enough transfers from the equitable share to cover existing service delivery obligations, while others would be receiving substantially more. In the most extreme case, Nyeri would experience the equivalent of a massive 75 percent funding cut. In order to avoid service delivery disruptions, inefficient reallocations or outright interruption, stop-gap funding would be required in the form of additional (conditional) transfers in an amount of KES 12.7 billion. In the second simulation, with a much increased total equitable share transfer, there would only be one ‘loser’ but a large number of counties would be experiencing significant surpluses (equivalent to over a 100 percent funding increase for 27 out of the 47) and the total fiscal burden may simply be impossible to accommodate in the national budget. 8.26 Another way to illustrate the challenges is to look specifically at devolved sectors. We focus on health spending, because health will most certainly be one of the most costly devolved functions, and a large proportion of existing health spending in counties is for staff salaries―an area where it is difficult to make expenditure cuts. As expected, the starting point is significant inherited disparities in current spending across Kenya’s counties. 8.27 Given existing disparities, a redistributive formula will create gaps at both ends of the distribution. To construct an estimate of inherited service delivery costs in health, an interpretation of the assignment of health functions in the Constitution is presented in Table 8-2. Based on this, salaries and other emoluments for 91 percent of the total 49,017 staffs on the government payroll of MoPHS and MoMS have been added, to correspond to staff currently carrying out district, rural and environmental health services, as well as the totality of the costs currently associated with the operation of primary health facilities. 8.28 Allocations under the equitable share transfer will diverge vastly from actual immediate needs. What the health spending data suggests is a major mismatch between the distribution of current spending―on future county functions―and the horizontal distribution implied by the equitable share formula. Figure 8-6 illustrates this mismatch. It contrasts the current (estimated) allocation of health 2 As discussed in chapter 5, the total amount budgeted for devolved functions in 2011/13 is KES 81 billion. However KES 16.5 billion is budgeted for DP AIAs. 98 Devolution without disruption – Pathways to a successful new Kenya Table 8-2: How the Constitution assigns health functions. Medical services Public health and sanitation County functions ▪ District and sub-district hospitals and ▪ Primary healthcare including: primary facilities - Family planning and meternal & child health ▪ Primary healthcare programs - Health education ▪ Clinical programs - Rural health and training centres ▪ Nutition ▪ Food control administrative services National functions ▪ National referral facilities ▪ Major disease control (e.g. malaria, leprosy, TB) ▪ Specialist hospitals (e.g. spinal) ▪ Kenya Expanded Program on Immunization (KEPI) ▪ KEMSA operational funding (excluding ▪ Kenya Medical Research Institute (KEMRI) drugs) ▪ National Public Health Laboratories ▪ Biomedical, hospital engineering and ▪ Port Health Control radiology ▪ Government Chemist ▪ Kenya Medical Training Centre (KMTC) Functions ▪ Provincial hospitals - operational funding shared by population among counties? with unclear ▪ HSSF - Grants to rural health facilities assumed to remain national? assignments ▪ KEMSA - Facilities assumed to manage funding for drugs? ▪ Health information systems - shared equaly between national and county governments ▪ Environmental Health - shared 50/50 between national and county governments? ▪ Provincial Administration - PMO costs shared by population among counties? ▪ Economic Stimulus Package - Additional stuff fully recruited and assigned to counties? ▪ NASCOP and NACC - Functions remain national? Source: World Bank calculations. spending to an alternative allocation of the same total amount following the CRA formula. In other words, it shows the difference between the per capita amounts that each county would be receiving under the proposed CRA formula, and the current costs attached to the delivery of devolved health services alone. The highly redistributive effect of the formula is clear. Figure 8-4: A massive reallocation of funds across counties under a �scally prudent scenario. Difference between current distribution of spending and projected allocation under a KES 97.5 billion equitable share scenario 3 2 1 Nyeri Nairobi Homa Bay Murang'a Kirinyaga Embu Kiambu Mombasa Nakuru Garissa Kisumu Kericho West Pokot Isiolo Elgeyo-Marakwet Laikipia Nyandarua Lamu Baringo Meru Taita Taveta Migori Tana River Machakos Siaya Wajir Tharaka Nithi Kajiado Makueni Busia Marsabit Samburu Nyamira Kisii Vihiga Kakamega Uasin Gishu Nandi Kitui Kwale Bungoma Trans Nzoia Bomet Mandera Turkana Narok Kili KES billions -1 -2 -3 -4 -5 Source: World Bank staff calculations. 99 Chapter 8: Allocating the county share equitably but carefully 8.29 There is a major risk that ‘leading counties’ will be unable to maintain existing service delivery levels. According to our simple comparison of (simulated) per-capita transfers and (current estimated) expenditure costs, half (23/47) of Kenya’s counties would not receive enough money from the equitable share transfer to meet the inherited costs of service delivery. In these counties, there will be a major challenge (unless additional transfers are delivered and own revenues are scaled up) to maintain essential services in operation. Streamlining service delivery costs cannot be done overnight. At the other extremes, historically under-privileged counties would receive allocations vastly in excess of immediate (to maintain existing services) needs. 8.30 While redistribution is in fact a core objective of the formula, un-intended negative effects could ensue. Equalization is of course, a deliberate aim of the formula―to correct the historical inequities in public spending in Kenya. But the transition to the new formula driven distribution needs to be managed carefully to ensure that: (i) health and other devolved services are not disrupted in the meantime; and, (ii) additional resources are effectively channelled toward expanding access to services in lagging areas. 8.31 As tempting as it may seem, increasing the percentage of the equitable share would provide only a partial solution and it would create additional risks. While it would reduce the risk of funding shortfall in leading counties, it could undermine sound Public Financial Management in historically under-served counties and create an unsustainable fiscal burden for the national government. 8.32 Mitigating strategies need to be worked out for both “cash poor� and “cash rich� countries. Table 8-3 highlights the contours of such strategies. To avoid major service delivery disruption and the risk of wasteful spending, detailed transition plans should be carried out for the main devolved sectors along three main dimensions: (i) service delivery; (ii) finances; and, (iii) human resources. Figure 8-5: Almost all counties “winners� under the CRA proposal but a major �scal burden on the center. Difference between current distribution of spending and projected allocation under a KES 203 Billion equitable share scenario 6 5 4 3 KES billions 2 1 Nyeri Kirinyaga Embu Lamu Homa Bay Isiolo Elgeyo-Marakwet Murang'a Laikipia Kericho Taita Taveta West Pokot Tharaka Nithi Mombasa Nyandarua Samburu Baringo Tana River Garissa Kisumu Kajiado Vihiga Siaya Nyamira Marsabit Kiambu Migori Busia Nandi Makueni Kwale Uasin Gishu Meru Machakos Nakuru Wajir Trans Nzoia Bomet Kisii Nairobi Kitui Bungoma Kakamega Mandera Turkana Narok Kili -1 -2 -3 Source: World Bank staff calculations. 100 Devolution without disruption – Pathways to a successful new Kenya Figure 8-6: Winners and losers of redistribution in the health sector. CRA-based allocation LESS current allocation (KES per capita) 900 800 Tana River KES 819 700 • Transfer under a 600 CRA formula exceeds actual costs 500 • Counties will have 400 Baringo surplus cash, which Kenya shillings might not be well KES-473 300 spent 200 100 0 • Current costs Tana River Turkana Lamu Garissa Mandera Samburu Tharaka West Pokot Bomet Isiolo Nyamira Marsabit Wajir Migori Siaya Vihiga Kwale Kajiado Kakamega Trans Nzoia Mombasa Kitui Nandi Makueni Busia Uasin Gishu Bungoma Embu Kisumu Nairobi Laikipia Murang'a Taita Taveta Homa Bay Kericho Nakuru Kirinyaga Nyandarua Machakos Meru Kisii Elgeyo Marakwet Nyeri Kiambu Baringo Narok Kili exceed transfer -100 under CRA formula -200 • Counties will not afford inherited staff -300 establishment costs • Public health -400 services could be disrupted -500 Source: World Bank staff calculations. Table 8-3: Risk mitigation strategies for "cash poor" and "cash rich" counties during the transition. A. “Cash-poor� – Transfer under CRA formula may not B. “Cash-rich� – Transfer under CRA formula represents cover actual costs significant increase in funding Types of transition challenges: Types of transition challenges: ▪ Service delivery – need to rationalize health services ▪ Service delivery – decide how to scale up services, through to prevent services from being disrupted. public provision or contracting out, and build capacity as well as ▪ HR – more staff than budget, need to revisit the staff investing in new health facilities where necessary. skills mix. ▪ HR – how to attract new staff and motivate existing staff. ▪ Finances – less budget, need to focus on priority ▪ Finances – increased budget, with possible low absorption effective interventions and efficiency gains. capacity, need to spend money well. Need strong advocacy within ▪ Management – capacity constraints, priority setting, county to secure funding for health. planning and results. ▪ Management - capacity constraints, priority setting, planning and results. Possible Counties: Possible Counties: 1. Mombasa 4. Elgeyo Marakwet 1. Tana River 4. Isiolo 2. Kiambu 5. Homa Bay 2. Kisumu 5. Lamu 3. Muranga Source: World Bank staff analysis. The Danger of scaling-up funding too quickly 8.33 The implementation of the proposed formula would result in dramatic increases in total resources transferred to “cash rich� counties. Because the formula is seeking to correct historical imbalances by reallocating money to historically marginalized counties, the biggest relative increases in funding will hit the counties least able to spend it (see Figure 8-7). Furthermore, much of the funding used to deliver existing services does not pass through district treasuries, so the actual increase in the volume of funding being managed subnationally is likely to increase even more dramatically than Figure 8-7 suggests. In contrast, many of the counties in the ‘cash rich’ group have local authorities that ran budget surpluses in 2009/10, as they did notfully spend their budgets (see Table 8-4). 101 Chapter 8: Allocating the county share equitably but carefully Figure 8-7: Some counties will receive major new funding. % difference between current distribution of spending and projected allocation under a KES 97.5 billion equitable share scenario 180 160 140 120 Percentage Increase 100 80 60 40 20 0 Meru Migori Taita Taveta Machakos Wajir Siaya Tana River Makueni Kajiado Busia Marsabit Kakamega Kisii Kitui Samburu Uasin Gishu Bungoma Nyamira Nandi Vihiga Kwale Trans Nzoia Bomet Mandera Turkana Narok Tharaka Nithi Kili Source: World Bank staff calculations. 8.34 The existing subnational PFM and service Table 8-4: Counties that will receive the biggest increases delivery systems have not been used to have least capacity to manage funds. managing such large volumes of funding. This Top 10 Local Authority Budget suggests that some marginalized areas may fail County Surpluses by county as percent of to spend the money allocated to them. This total revenue, 2009/10 could have the results that: (i) increased funding Tana River 45 fails to translate into any tangible improvement Wajir 43 in service delivery, which may even get worse Marsabit 25 at first; (ii) without careful attention this Tharaka Nithi 22 flood of cash without associated attention Kajiado 18 to system building may lead to fiduciary TaitaTaveta 17 mismanagement; and; (iii) the combined Kiambu 14 effect may be to undermine public confidence Kilifi 14 intergovernmental financing arrangements, Trans Nzoia 14 leading to a backlash. UasinGishu 13 Source: World Bank calculations. Phasing in the transition over time 8.35 In cash-rich counties, absorption bottlenecks could justify transferring function and resources over time. The Constitution and the Transition to Devolved Government Act provide that functions and funds should be devolved to the counties asymmetrically, as they develop capacity to assume them. While it is not explicitly stated in either law, it follows logically that funding should be held back until the associated functions are transferred. One possible way of dealing with this challenge is to ensure that, at least for the 2012/13 financial year, the equitable share is clearly budgeted for against each county through a national law setting out county votes. The money could be spent through national systems, but only with sign-off by a nominated official at the county level. This would help to ensure that devolved funds are clearly shown in budget legislation, while at the same time, avoiding to pump 102 Devolution without disruption – Pathways to a successful new Kenya vast sums of money into counties with no established PFM systems. This challenge will also require a rapid and systematic effort by the national government, to support counties to develop adequate PFM, HR, and service delivery capacity. 8.36 In cash-poor areas, stop-gap funding provisions could be introduced. In the short run the priority is to maintain existing service delivery levels, while equalization should be pursued over time. Box 8-6 provides an illustration of how both imperatives were safeguarded when Papua New Guinea decentralized. The same basic approach could be pursued in Kenya. The first step is to establish how much each county would need to maintain service delivery at existing levels, and to estimate funding gaps, once the equitable share transfers are factored in. These gaps could then be met via additional transfers, but fixed in nominal terms (adjusted for inflation) over a number of years. Over time, the redistributive elements of the equitable share transfers would dominate, as the transfer is calculated in proportion of domestic revenues. Ideally, gap-fill transfers would be phased out at a time when, equitable share transfers will have increased significantly, to allow most counties to meet their service delivery obligations, and when counties with the most glaring gaps will have had time to restructure and streamline their service delivery arrangements. Conclusions 8.37 Given the major transition challenges associated with the current proposed CRA formula, it would make sense to maintain the equitable share at, or around, 15 percent. This is not to say that aggregate county needs will not exceed this amount, but that meeting these needs in the least possible distortionary way will require mobilising additional transfer instruments. These could be conditional or unconditional, as long as they are allocated according to a different formula―at least in the transition period―taking into account funding shortfalls in counties which currently enjoy higher-than-average service delivery levels, and absorptive capacity constraints in counties that start from a very low level. Box 8-5: Is the 15 percent a minimum amount even during transition. The Constitution does not provide a direct answer to the question whether the equitable share (whether set at the minimum 15 percent or larger) may be reduced during the transition period. Nonetheless, the only reasonable interpretation is that it can be reduced. It would not be consistent with the concept of a phased transition to expect the national government to continue to fulfill functions while transferring the budget for such functions to counties. This reduction may mean that less than 15 percent of the revenue raised nationally is alocated to counties during the transition. The Constitution expects the allocation of functions to counties to be asymmetrical. Accordingly, the amount given to individual counties may also be varied during transition to reflect the functions each county has. However, with one exception (mentioned below), the equitable share must be unconditional even during transition. There may be political resistance to reducing the equitable share, especially below the 15 percent mark. A work around may be achieved by budgeting the full amount of the county allocation from the equitable share in the transfers to the county, but setting the condition that in circumstances in which the national government is continuing to deliver some functions, a propotion of those funds may applied to the defray the costs to the national government of carrying out those functions. Source: Legal option by Christina Murray. 103 Chapter 8: Allocating the county share equitably but carefully Box 8.6: How Papua New Guinea gradually introduced equalization. Papua New Guinea introduced a new intergovernmental financing system in 2008. The political negotiations involved in getting a sufficient number of MPs to pass the required constitutional amendments meant striking a deal to ensure no province would lose out as a result of the changes. Mechanisms of this kind are often called ‘hold harmless’ provisions. The new arrangements provide for provinces to share a percentage of national revenue called the Equalisation Amount. Over a five-year transition period, the provincial percentage share has gradually increased, starting from a base which gave provinces a slight increase over the previous year. The distribution of the Equalisation Share is based on a two-step process. First, an amount equal to the nominal value of the amount each province had received in the year before the new system began, was distributed. Second, the remaining balance is distributed on an equalisation basis. As the amount of the Equalisation Share increases each year, there is more available to equalise. At the end of the transition period, these ‘hold harmless’ arrangements stop, and the whole pool is distributed on an equalisation basis. Initially, it was feared that this would result in some provinces’ funding suddenly reducing, but it was decided to deal with this closer to the event. As it happens, these fears have not been realised. Because of strong national revenue growth, the equalisation pool has grown, and the amount available for all provinces is sufficient to ensure that any decrease is very small, compared to overall levels of funding. Source: World Bank staff analysis. 8.38 Limiting the size of the equitable share transfer would also maximise the available fiscal space to be mobilized via conditional grants. Conditional grants could be used to fill financing gaps at county level, but also promote the provision of critical programs and support a system of performance management at the county level. 8.39 It will also be critical to ensure that some funding is made available to maintain urban service delivery. Not only is the equitable share unlikely to be enough to meet the totality of county needs, especially in the more privileged ones (which are also the most urban), but they will also be entirely untied. At the same time, most counties―which will be responsible for resourcing urban services―will be dominated by rural constituencies that may not view urban development as a priority. Given the role that Kenya’s cities will continue to play in supporting economic growth, it will be key to find ways to secure urban service delivery and financing under the new set-up. 104 Devolution without disruption – Pathways to a successful new Kenya CHAPTER 8: KEY INSIGHTS / RECOMMENDATIONS. The CRA has recommended a formula for allocating the equitable share (constitutionally floored at 15 percent of national revenue) across Kenya’s forty-seven counties. The proposed formula takes into account the constitutional emphasis on equalization across counties. It is a good start, but it also has important gaps. In particular, its highly redistributive nature means that increasing funding to ‘lagging’ regions will necessarily imply fiscal stress for the national government and or for relatively better off counties. There is a basic dilemma: o If the aggregate pool to be transferred is just enough to cover existing aggregate spending on devolved functions, reallocations toward historically under-served counties would leave other counties under-funded, and risk interrupting service delivery. o If the pool is increased above and beyond needs, ‘lagging’ counties may face severe absorptive capacity constraints (to spend the windfall well and transparently), and the national government would have to make cuts in the services it currently delivers. *** Given the experimental nature of the CRA formula, and lack of clear estimates of actual county needs and capacity, it may make sense to limit the equitable share transfer at─ or close to─ 15 percent during a transitional period. Unmet needs at county level could be filled via other unconditional or conditional instruments. In counties which have been historically over-serviced (relatively), the key would be to avoid service delivery interruptions by extending ‘stop-gap’ funding to be phased-out overtime. In counties which have been historically under-serviced and marginalized, it will be vital to build up the capacity, to handle vastly increased funding efficiently and transparently. 105 CHAP TE R N IN E Strings attached? The scope for conditional transfers 9.1 It is critical that the intergovernmental transfers system is designed in light of explicit choices regarding conditional funding. This is essentially for three reasons. Firstly, devolution will occur in a tight fiscal environment, which means that transfers to county government―if not planned properly and costed altogether, rather than in a piecemeal fashion―will have a significant impact on the national government’s ability to deliver core services and functions. Secondly, because Kenya’s Constitution mandates that at least 15 percent of national revenue should be transferred to counties on an unconditional basis, the scope for conditional instruments is significantly reduced to start with. Finally, consideration of the equalization effects of conditional programs ought to be included in the design of the unconditional equitable share transfers. 9.2 There are powerful reasons why a significant share of transfers (above and beyond the 15 percent minimum) should be made on a conditional basis. While untied funding is vital to give meaning to local autonomy over allocative choices, conditional funding will also be important to support Kenya’s devolution. Conditional instruments can be mobilised to ensure that key national priorities are maintained across Kenya, or to address region-specific needs (including possibly marginalization). In addition, they could be designed strategically to mitigate the shortcomings of the equitable share transfer formula (for instance to provide additional ‘stop-gap’ support to service delivery) in the initial years of the transition to devolved government. Finally, conditional grants can constitute the backbone of a subnational performance management system. The rationale for conditional transfers in Kenya 9.3 Conditional transfers are a typical feature of devolved funding arrangements, despite the fact that they limit the extent to which local governments can freely decide over the types, quantities and manner in which services are delivered. Along with unconditional transfers and payments, they complete the range of instruments through which the central level can Figure 9-1: Instruments to support local service delivery. support service delivery at the local level. TRANSFERS There are also many differing, and often highly complex, variants of conditional GRANTS PAYMENTS transfers in use. As Bird and Smart for functions that are the for functions that are the observe: ‘in almost every jurisdiction, responsibility of the recipient governement responsibility of the transfering governement the world of intergovernmental grants turns out to be a complex and convoluted UNTIED confusion of labels, intentions, and realities.’1 Establishing clear definitions TIED and consistent terminology is therefore - Matching essential to any discussion of conditional - Non matching transfers. Figure 9-1 provides a simple Source: Adapted from Searle and Martinez-Vazquez (2007). 1 Bird and Smart (2009). 106 Devolution without disruption – Pathways to a successful new Kenya typology of conditional transfers, while Box 9-1 provides some definitional clarity, and both define the terminology used throughout this chapter. 9.4 While the extent to which “real� devolution is achieved can arguably be measured by the share of the funding that is transferred unconditionally, moving too quickly toward that goal can have adverse effects. Early ‘radical decentralizers’ like Uganda and Indonesia have gradually re-introduced conditionality, in response to the perceived failings of rapid decentralisation of unconditional funding.2 Clearly, a balance needs to be struck regarding the appropriate extent of conditional transfers. Hence, the main questions are really how much conditionality, and for what purpose. There are many types of conditional transfers allowing differing degrees of subnational autonomy depending on design (see Box 9-1). Box 9-1: Typology of conditional grants and their impact. ▪ Block transfers work as general purpose grants except that they are earmarked to a specific sector or sub-sector. Once this basic condition is fulfilled, the recipient has full discretion to allocate the funds across specific uses. For instance block transfers for education could be used to supplement teachers salaries in county X, but to purchase textbooks in county Y. Like general purpose transfers they have only an income and no substitution effect. This being said the empirical evidence is that, despite the fact that money is fungible, block transfers somehow stick to the sectors they have been earmarked to (a phenomenon known as 'flypaper effect'). ▪ Specific purpose transfers are different in that they typically restrict the choice of expenditures that can be financed (input based conditionality) or require that results benchmarks be met (output based conditionality). The latter are preferable in the sense that they preserve local autonomy over program choices. They are best suited to subsidize activities which are considered high priority by the central government, but low priority at the local level. Conditions can be specific to the function to which the transfer is provided or not. ▪ Specific purpose grants can include a matching component whereby the recipient must match the transfer out of own resources (following a pre-specified ratio). While matching requirements encourage local ownership, they may increase inequities to the extent that better off communities will have greater access to additional state resources. There the subsidy (because the grant subsidizes local spending) has not only an income effect, but also a price effect (making the subsidized good cheaper). Matching grants are particularly appropriate for addressing spillovers, precisely because these spillovers distort the cost/benefit of service provision at the local level. Source: Adapted from Boadway and Shah (2007). 9.5 It is important to stress that conditional transfers do not necessarily undermine devolution, but can in fact support it. In other words, there are legitimate reasons why the centre may wish to influence local spending, both in terms of uses and levels. They boil down to efficiency and equity considerations. On efficiency grounds, the central government may wish to promote expenditures on decentralized services (or investments), whose benefits spill-over across jurisdictions because local governments would otherwise under-provide them. Equity considerations may come into play when local governments provide public services (such as health), which have a redistributive role, and are therefore of strategic national interest. 9.6 It may make sense, in the early stages of devolution, to define a clear framework for the evolution of conditions, including a plan for phasing out conditionality. While some degree of conditionality is needed in the early stages of devolution, it may make sense to agree on a set of principles and rules for gradually increasing the share of transfers to be made unconditionally. That would assuage 2 Easton et al. (2010). 107 Chapter 9: Strings attached? The scope for conditional transfers the fears of the national government over ‘loss of control’, as well as providing guarantees that local autonomy will be respected. Sierra Leone provides an example of such a strategy (see Box 9-2), where conditionality on grants to local councils is to be phased out first, in terms of the degree of detail within the conditions, and then in terms of the conditions themselves. Box 9-2: A strategy for phasing out conditionality. Sierra Leone established local councils in July 2004. The legal and administrative framework is structured around a strategy that will gradually: 1. commence with a separate tied grant for each function devolved to local councils and a high level of detail on the associated conditions; 2. reduce the level of detail within the conditions, as councils are capable of accepting more authority and autonomy, until most of the funding for each function is a general tied (block) grant; 3. maintain sufficient funding through detailed tied grants, to give the central government capacity to achieve ‘national’ objectives as opposed to local objectives; and, 4. eventually combine most of the general tied grants into large untied grants, with the tied grants aimed at achieving national objectives continuing to run in parallel. Source: Adapted from Searle and Martinez-Vazquez (2007). 9.7 A ‘transition’ strategy may also be appropriate for Kenya. Kenya’s policy-makers need to strike a fine balance between two imperatives: seizing the moment of relative national consensus to empower counties as much as possible and, at the same time, avoiding the temptation of overloading counties in ways that could set them up for failure, and lead to a recentralization backlash. The goal of preserving county autonomy is clearly embodied in the constitutional guarantee that at least 15 percent of national revenue will be transferred on an unconditional basis. At the same time, the Constitution also provides that national government could take over devolved functions, in case of county failure. 9.8 The equitable share formula presented by the CRA is a good starting point, but it is experimental. The formula for allocating the equitable share across counties relies on high level proxies of county needs and costs. How accurately it will reflect the true costs faced by counties remains to be seen, but it is inevitable that discrepancies―both upward and downward―will occur, especially in those counties which have been historically under- and over-privileged. Moreover, counties’ capacity to deliver essential services (particularly in historically marginalized areas, where service delivery is particularly challenging) is as yet mostly untested. Given this, and until the formula for the horizontal allocation of the equitable share can be refined on the basis of experience, it may not make sense to increase the volume of funding being channelled through the equitable share, much above the constitutionally mandated floor of 15 percent of audited national revenue. Instead, additional funding could be provided on a conditional basis (with a degree of specificity to be determined) so that resources can be aligned as closely as possible to actual service delivery obligations, and so that key programs can be protected. 9.9 The equitable share transfer is explicitly designed to redistribute resources and promote equalizationacross counties, but within counties there is no obligation to allocate their resources in accordance with these objectives. Minimum national standards may well exist as policies (such as requiring a mandated number of hospital beds per thousand people) but will remain abstract in many counties if these mandates remain unfunded by the centre, or if they can’t effectively be monitored 108 Devolution without disruption – Pathways to a successful new Kenya and enforced. To ensure that minimum service delivery standards prevail throughout the country, complementary conditional grants could provide appropriate incentives for counties to allocate spending to services that matter for equalization. For instance, grants tied to the provision of health services could ensure that all counties allocate a minimum of resources to health care. They could also seek to ensure that positive externalities (such as from immunization) are fully factored into allocation decisions at the local level. South Africa provides an example of a grant designed to address positive externalities, with matching funds provided to teaching hospitals to encourage enrolment of non-local students, and use of hospital facilities by non-residents.3 Supporting national priorities at the local level 9.10 Kenya already allocates ‘tied funding’ at the local level: a first order priority should be to decide on the future of these funds. This is important because three of them taken together accounted for KES 38 billion (close to US$ 450 million)―or over 8 percent of 2010/11 ‘national revenues’. If these programs were maintained, and effectively became conditional grants, the share of national revenue to be transferred via the equitable share would need to be correspondingly less. Box 9-3: Existing funding arrangements that would be conditional if continued. To the extent that they remained earmarked to specific uses, these programs would have to be financed through conditional grants and their costs would not be included in the calculation of the county equitable share: Road Maintenance Levy Fund (RMLF): County share KES 11.3 billion. Road maintenance is currently funded through the Road Maintenance Levy Fund, financed by a fuel excise, and spent through the Kenya Urban Roads Authority (KURA) since 2010/11 (it had previously been spent by local authorities). If this funding is to continue to be tied to road maintenance, it would need to be an earmarked grant. Constituency Development Fund (CDF): KES 14.4 billion. While the future shape of CDF is unclear, if it continues to be dedicated for spending only on specific projects, it will effectively constitute an earmarked grant. Local Authorities Transfer Fund (LATF): KES 12.3 billion. Urban services are partly funded from the national government’s LATF. If the national government wishes to earmark funding for urban service delivery, this would not be factored into the calculation of the county equitable share. Source: World Bank staff analysis. The case and options for capital grants 9.11 Infrastructure investments would be obvious candidates for conditional funding on account of both efficiency and equity considerations. International experience suggests that: (i) Infrastructure gaps drive migration patterns (particularly labour migration, from poorly endowed and rural areas to better endowed and urban areas), above and beyond what would be optimal.4 In other words, where public service infrastructure is unequally distributed across space there are fiscally induced migrations that create externality costs such as congestion, crime, pressure on infrastructure, and environmental degradation, that could be alleviated by providing more equal and spatially even access to basic infrastructure; (ii) Subnational governments tend to spend a high proportion of their budgets on wages and salaries and relatively little on investment. There are political economy reasons for this, as public jobs are 3 Shah (2007). 4 In South Africa such migration is commonplace from rural areas to large urban centers such as Cape Town or Johannesburg that enjoy better housing, education and health infrastructure. 109 Chapter 9: Strings attached? The scope for conditional transfers a preferred solution for patronage, as well as logistical reasons, as investment spending typically requires much more complex budgeting and implementation. Yet public infrastructure is a major input into the production of capital intensive services such as education, health and transportation for which the national government may want to ensure minimum and uniform standards across the country. Other than through conditional grants, there is simply no way to ensure that similar amounts will be devoted to maintaining and expanding infrastructure at the local level. 9.12 Capital grants, motivated by the objective to provide minimum uniform standards of service delivery, would be justified from an equalization perspective and address negative externalities of ‘excessive’ migration. In Kenya, years of inequitable service delivery have distorted the distribution of public infrastructure, so that the transition to more equitable service delivery will require significant infrastructure investment, and cannot be achieved overnight. Currently, Kenya ‘earmarks’ over KES 11 billion (US$ 130 million) to road maintenance through the Road Maintenance Levy Fund (financed by a fuel excise). At minimum, there should be a decision on whether the RMLF―at least its rural and urban roads components―should be converted into a conditional grants program, and whether the current allocation criteria should be retained. A more ambitious agenda would be to extend the scope of the transfer program, potentially consolidating RMLF and CDF into a single transfer scheme. 9.13 A capital grants scheme could be based on a transparent formula. Alternative options―such as project specific grants (Box 9-5)―may go against the equalization objective. But designing such a scheme would be difficult in view of two basic requirements. First, it would be difficult to avoid distorting incentives for local provision of infrastructure―i.e. under-provision out of own─or unconditional sources, in the expectation that transfers from the centre would fill the gap. This is usually addressed by requiring matching contributions from the recipient,or by targeting central funding to new projects that would otherwise not be funded or by monitoring local provision and providing incentives to reward it. Second, it is also hard to ensure that the complementary inputs to service delivery (namely staff and operating capital) are provided in parallel by the local government. In Kenya, the experience of CDF has demonstrated this to be a major bottleneck, with the potential to dramatically reduce the returns on the initial capital investment. A simple capital grants scheme model (Box 9-4), developed by Petchey and Macdonald allows to maintain overall increases in the capital stock, while closing infrastructure gaps across regions. The case and options for urban service grants 9.14 Kenya’s new devolution arrangements make county governments responsible for financing urban service delivery. Urban service functions (water, sanitation, parks, solid waste management and transport) are constitutionally assigned to county governments, along with the tax bases that previously sustained municipal and town councils (property tax, entertainment tax and business licensing). Also, instead of elected municipal and town councils as presently exist, the new county governments will establish appointed city and municipal boards and town committees (referred to as urban boards) to run urban services.� 9.15 The urban boards will depend on the county government for funding, but there is a risk that they will fall short of needs. The boards could be resourcedeither through transfers from the county budget, � The new urban boards will be appointed under the Urban Areas and Cities Act. The Act makes a distinction between city and municipal boards and town committees in terms of their powers and functions. City and municipal boards have corporate status; town committees do not. This makes it unclear to what extent town committees will function as separate budgetary and staffing entities. This paper ignores this issue, and assumes that all urban areas with a population over 10,000 will be sufficiently empowered to administer urban services with relative autonomy. 110 Devolution without disruption – Pathways to a successful new Kenya Box 9-4: A simple model for allocating capital grants in developing countries. A simple grant scheme model was developed (Petchey, Macdonald and Josie, 2004) to specifically respond to needs of developing countries with: (i) low overall levels of infrastructure; and, (ii) important regional disparities. The model relies on an endogenously determined per capita public capital standard, and sets out a path for ‘backlog’ regions to catch up to the standard, while also allowing ‘surplus’ regions to continue increasing their per capita capital stock. The model involves creating a per capita capital stock index for each subnational unit and then ranking them into two categories as ‘backlog’ or ‘surplus’, depending on where they fall relative to the national average (or whatever other standard). The total pool of funds available for transfer is divided into (i) a backlog component to address spatial disparities, and (ii) an economic efficiency component to ensure that public capital continues to rise. The respective shares of these components are decided as a policy parameter depending on how quickly equalization is sought, and its perceived importance, relative to the goal of increasing capital across all units. The backlog component is allocated across units according to their relative distance from the average. The economic efficiency component can be allocated simply on a per capita basis. In year two of the program, the backlog indexes are recalculated to take into account changes as a result of the transfer, and the same process is followed. At the end of the process (year “t�) all regions/units will converge to the per capita uniform standard (average) set in period 1. In Kenya, the most difficult requirement would be to calculate the per capita public capital stock of each county. A way around these difficulties would be to weigh the total capital stock estimate, according to county proxies such as county GDP. Two critical decisions include: The total size of the pool and the relative shares of the two components. The larger the ‘backlog’ component share, the larger the pool must be in order for per capita public capital to continue to grow in ‘surplus’ areas. Source: Adapted from Petchey and Macdonald (2007). or delegation of power to collect revenues (either as charge for services or by collecting and retaining other county taxes). Since the new bodies are management bodies, not governments, they cannot impose taxes and charges, unless authorized by a national or county law. In this context, there is a risk that urban services may be inadequately funded under the new arrangements. Because rural residents will dominate most counties, county governments may choose to preference rural services, and redistribute the revenues raised from urban residents, away from recurrent costs of urban services and towards other purposes. This may jeopardize the economic development potential of Kenya’s urban areas. 9.16 County governments should delegate to urban boards the revenues currently collected by local authorities but this is unlikely and insufficient. Indeed current LA revenue sources will be the only source of own-revenues for county governments, which would be left entirely dependent on transfers without them; the incentives they face to resource urban areas (both from a fiscal and political point of view) are minimal. Urban areas will remain transfer dependent in any event. Municipal functions are currently being funded by own source revenues (to become county revenues) and significant transfers from the centre under the LATF program, which accounts for over KES 12 billion. 111 Chapter 9: Strings attached? The scope for conditional transfers Box 9-5: Five design issues for earmarked urban grants. How much should the total pool be? At minimum, the total pool (including central transfers and county contributions) should be enough to maintain existing service levels. Their current cost can be estimated by adding the total own source revenues of local authorities to LATF transfers. That total should be discounted to account for ‘non core’ costs currently incurred by LAs (such as allowances to ward councilors) and minor capital spending on projects unrelated to urban functions through LATF (such as education). The total would also have to exclude current spending on/by existing councils that fall below the new 10,000 threshold as well as Nairobi, Mombasa and Kiambu, where county resources will be overwhelmingly spent on urban functions. How to allocate the grant between counties? A formula could be designed to close the gap in each county between the (potential) tax revenues collected from urban residents, and the overall cost of maintaining service provision. Alternatively it could simply follow the current LATF formula that currently allocates most of the funds according to urban population shares. How much should counties continue to allocate to urban services themselves? The main condition of the grant should be that counties continue to fund ‘core’ urban services out of own revenues. The exact amount should reflect county fiscal capacity, and the cost of urban services there. The simplest way to capture both these objectives is by requiring counties to allocate the amount of urban revenues they raise to urban services expenditure. Earmarking the grant this way would require counties to do no more than return the revenues they raise from urban areas, in the form of services. The initial allocation of the grant would be based on the previous revenue capacity of city, municipal and town councils in that county, with revisions based on subsequent reporting of revenues generated. What other conditions should be attached to the grant? In order to be effective, the urban services grant should be earmarked at least to urban services. A very simple form of earmarking would be to provide a menu of purposes that the grant could be used to fund (i.e. a list of core urban services). Earmarking could also include more detailed conditions such as: (i) minimum service standards; (ii) specific items that must be funded; (iii) maximum share of expenditure on staffing or other administrative overheads; (iv) procedures for engaging citizens in planning, budgeting or evaluation; (v) performance conditions (for a proportion of the grant); and/or, (vi) reporting requirements. How should the grant be paid? Two options include: (i) direct payment to the county government (with the risk that funds could get ‘stuck’ there; or, (ii) a ‘pass-through’ option where transfers would be budgeted at county level, but paid directly to the urban boards. The latter however would limit the discretion of counties to allocate the funds across urban areas within them. Source: Adapted from Petchey and Macdonald (2007). 9.17 The national government could help to ensure that urban services are adequately provided through earmarked urban service grants. Grants might be paid either to the county governments or to the bodies that manage the services (urban boards). However, most county revenues are fungible and can be allocated to any purpose. An earmarked grant might have the effect of allowing counties to reduce their own funding to urban services. County governments might abandon their own responsibility for taking care of urban services, and blame the national government for inadequate service delivery in urban areas. 9.18 Any earmarked transfer should be designed to ensure that counties maintain their own level of funding for urban services. This could be achieved through an additionality clause, which would bind counties to maintain a certain level of funding to urban services. A county that did not allocate an adequate amount would not receive the earmarked grant. Box 9-6 below discusses the main design issues for such an earmarked grant. 112 Devolution without disruption – Pathways to a successful new Kenya Box 9-6: Other types of conditionality. While conditional grants are typically tied to specific service delivery objectives (or financial conditions in the case of matching grants), there are additional options: ▪ Tied grants can include administrative conditions on the environment in which services are being provided (such as doctors’ qualifications or the existence of standards of procurement) and can therefore influence the systems of delivery at county level. ▪ Information conditions can also be specified (basically data collection and sharing requirements – such as are currently attached to LATF) to provide all stakeholders (central and local governments as well as beneficiaries) with better information on: (i) the effectiveness of the grants against specific target outcomes; as well as, (ii) a better understanding of the service delivery environment in counties. Ultimately, this can provide a better base for service provision planning, and for comparing standards of service across jurisdictions. Source: Adapted from Searle and Martinez-Vazquez (2007). Filling gaps in the transition 9.19 The equitable share formula, through which most funding for Kenya’s counties will be channelled, will create a number of funding gaps. Because it uses high level proxies of needs and costs and includes only simple indicators (there are very good reasons for that), the formula is not complex enough to account for the specific needs of particular counties, communities or regions. Until the formula is eventually revised with more and better information on actual needs faced by counties, time-bound discrete conditional grants programs could be used to facilitate the transition. Specific needs include funding for: (i) services with significant spillovers; (ii) programs to address region-wide patterns/drivers of slow development (such as faced by the arid and semi-arid regions of Kenya); and, (iii) stop-gap temporary transfers to allow for adjustment in counties likely to experience a significant drop in financing from the national government in the new architecture. 9.20 Provincial hospitals constitute a good example of county services with significant spillovers. It is not yet clear which level of government will be responsible for them under devolution. They may not receive adequate funding if left entirely at counties’ discretion. There are only eight provincial hospitals, which operate as secondary referral facilities providing more advanced services than can be offered at the two hundred and eighty or so district hospitals. They are correspondingly much more expensive to operate, but if responsibility for provincial hospitals is devolved to county governments, this additional cost will only burden the eight counties in which these hospitals are located. If the equitable share is allocated according to a simple formula, it is unlikely that such a formula could be constructed to give counties with provincial hospitals increased funding to correspond to this increased cost. Furthermore, funding provincial hospitals through unconditional funding invites the possibility that the county in which the hospital is located will either de-prioritise this expense, with consequences for citizens in neighbouring counties who cannot influence the decision, or impose cost and other barriers to non- county residents, using the service. 9.21 Some of Kenya’s counties, typically clustered in the arid and semi-arid north, face specific challenges to development. These challenges are simply not well address by a blanket national formula, because they have to do not only with low starting points in terms of infrastructure and access to services, but also to deeper patterns of marginalization, and to the unique geographic conditions and socio- economic makeup in these regions. The Constitution specifically provides for an ‘equalization fund’, equivalent to no less than 0.5 percent of total revenues. The fund, however, is too small to make any meaningful impression on the lagging service infrastructure gaps, and it is likely that other instruments 113 Chapter 9: Strings attached? The scope for conditional transfers will need to come into play. Conditional grants targeted at specific regions or groups of counties could be advanced on a project basis, so as to maximize efficiency gains (without hurting equity objectives), tied to infrastructure (to avoid being absorbed into unproductive consumption spending), and possibly matched to an appropriate county government contribution; in order to encourage efficient use of the funds and ownership of the resulting assets. 9.22 Some counties will experience significant funding shortfalls during the transition and service delivery could suffer. The bulk of transfers to county governments will take place through the general purpose equitable share, creating ‘cash rich’ and ‘cash poor’ counties (in comparison to their pre-devolution funding shares). There will be challenges in each of these groups. The newly ‘cash rich’ counties―i.e. those which were historically underserved―will experience a windfall gain relative to their inherited service delivery obligations. There, the main challenge will be to ensure that these additional resources are well spent. Additional grants can play a role in ensuring that this happens in the context of a local performance monitoring system (see below). As for the newly ‘cash poor’counties―i.e. those that that were relatively over-served on a per head basis prior to devolution―(there will be a greater or lesser number of those depending on the magnitude of the equitable share), they will possibly find it difficult to maintain funding for service delivery at existing levels. The result would be fiscal stress and the need to make painful potentially counter productive adjustments (such as by cutting non-salary operating expenditures on drugs and supplies in the health sector, while maintaining staff and facilities in operation). 9.23 Targeted conditional transfers could be mobilised to fill these gaps. They could be set at a fixed nominal amount―corresponding to the estimated shortfall or a fraction thereof―and limited in time so as to smoothen the adjustment process, while avoiding to create an entitlement, that would defeat the purpose of equalization. Likewise, they could be tied to specific uses or sectors―such as those most likely to be cut under fiscal stress scenarios. Depending on the nature of the functions assigned to county governments, conditional grants may also be needed to address asymmetric function assignments. Stimulating performance at the local level 9.24 As part of a subnational performance monitoring system, grants can help to spur healthy competition between counties. Performance-based grant systems (PBGSs) aim to provide subnational governments with appropriate incentives to improve whichever aspect of performance is targeted. This way, they seek to reduce some of the downside risks associated with devolution, and to improve the efficiency and effectiveness of devolved governments. Performance based grants can be grouped into two broad categories: (i) compliance grants based on absolute (minimum) benchmarks with an ‘all or nothing’ allocation; and, (ii) those based on relative performance with grants allocated in proportion to performance. Compliance based grants condition access to funding on meeting basic minimum standards, and are therefore focused on measuring a subnational government’s ability, to perform its basic functions. They often involve capital grants so that failure to meet performance benchmarks does not undermine subnational governments’ ability to meet their recurrent obligations. In practice, sanctions for non-compliance or weak performance have often been altered to ensure minimum funding levels, albeit with more aggressive mentoring and supervision. Such systems are intended to be integrated into national inter-governmental fiscal transfer arrangements, rather than added on separately without regard to their impact on overall vertical and horizontal distribution of resources.� � Steffensen (2011). 114 Devolution without disruption – Pathways to a successful new Kenya 9.25 As well as being either ‘absolute’ or ‘relative’, performance grants have a further two main dimensions along which they vary. First, they can be tied to sector-specific service delivery objectives (e.g. placing conditions on health sector outputs), or they can be granted on cross-cutting administrative and/or information conditions to encourage better systems and, ultimately, better governance at the local level (e.g. conditions focusing on revenue collection, planning and budgeting, accountability, general governance). A second important dimension of a PBGS is the extent to which funds are discretionary (i.e. how strict the earmarking is). 9.26 In practice, most PBGSs have typically focused on improving cross-cutting systemic conditions relating to Public Financial Management (PFM), for example planning and budgeting, and have allowed subnational governments quite broad discretion in sectoral allocation decisions, limiting earmarking, to ensure a focus on capital investment rather than recurrent spending. As well as underpinning Kenya’s own Local Authorities Transfer Fund (LATF), this approach has been used widely across Africa (pioneered in Uganda with support from United Nations Capital Development Fund (UNCDF), and later developed in Ghana, Mali, Sierra Leone and Tanzania) and in Asia and the Pacific (Bangladesh, East Timor, Indonesia, Lao PDR, Nepal and Solomon Islands). These schemes have typically been heavily donor financed and designed (broadly following a model developed by UNCDF), with some evidence over time of increased government buy-in signalled by increased budget allocations�. 9.27 By contrast, output-based grants place a heavy emphasis in their conditions on sector-specifc results. While having some degree of sectoral earmarking, they typically allow significant discretion in how funds are spent (they do not specify exactly what inputs can be purchased), preferring to empower subnational governments, to make the funding Figure 9-2: The main dimensions of performance-based grant design. allocation and programme design decisions they deem Extent of discretion in use of funds? necessary, to achieve the targeted results. They seek Discretionary Heavily earmarked to focus recipients’ attention on the results chain, and to Focus of Cross-cutting Most PBGSs have cross-cutting think through how the sector conditions on (overall PFM, conditions and are discretionary type of governance, across sectors but earmarked to service delivery system will performance etc.) capital investment support improvements in improvement each link of the chain, until sought results (outputs) improve. Output-based grants focus They do not seek to target conditions on sector spectoral earmarking, but the focus is on outcomes since these are outputs rather than inputs usually to some extent, Sectoral out of the hands of public sector managers.� Figure Source: World Bank staff analysis. 9-2 illustrates these two dimensions graphically. � Ibid. � Shah (2007). 115 Chapter 9: Strings attached? The scope for conditional transfers 9.28 There are three common elements to most PBG schemes built along the lines of the UNCDF model/ blueprint: (i) the capital grant scheme; (ii) a performance assessment process; and, (iii) capacity building of the recipient subnational government. The capital grant scheme provides for multi- purpose grants, which although linked to capital investment projects, are largely discretionary. A crucial issue here is that transfers must be sufficiently large to provide sufficient incentives for local governments to meet the associated conditions, although relatively small grants of around US$ 1-4 per capita may be sufficient to achieve this. The performance assessment process is comprised of indicators, which are measured annually and usually track processes and systems (e.g. for Public Financial Management) and can be divided into two broad categories: (i) binary ‘yes or no’ measures which measure fulfilment of minimum conditions; and, (ii) performance measures which allow for different gradations of performance to be tracked. Thirdly, the local government capacity building component usually comprising of both demand-and-supply-driven interventions, allows subnational governments to respond to weaknesses identified in the performance assessments.9 9.29 Steffensen (2011) reviews lessons from international experience with Performance-Based Grant System (PBGSs) in fifteen countries. Achievements and benefits cited in the study include improved core administrative functioning and Public Financial Management, with some evidence of increased own-source revenues, where fiscal effort is a targeted aspect of performance. Transparency and accountability of subnational governments has also improved, as has the efficiency and effectiveness of capacity building support programs to subnational governments, as they have been able to respond directly to weaknesses revealed through performance assessments, and their results in turn have been measureable by the same means. PBGSs are typically heavily donor driven, and have indeed helped to facilitate improved coordination between donors. There are also indications that PGBSs have led to better infrastructure and service delivery outputs, although it is stressed that it is too early in the life of most such schemes to say conclusively. 9.30 The report also identifies a number of common challenges revealed from experience with PBGSs. First, because they tend to focus on performance improvement in cross-cutting systemic reforms, it is almost impossible to say conclusively whether they havelead to tangible improvements much further down the results chain, at the outcome and impact level. Second, PBGSs are vulnerable to external factors which can undermine their efficacy, including conflict, poor revenue assignment, parallel funding streams which dilute the PBGS incentive effect, and disbursement delays. Weak management capacities at the central level can also hamper implementation, and ultimately there has to be political will to withhold funding to non-performing subnational governments―without this the whole exercise is undermined. On the technical design side, choice of performance measures is often crucial, with poor choices leading to systems which are either unfair (because they hold governments accountable for things beyond their control), or because they generate perverse incentives, by encouraging governments to ‘game’ the system by hitting the target rather than providing the underlying improvement it is in fact trying to measure. Overall the author stresses that PBGSs should not be seen as a panacea for all the problems associated with decentralisation, and concludes by suggesting a number of core pre-requisites that must be in place for a PBGS to be most useful and effective, together with recommendations for introduction (see Box 9-8). 9 Steffensen (2011). 116 Devolution without disruption – Pathways to a successful new Kenya Box 9-8: Pre-Requisites for PBGS and recommendations for implementation. Pre-requisites: ▪ Strong policy support for performance incentives and political will to sanction weak performers. ▪ Document strengths and weaknesses of previous approaches through solid analytical work. ▪ Robust, careful and participatory PBGS design. ▪ Subnational government must have sufficient autonomy to genuinely control own performance. ▪ Capacity building arrangements must be linked to performance with a sensible mix of supply-and-demand-side interventions. ▪ PBGS operations, measures and outcomes are highly transparent and publicly disclosed, in particular the results of regular performance assessments. ▪ Support to subnational governments by central government and donors must be stable, timely, long-term, predictable and well coordinated. Recommendations for design and implementation: ▪ Invest sufficient resources and time in proper design, as PBGS are technically demanding. ▪ Ensure links to other dimensions of decentralization (e.g. IGFT system, PSR and HR). ▪ Ensure that all stakeholders understand potential benefits and challenges from the outset. ▪ If using pilots, design them strategically and realistically to optimize opportunities for roll out. ▪ Indicators should be appropriate, measure performance attributable to the SNG, measure both key performance of the SNG and PBGS objectives, and derive from statutory or regulatory frameworks. ▪ Start simple, focusing on critical and core SNG performance areas, and refining things over time. ▪ All guidelines and procedures must be clear, coherent, user-friendly and widely disseminated. ▪ Credible performance assessment that is aligned with the SNG’s planning and budget cycle. ▪ Establish effective coordination bodies to endorse assessments and oversee implementation. ▪ Integrate a capacity building support strategy, with both supply-and-demand-side interventions. ▪ Establish sound M&E systems, to track results and provide basis for ongoing PBGS adaptation. Recommendations for implementation: ▪ Minimise exemptions from the ‘rules of the game’, which compromise the entire system. ▪ Ensure transparency and communication at all phases of implementation. ▪ Provide technical and capacity building support to core agencies responsible for PGBS implementation. ▪ Ensure PGBS is well anchored in central policy-making bodies. ▪ Regularly review and follow up on implementation arrangements. ▪ Follow up and use M&E information, address complaints and regularly adjust the system in transparent ways and in consultation with all stakeholders. Source: Steffensen (2011). 117 Chapter 9: Strings attached? The scope for conditional transfers CHAPTER 9: KEY INSIGHTS / RECOMMENDATIONS. It is critical that the intergovernmental transfer system is designed in light of explicit choices regarding conditional funding because of: (i) Kenya’s tight fiscal environment; (ii) the constitutional guarantee that the equitable share will be purely unconditional; and, (iii) specific merits of conditional funding including for equalization purposes Conditional instruments could be mobilized essentially for three purposes: o to guarantee that national priorities will continue to be funded under devolution; o as part of a transition ‘hold harmless’ strategy to maintain service delivery in historically over-served counties; o to underpin a subnational performance management system. *** Given the radical and experimental nature of Kenya’s devolution, it may make sense to maximize the scope for conditional funding (and minimize that of unconditional resourcing), if only to ensure that essential programs will remain funded. To preserve the spirit of the Constitution, this could be done as part as an explicit transition strategy, with a clear plan for phasing out conditionality over time. Conditional grants programs could be considered specifically for: o capital projects, possibly folding together the current RMLF and CDF programs; o supporting urban service delivery; o addressing the specific unmet needs related to services with externalities and regions specific challenges; o providing temporary stop-gap support to counties at risk of experiencing severe fiscal shortfalls during the transition. Conditional grants could also be considered to foster inter-county performance, and promote a system of performance monitoring at the local level. 118 CHAP TE R TE N Making the most of the equalization fund 10.1 The equitable share and―possible―additional conditional grants may be insufficient to address deep spatial inequities. The bulk of equitable share transfers will go to funding existing services, as well as the administrative costs for setting-up and running counties, and the relative crudeness of the formula (in which poverty is a minor component) does not allow for targeting specific areas or gaps over time. In turn conditional transfers will have to be over and above the constitutionally mandated 15 percent, subjecting them to the overall―tight―budget constraint, and these limited funds will need to be allocated strategically across a number of possible alternative uses and objectives. 10.2 The equalization fund is meant to overcome these constraints, but there will be limits to what it can achieve. Article 204(2) of the Constitution, calls for an Equalization Fund to be used to provide basic services―water, roads, health facilities, and electricity―to marginalized areas to the extent necessary to bring the quality of those services in these areas to the level generally enjoyed by the rest of the nation. The fund amounts to a minimum of 0.5 percent of total nationally-generated revenue, computed on the basis of the most recent audited revenue accounts, as approved by Parliament. As per the Constitution, budget estimates submitted to Parliament will include estimates of spending from the fund. How much does the Fund amount to? 10.3 The Equalization Fund amounts to very little money, too little to make a meaningful impact on marginalization. Based on national revenue estimates for 2011/12, the Equalization Fund amounts to approximately KES. 3.6 billion at minimum.1 This amount is roughly equal to the annual salary costs of Kenyatta National Hospital, but less than that of the University of Nairobi. Figure 10-1: How much does the Equalization Fund amount to (KES billion)? In comparison to the CDF (KES 1784 5 billion), the Equalization Fund is equally marginal, amounting to only a little 4 more than the amount by which the CDF increased between 2010/11 and KES billion 3 2011/12. The entire annual allocation for the Equalization Fund can construct 2 only about forty five kilometers of tarmac road!2 These comparisons 1 demonstrate just how materially 0 unimportant the Equalization Fund is CDF increase (2010/11 Ministry of Northern Equalization Fund Kenyatta National University of Nairobi vs. 2011/12) Kenya budget (2012/13) Hospital annual salaries annual salaries likely to be (see Figure 10-1), unless it Source: World Bank staff analysis. is deployed strategically. 1 Excluding all donor and domestic borrowing, as well as appropriations-in-aid (AIAs), the total nationally-generated revenue in 2011/12 is projected at KES 725.8 billon. 2 Based in indicative average road construction costs obtained from the Kenya Roads Board (KRB). 120 Devolution without disruption – Pathways to a successful new Kenya 10.4 Recent disparities in the actual allocation to the Equalization Fund not only highlight the need for strict adherence to the applicable Constitutional provision; they alsomagnify the fund’s smallness. Through the Budget Policy Statement released in April 2012, the Government allocated KES 6.7 billion to the Equalization Fund, an amount equivalent to 0.8 percent of projected shareable revenue for the 2012/13 fiscal year3 (see Table 10-1). The Budget Statement presented to Parliament by the Finance Minister on June 14 2012 contained a different allocation, of KES 5.5 billion, to be split between the 2011/12 and 2012/13 fiscal years4. The allocation represented 0.3 percent Table 10-1: How much will actually be allocated to the Equalization Fund? of the shareable revenue in each All figures in KES Billions 2011/12 2012/13 2013/14 2014/15 unless otherwise stated year. Clearly, the policy adopted will need to remain consistent with the 1. Shareable revenue 725.8 868.1 999.4 1,130.5 Constitutional provision as far as the 2. National Government 684.1 701.4 806.2 909.4 allocation Equalization Fund is concerned. Yet 3. Transfer to County 41.7 166.7 193.2 221.1 even where the allocated amount Government exceeds the constitutional minimum of which: Equalization Fund 0.0 6.7 4.3 5.0 of 0.5 percent, the point raised earlier Equalization Fund as 0.0% 0.8% 0.4% 0.4% regarding the smallness of the fund still percent share prevails. Moreover, the fund will grow Source: MoF (2012). marginally over its 20 years lifespan. 10.5 Such a small fund might better serve as ‘seed money’, to leverage additional resources into a bigger, more effective fund. One possibility would involve the passage of an ordinary law to allow additional resources to be added to the Equalization Fund, including contributions from development partners. What objectives should guide its allocation? 10.6 The object of the Equalization Fund is still unclear. The Constitution defines marginalized people, but the Equalization Fund seems to aim at marginalized areas, which the Commission on Revenue Allocation (CRA) is tasked to identify.5 Two dimensions which CRA could consider are: (i) areas in which basic services fall below the level generally enjoyed by the rest of the nation; and, (ii) the location of marginalized groups and/or communities (see Box 10-1). The objective of the fund could become more clear depending on which of the two aspects gets the most weight in CRA’s identification criteria. 10.7 Clarifying the fund’s objective will involve establishing whether areas of poorer access to services align to areas where marginalized groups are located. A recommended approach is to construct a ranking of counties using a composite index of service access. 10.8 The fund should not aim to give all areas access to the same infrastructure, instead it should focus on key service gaps. Although many Kenyans expect that devolution will trigger economic development in their respective counties, growth and economic activity (no matter what infrastructure is being built) will continue to be―and increasingly so―spatially concentrated. Therefore, infrastructure investments 3 As defined in the CRA Act, shareable revenue includes income taxes, import duty, VAT, excise taxes and non-tax revenues. 4 According to the Budget Statement, the Equalization Fund allocation in 2011/12 will be utilized on water, roads, health facilities and electricity; while the allocation for 2012/13 will be set aside awaiting the determination and publication of CRA’s criteria for identifying marginalized areas. 5 According to Article 216(4), CRA shall determine, publish and regularly review a policy in which it sets out the criteria by which to identify the marginalized areas for purposes of the equalization fund. 121 Chapter 10: Making the most of the Equalization Fund Box 10-1: Who are the marginalized communities? Article 260 of Kenya’s Constitution defines “marginalized community� as: (a) a community that, because of its relatively small population or for any other reason, has been unable to fully participate in the integrated social and economic life of Kenya as a whole; (b) a traditional community that, out of a need or desire to preserve its unique culture and identity from assimilation, has remained outside the integrated social and economic life of Kenya as a whole; (c) an indigenous community that has retained and maintained a traditional lifestyle and livelihood, based on a hunter or gatherer economy; or, (d) pastoral persons and communities, whether they are: (i) nomadic; or (ii) a settled community that, because of its relative geographic isolation, has experienced only marginal participation in the integrated social and economic life of Kenya as a whole. Article 260 also defines “marginalized group� as a group of people who, because of laws or practices before, on, or after the effective date, were or are disadvantaged by discrimination on one or more of the grounds in Article 27 (4); according to Article 27 (4)-(4), “the State shall not discriminate directly or indirectly against any person on any ground, including race, sex, pregnancy, marital status, health status, ethnic or social origin, colour, age, disability, religion, conscience, belief, culture, dress, language or birth�. Source: Kenya Constitution. should be targeted first and foremost, at social services delivery, to ensure that all Kenyans have access to the same opportunities (possibly involving migration). Following the World Development Report (WDR, 2009), the focus should be on ensuring that the well-being of individuals does not depend excessively on location. Essentially, the WDR proposes that public investment should target people not places. To a large extent this is also the message which underpinned Sessional Paper #10 of 1965. What will the fund finance and where? 10.9 Is the fund intended only for infrastructure? It is unclear whether or not the fund will be restricted to funding capital development (as is the case with the CDF); or, whether it can be also spent on recurrent expenditures. If the former option is retained (as the Constitution appears to indicate), there is a significant risk―given the experience of CDF and the weak planning and budgeting capacity in ‘lagging’ counties―that capital spending decisions will be made in a vacuum, potentially creating un- planned liabilities (as complementary funding will be needed for operating the facilities), or failing to result in social services delivery improvements (if these additional funds cannot be mobilized). Should the fund be eligible for recurrent spending, then it might be better utilized. For instance to top-up the wages of critical human resources (such as health workers in remote areas). Indeed, the inability of these regions to attract and retain qualified doctors, nurses and teachers, is a leading factor behind poor social outcomes, a characteristic feature of Kenya’s North. 10.10 If the fund crowds-out existing funding for lagging areas its impact could be nil or negative. There are currently a number of funds and programs that implicitly or explicitly seek to promote greater access to services in lagging regions. Examples include, programs under the Ministry for the Development of Kenya and other Arid Lands, as well as, under the Ministry for Special Programmes. The former coordinates activities in at least ten arid counties (see Figure 10-2), with an annual budget of approximately KES 4 billion, roughly the size of the Equalization Fund.6 The latter manages Disaster Risk Reduction (DRR) programmes, many of them also in the arid lands. The future of these programs─and of the ministries that oversee them―is unclear. After decentralizing a considerable portion of Kenya’s budget, and faced with a tight budget constraint it is possible that the national government will find it necessary to 6 (i) The Ministry’s focus includes: infrastructural development; township planning; promotion of livestock and related industries; water supply; natural resources management and mineral resources exploration; tourism development; human resources development; irrigation; and tapping solar and wind energy. (ii) Presidential Circular No. 1/2008. 122 Devolution without disruption – Pathways to a successful new Kenya downscale current funding targeting marginalized regions. If so, the equalization fund would not bring additional resources to these areas. Therefore it is key that the fund’s objectives and priorities are conceived as part of a comprehensive set of instruments, targeting lagging regions, with a clear goal of additionally. 10.11 Given the small size of the fund it may be wise to define its geographic focus narrowly. If CRA selects counties currently covered by the Ministry of State for the Development of Northern Kenya and other Arid Lands―this is the area commonly called arid and semi-arid lands (ASALS)―it would represent eighty per cent of the country’s land mass (see Figure 10-2). Clearly, distributing the rather small Equalization Fund across such an expansive and diverse region, would undermine the fund’s effectiveness. More generally the limited size of the fund calls for focus, in space, targets and possibly time (i.e. by allocating the bulk of funds to specific areas / projects in any given year on a rolling basis as opposed to spreading it across to many dispersed programs). Table 10-2: Northern Kenya’s main policy challenges and counties affected. Northern Kenya Pastoralism Aridity Baringo, Garissa, Ijara, Isiolo, Lamu, Baringo, Garissa , Isiolo, Kajiado, Baringo, Garissa, Isiolo, Kajiado, Kilifi, Mandera, Marsabit, Samburu, Tana Mandera, Marsabit, Narok, Samburu, Kitui, Kwale, Laikipia, Lamu, Makueni, River,Turkana, Wajir and West Pokot. Tana River, Turkana, Wajir, West Pokot Mandera, Marsabit, Narok, Nyeri, and parts of Laikipia. Samburu, TaitaTaveta, Tharaka-Nithi, Tana River, Turkana, Wajir, West Pokot, and parts of Meru. Source: Ministry of State for Development of Northern Kenya & other Arid Lands. 10.12 Consideration of capacity may call for direct administration by the center. The counties which will be most in need of the fund are likely to also be lacking the capacity to plan and budget for its use,and to carry-out spending. The Constitution does not dictate whether the fund should be spent directly by the center (although earmarked to specific uses) or transferred to counties as conditional grants. While direct spending would limits the decision-making autonomy of county governments, it can bring about the advantage of better coordination, focus and possibly management, hence, higher effectiveness. Though small, the Equalization Fund can still be effective 10.13 A crucial first step would involve identifying the primary policy challenge that the fund seeks to address. If the Fund’s primary policy challenge is to reduce regional inequality and help the North catch up with the rest of the country, then specific causes and dimensions of this inequality must be clearly delineated. This is not easy, considering that Kenya’s north compares poorly with the national average, against almost every indicator of human development―access to health care, education, energy, water, financial services and justice. Determining which specific indicator to use might be a challenge. Also, while some causes of regional inequality have been due to biased public policies, there are additional complex factors at play, including the pastoralist lifestyle, the dominant production system in the ASALs, which exposes local communities to environmental and economic vulnerabilities, holding them further behind the rest of the country. Addressing such factors through policy is not easy. But one of the most debilitating features of Kenya’s north, its aridity, remains the most tricky to resolve, least of all, through a skimpy fund with a medium-term lifespan. 123 Chapter 10: Making the most of the Equalization Fund Figure 10-2: Counties covered by the Ministry of Northern Kenya. Turkana Mandera Marsabit Wajir West Pokot Samburu Isiolo Baringo Laikipia Tharaka Nithi Nyeri Garissa Embu Narok Kitui Tana River Makueni Lamu Kajiado Kilifi Taita Taveta Kwale KEY ARID COUNTIES SEMI-ARID COUNTIES Kilometers EXPANDED COAST EWS COUNTIES 0 35 70 140 210 280 Source: Ministry of State for Development of Northern Kenya & other Arid Lands. 124 Devolution without disruption – Pathways to a successful new Kenya 10.14 Generally, the smallness of the Equalization Fund means that it can best be used to leverage more spending, by focusing on systemic strengthening and capacity building. This will be more effective than actually trying to construct facilities and infrastructure projects, or to purchase durable equipment. Should a capital grants program be rolled out at the national level, the equalization grant could help the more challenged counties to: (i) provide the matching capital; and, (ii) prepare project applications (technical and financial proposals)―in the case where a project based approach is chosen―and build capacity implement complex infrastructure projects. CHAPTER 10: KEY INSIGHTS / RECOMMENDATIONS. The equalization fund , such as it has been designed will be insufficient to address the vast needs of Kenya’s marginalized communities, and of the counties in which they live. Even increase in the fund beyond the constitutional 0.5 percent minimum, such as contemplated by the BPS, would fall short of needs. Moreover the impact could be nil―or even negative―if it crowds out existing programs (such as those under the Ministry for the Development of Kenya and other Arid Lands, and the Ministry for Special Programs) whose fate under the new dispensation is still unclear. *** The priority would be to clarify the object of the fund and particularly: o the definition of ‘marginalized areas’ : too broad a definition would spread the meager resources of the fund too thin to have an impact; and, o the scope of interventions which should be narrowly focused on alleviating key service bottlenecks It may make sense not to limit the fund to financing infrastructure. Instead cataclytic interventions could consist in: o addressing staff incentives problems in working in remote localtions; and, o capacity constraints in applying for capital funding (under a capital grants program) and for managing capital projects. Considerations of local capacity may call for direct administration by the center at least temporarily, to guarantee maximum efficiency in the use of funds. 125 PART IV Translating the vision with a smooth transition CHAP TE R E L E V E N Promoting intergovernmental coordination 11.1 Devolution complicates the management of government. It potentially diffuses accountability between different levels of government; introduces the possibility of mismatched resources, responsibility and authority; and generates conflict over policies and priorities. Mechanisms for intergovernmental relations are key to resolving these collective action problems. 11.2 In Kenya, intergovernmental relations will be particularly important for coordination of service delivery. The Constitution mandates shared responsibility for some important areas of service delivery; with the national government generally responsible for policy, and county governments in charge of implementation. But the national government has limited fiscal or supervisory levers with which to influence the achievement of national policy priorities at county level. This is why the importance of maximising the scope for conditional grants was emphasised in earlier chapters. In the absence of significant fiscal or supervisory mechanisms to achieve national policy goals, the institutions of cooperative government become all the more important. 11.3 Achieving trust between central and county governments, and between different county governments, will not be easy. Given Kenya’s long-standing and recent history of conflict and the role of devolution in responding to it, there are arguments for giving a higher priority to intergovernmental relations, than might be the case in other newly decentralised systems. The policy process through which the detail of devolution is being decided has been characterised by distrust: between civil society and government; between county-level actors determined to maximise their autonomy and entrenched central government interests; and over the allocation of resources to different parts of the country. There is far more ‘heat’ in these disputes than typically accompanies decentralisation, and they are likely to escalate, as implementation begins in earnest. 11.4 The legal framework for devolution foreshadows an extensive architecture for intergovernmental relations. However, the framework as designed focuses mainly on relations between governors, and around the budget process. Yet the real impacts of intergovernmental relations―positive and negative―will be felt in the sectors. International experience suggests that friction between levels of government can undermine effective service delivery. This is particularly pronounced in the health sector, arguably the largest (in terms of staff), and most important sector devolved to county governments. Health service delivery is highly vertically integrated and meeting international policy goals will be dependent on service delivery controlled by county government. Positive models from other countries suggest how these challenges can be managed, but implementing these approaches will require commitment and resourcing. The sheer number of counties will challenge effective decision-making. Creative approaches will be required to ensure that the bodies can still make decisions effectively. 11.5 The future of provincial administration presents additional challenges. Kenya’s powerful provincial and district administrations provide a direct conduit between the Office of the President and the lowest level of government engagement with citizens. Efforts by some to dismantle this structure in the course 128 Devolution without disruption – Pathways to a successful new Kenya of designing the system of devolution, have so far been resisted successfully. Conflict between district administration and county governments could undermine effectiveness and accountability. However, a substantial rump of national functions is likely to remain, and the arrangements for coordinating them, both between national agencies and with county governments, remain to be decided. Why intergovernmental relations is important in Kenya 11.6 Kenya’s Constitution promotes a cooperative model of government, in which no level of government is superior to the other. Kenya has borrowed from the model of government in South Africa, and from the German model that in turn informed it, the concept of separate but interdependent spheres of government. In order to carry this vision into reality―and for several practical reasons― intergovernmental relations will be more important in Kenya than in many other devolved systems. 11.7 National government powers to interfere in county affairs are circumscribed by the Constitution. The Constitution provides that the sovereign power of the people is exercised at national level and county level.1 Governments at national and county levels are distinct and inter-dependent, and should cooperate with each other.2 Article 189 of the Constitution spells out the requirement for mutual respect, implementation of laws of the other level of government, policy coordination, and the use of dispute settlement procedures (which are provided in the Intergovernmental Relations Act). Conversely, the national government’s power to interfere in how counties exercise their devolved powers and functions is limited: • A law may empower the national government to assume responsibility for functions, if a county fails to perform them.3 However the law that implements this provision has not yet been enacted. • The President may suspend a county government in an emergency arising out of internal conflict, or in other exceptional circumstances.4 E xceptional circumstances are provided for in the County Governments Act, which allows a petition for suspension of a county government to be entertained, if 5,000 registered voters sign it.5 • Although the national government can make laws on any matter,6 national laws prevail over county laws only if they meet certain criteria including: (i) they must apply uniformly across Kenya; (ii)the national law must be aimed at preventing unreasonable action by a county that is prejudicial to national economic, social or security interests of Kenya or another country; (iii) the law covers a subject matter that cannot effectively be regulated by individual counties; (iv) the law deals with a matter that requires uniformity across counties, and establishes norms and standards or national policies; or, (v) a national law is needed for reasons of national security, economic unity, protection of common markets, promotion of economic activities across county boundaries, equal opportunity or protection of the environment. If these conditions are not met, the county law will prevail.7 11.8 The capacity of the national government to influence county policy and spending priorities through conditional grants is likely to be limited. The majority of funding to county governments through the minimum 15 percent equitable share is likely to be (and should be) unconditional, and there are relatively few powers provided to the national government to control the way counties spend funds. 1 Constitution of Kenya, Article 1. 2 Constitution of Kenya, Article 6(2). 3 Constitution of Kenya, Article 190(4). 4 Constitution of Kenya, Article 192. 5 County Governments Act 2012, Section 123. 6 Constitution of Kenya, Article 186(4). 7 Constitution of Kenya, Article 191. 129 Chapter 11: Promoting intergovernmental coordination 11.9 Intergovernmental coordination will be an every-day issue, because the main service delivery responsibilities are split across government. The respective functions and powers of the national and county governments are spelled out in the Fourth Schedule of the Constitution. Virtually the only sectoral area where national government has not been assigned any policy responsibility is the delivery of urban services, like street lighting, traffic and parking, animal control and welfare, fire fighting and the like. Table 11-1 shows the sectors where national and county governments will both play a role. Note that not all functions are assigned under the Fourth Schedule, and unassigned functions are the responsibility of the national government. Table 11-1: Selection of concurrent areas of service delivery responsibility. Sector National responsibilities County responsibilities Health Health policy, national referral facilities. County health services, ambulance All other health functions not mentioned services, primary health care, licensing in Fourth Schedule. of food preparation and sale, veterinary services, cemeteries and funeral parlours, refuse removal and solid waste disposal. Agriculture Agriculture policy. Crop and animal husbandry, livestock sale All other agriculture functions not yards, county abattoirs, plant and animal mentioned in Fourth Schedule. disease control, and fisheries. Energy Energy policy including electricity and gas County planning, including (e) electricity reticulation and energy regulation. and gas reticulation and energy reticulation. Education Education policy, standards, curricula, Pre-primary education, village examinations and the granting of polytechnics, homecraft centres and university charters. childcare facilities. Universities, tertiary educational institutions and other institutions of research and higher learning and primary schools, special education, secondary schools and special education institutions. Trade and commerce Regulation of banking, insurance and Trade development and regulation financial corporations. including markets, trade licenses, Labour standards. fair trading (but not regulation of Tourism policy and development. professions), local tourism, and cooperative societies. Disaster management Disaster management. Disaster management. Water Water protection, ensuring sufficient Storm water management in built up residual water, hydraulic engineering and areas; and water and sanitation services. safety of dams. Source: Adopted from the Kenya Constitution 2010. 11.10 Interaction between the two levels of government is likely to be different from sector to sector, and exactly what each level will do still needs to be worked out. In the energy sector, for example, the role of the county appears to be limited to planning, while the national government may look after actual service delivery. In the education and health sectors, both levels are involved in service delivery, but involving different components of the service network; in the case of heath the majority of institutions are a county responsibility, but in the education sector the reverse is the case. Both levels of government have concurrent responsibility for disaster management. 130 Devolution without disruption – Pathways to a successful new Kenya 11.11 Effective intergovernmental coordination will require both a change of culture, and the building of capacity of both levels of government. It is common for line ministry staff in formerly single tier systems to resist real decentralisation of their functions. Partly this originates from a natural reluctance to lose powers and control of resources―in some cases, because these may offer rent seeking opportunities― but also because they fear that service delivery standards will be compromised, or that service delivery will be subject to political interference. Ministry staff can also have great difficulty reorienting from a service delivery role to a policy role, and in understanding how to use the new tools at their disposal to influence policy outcomes, like standards, expenditure norms and performance monitoring. 11.12 The most significant changes required will be at county level. At the county level, sectors are currently managed in a highly centralised way. Staff are recruited nationally, the majority of budget is managed nationally, and human resource management is centralised. This means that few of the de-concentrated staff located in districts now have the capacity to engage in meaningful, resource- constrained budgeting processes, or independent formulation of policy or legislation. Although the focus initially is likely to be on building the capacity of those at county level, national-level staff should also be retrained to embrace their new roles. In important sectors like health, this is likely to be just as important for the success of devolution as the training of county staff. Structure of intergovernmental relations in Kenya 11.13 Two laws establish the framework for future relations between levels of government. The Intergovernmental Relations Act (passed by Parliament in February 2012) establishes a number of intergovernmental mechanisms. In addition, a National Budget Economic and Budget Council is established under the Public Finance Management Act, which is still before Parliament. The system is described in Box 11-1. The detail of how these forums will work is still to be worked out. 11.14 A particular challenge for Kenya will be how to get this system working with so many counties. Many of the international models for intergovernmental cooperation operate with much smaller numbers of subnational units. Australia has 8 states; South Africa has 9 provinces; Canada has ten provinces and three territories; and Germany has 16 lander for a population of 82 million. Even Brazil with a population of almost 200 million has only 26 states and 1 federal territory―an average of over 7 million persons for each state. In contrast, Kenya’s counties have an average population of less than 1 million. Involving such a large number of counties lessens the possibility of agreement. Intergovernmental bodies provide an important forum for building relationships, but they should also be able to take decisions. It may be useful to consider how this is achieved in Nigeria, which has 38 subnational units, and where intergovernmental relations mechanisms are relatively well developed. In Papua New Guinea, the 19 provinces established themselves into four regional councils, that served to unify regional positions in preparation for participation in annual Premiers’ Council Conferences, at which national and provincial governments reached joint positions on policy issues. 11.15 Coordination of policy formulation is easier to achieve than coordination of implementation. The next section focuses on the importance of expanding the scope of intergovernmental relations into the sectors. In some countries, the main objective of intergovernmental coordination is joint policy development. While this will be very important in Kenya, perhaps even more important is coordination at the level of implementation, in the areas where both levels of government share service delivery responsibilities. The first area of policy coordination that will be required is the unbundling of the 131 Chapter 11: Promoting intergovernmental coordination Box 11-1: Kenya’s intergovernmental relations architecture. National and County Government Coordinating Summit (Intergovernmental Relations Act 2012) • Chaired by the President (or Deputy President if the President is absent). • Vice-chair is the Chair of the Council of County Governors. • 47 Governors. • Forum for consultation and cooperation; promotion of national cohesion and unity; consideration and promotion of national interest; consideration of reports from other intergovernmental forums; evaluating the performance of national or county governments; coordinating and harmonizing development of national and county government policies. • Facilitating and coordinating the transfer of functions from and to either level of government. • Meets twice a year. • Reports annually to National Assembly, Senate and county assemblies. Intergovernmental Relations Technical Committee (Intergovernmental Relations Act 2012) • Chairperson and eight members recruited competitively and appointed by the Summit. • Responsible for day-to-day administration of the affairs of the Summit, implementing the Summit’s decisions, and taking over the functions of the Transition Authority after it is dissolved. • May establish sectoral working groups or committees for the better carrying out of its functions. • Has a secretariat, headed by a Secretary appointed with the approval of the Summit. • Reports quarterly to the Summit, and is accountable to it. Council of County Governors (Intergovernmental Relations Act 2012) • Governors of forty-seven counties. • Forum for: consultation; dispute resolution; sharing information on county performance; considering matters of common interest; formulating inter-county agreements and monitoring them; and facilitating capacity building. • Meets twice a year. • Reports to Summit, Senate and National Assembly. Intergovernmental Budget and Economic Council (Public Finance Management Act 2012) • Chaired by Deputy President. • Members: Cabinet Secretary Finance, Cabinet Secretary Intergovernmental Relations, representatives of Parliamentary Service Commission, Judicial Service Commission, Chairperson of Commission on Revenue Allocation, Chair of Council of County Governors, and forty-seven county Executive Committee members for finance. • Forum for: consultation; dispute resolution; sharing information on county performance; considering matters of common interest; formulating inter-county agreements and monitoring them; facilitating capacity building. • Functions: review various fiscal and budget documents including Budget Policy Statement and Budget Review and Outlook Paper, forum for consultation on economic, budgetary, financial management and integrated development matters, consult on legislation with a financial implication for counties, consider recommendations of the Commission on Revenue Allocation on equitable distribution of revenue and allocation of revenue among county governments, and agree schedule for disbursement on the basis of cash flow projections. • National Treasury provides secretariat. Source: World Bank staff analysis. 132 Devolution without disruption – Pathways to a successful new Kenya constituent elements of the different functions, to ensure clarity and a common understanding about what each level of government will do (see Chapter 5). As part of this exercise, there could also be a joint planning process to work out how day-to-day coordination around implementation should happen. Models for sector coordination 11.16 Good intergovernmental relations will make the most tangible difference in individual sectors. In both Germany and South Africa, cooperative approaches to intergovernmental relations encourage a high degree of integration between levels of government, but have also created problems with priority assignment (Germany)8, and clear specification of roles and responsibilities (South Africa).9 The greater the degree of integration of responsibilities, the greater the need for coordination. 11.17 Sectoral coordination bodies can provide a forum for on-going negotiation of function assignments. It is likely that the initial function assignments in some of the key sectors will be subject to refinement and revision, as the two levels of government begin to work together. Function assignments are dynamic, and it is quite common for various aspects of functions to be recentralised, then decentralised again, in an on-going cycle. The most important objective is that there is clarity, and that the changing function assignments are reflected in changes to resource allocation. A radical overhaul of intergovernmental management arrangements in Australia, led to the negotiation of National Agreements between commonwealth and state ministers in various sectors enshrining agreements on mutual priorities in five key areas of state service delivery. The largest sector of the five is health. Although the initial agreement was only negotiated in 2009, within two years it was already being amended in consequence of a major reform that will see the Commonwealth Government take over funding responsibility for some key aspects of public hospital services. Both the National Agreement and the Commonwealth Federal Financial Relations Act were amended to implement this change in assignment of expenditure responsibilities (see Box 11-2). 11.18 Service standards, expenditure norms and performance frameworks are more likely to be accepted if they are negotiated jointly with county governments. The Australian National Agreements include expected outputs and performance benchmarks, agreed jointly by the states and the Commonwealth. The Counsil for Australian Government (COAG) Reform Council, an independent body comprised of appointees who are not politicians, monitors achievement and publishes data and analysis on their website. This helps encourage a culture of performance among states, even though the funding for service delivery is not tied to specific spending or performance outcomes. A national database is being assembled that includes performance information on all the Agreements (see Box 11-2). Because all the agreements fit under the common umbrella of the Intergovernmental Agreement on Federal Financial Relations, they adopt a common approach, layout and presentation of information. This helps to avoid sector-specific approaches that would add to the overall cost of implementation, if each sector were different. In contrast to the Australian experience, a number of countries have experienced problems in achieving compliance with service standards, because they have been developed unilaterally and are either unrealistic or unaffordable. Indonesian decentralisation laws envisaged minimum service standards, but line ministries were left to develop individual regulations which were in many cases too ambitious to be feasible within the limits of regional government resources. 8 OECD (1997). 9 Steytler and Fessha (2005). 133 Chapter 11: Promoting intergovernmental coordination Box 11-2: Intergovernmental Relations in Australia. The Commonwealth of Australia was formed in 1901 by the federation of six separate British colonies. The Constitution of Australia gives specific enumerated legislative powers to the Commonwealth government, with residual powers― including most service delivery responsibilities―remaining with the States. A decision of the High Court in 1942 gave the Commonwealth control over the most important sources of revenue. Since then, Commonwealth influence over policy has increased steadily, in part through the use of the Commonwealth’s fiscal power to control state priorities through tied grants called Specific Performance Payments (SPPs). SPP grants proliferated into a confusing and only partially effective matrix, often with conflicting objectives, until a comprehensive reform of intergovernmental fiscal relations was implemented, through the Intergovernmental Agreement on Federal Financial Relations that took effect on 1 January 2009. Intergovernmental financing The new intergovernmental financing arrangements streamlined intergovernmental financing by clarifying responsibilities, providing flexibility in the way services are delivered, incentives for improved performance and mutual accountability of both levels of government. The number of SPPs was rationalised drastically. There are now three broad financing instruments: (a) GST revenue grants funded by sharing 100 percent of Goods and Services Tax (GST), distributed in an equalising way between states on the advice of the Commonwealth Grants Commission, plus general purpose financial assistance; (b) Specific purpose payments to fund State service delivery responsibilities (health care, schools, skills and workforce development, disability services and housing). For each of these areas there is a National Agreement specifying service outcomes and monitoring arrangements. Although the agreements specify service outcomes, the payments are not dependant on specific conditions being met, other than that the funds must be spent in the sector. In effect, they are block grants. (c) Partnership payments are provided in respect of specific national reform priorities implemented by States, and fund (i) capital projects, (ii) facilitation of reform, and (iii) reward payments for exceptional performance. Each partnership payment is the subject of a National Partnership Agreement. Roles and responsibilities of intergovernmental bodies The apex body for coordination between levels of government is the Council for Australian Governments (COAG). Under COAG, the Ministerial Council for Federal Financial Relations oversees the implementation of the intergovernmental financing agreement. Sector Ministerial Councils (for example, the Council of Australian Health Ministers) reach agreement on the policy priorities reflected in National Agreements, which specify the responsibilities of the States but do not require spending of any specific amounts. Payments of SPPs and Partnership payments are made on a monthly basis by the Commonwealth Treasury to each State Treasury, which then distributes the funds to respective implementing agencies in the state. This helps to reinforce the idea that the first accountability for service delivery is to State Parliaments and citizens. Commonwealth sector departments (ministries) are primarily responsible for the policy dimensions of national agreements, and for providing advice to Treasury on the amount of payments. The achievement of mutually agreed outcomes and performance benchmarks, is monitored by the COAG Reform Council, an independent body which publishes performance information along with analytical reports, assesses eligibility for performance payments under the Partnership Agreements, and monitors the overall pace of reform. Source: World Bank staff analysis. 134 Devolution without disruption – Pathways to a successful new Kenya 11.19 Conditional grants lubricate intergovernmental coordination. When payment of tied grants forms a part of the intergovernmental arrangements, both levels of government have a much stronger reason to cooperate. Nigeria’s intergovernmental coordination arrangements benefited from the introduction of grants tied to tiered planning processes at national, state and local levels. In Uganda decisions on allocation of conditional grants form a central element of the national Medium Term Expenditure Framework planning process. Challenges of parallel systems 11.20 It is not yet clear how the existing system of provincial and district administration will relate to county governments. A system of provincial and district administration, initially established during the period of British colonial administration, has formed the backbone of central government administration in rural areas, since Independence. In keeping with the British model of de-concentrated administration, District Commissioners do not have direct control over staff of other ministries, but they are unquestionably the most senior public servant in the district, command a section of the police force (the administration police), and are responsible for liquor licensing, registration of persons, maintenance of security, and dispute resolution. Below each district commissioner, further geographic subdivisions (divisions, locations) are under the administration of a hierarchy of district officers, and chiefs who report to the district commissioner. The Constitution requires the system of provincial administration to be restructured, in line with the system of devolution within five years. 11.21 The role of district commissioners in counties is one of the most contentious aspects of the devolution arrangements. The coming into force of the County Governments Act, the main legislative instrument defining the political and administrative structures of county governments, was delayed for almost five months, because the President refused to assent to two provisions that affected the respective role and relationships of county governments and provincial and district administrations.10 The first was a provision under which the County Governor would chair the “county equivalent of the National Security Council�.11 The second was an amendment that had the effect of seconding district commissioners, district officers andchiefsto county governments.12 These proposals would have effectively put the county government in charge of both the staff and security functions of the district administration. The President, to whom the provincial and district administration staff report,13 clearly did not agree with this proposal. The impasse was resolved in mid-2012, when Parliament agreed to further amendments that removed these two provisions, and the County Governments Act was finally passed. 11.22 The relationship between district administrations and county governments will define intergovernmental relations in a day-to-day sense, as much as do relationships within sectors. A key question is how the remaining national functions not devolved to county governments will be organised. Table 11-2 shows the departmental heads represented in a typical district. Many of them relate to functions that are not devolved under the Constitution, although they might be delegated. In some cases, functions that are best performed at local level (like social welfare, children’s and gender programs) have not been devolved by the Constitution, although the 2012/13 budget indicates the district-based staff of these ministries are expected to be devolved. In other cases the functions of existing department heads might be split (like adult education). It will be important to define a new architecture both for day-to-day relationships between national, and agencies in each county, but also for arranging the functions that remain national. 10 Kenya Parliament, Hansard, 29 February 2012 (PM), page 1. 11 Clause 31(1) County Governments Bill. 12 Kenya Parliament, Hansard 23 February 2012 (PM), page 135. 13 The Ministry of State for Provincial Administration and Internal Security is under the Office of the President. 135 Chapter 11: Promoting intergovernmental coordination Table 11-2: Departmental heads represented in a typical district. District Department County National Comment on function assignment assumptions Head 1. District Accountant Administration functions are incidental to functional responsibilities. 2. District Internal Auditor Ditto. 3. District Development Ditto. Planning functions clearly given to county governments Officer under the County Governments Act. 4. District Procurement Ditto. Officer 5. District Information Ditto. Officer 6. District Medical Officer District level health facilities and below devolved. Health 7. District Public Health Primary health and regulation of food sales devolved. Officer 8. District Education Officer Only pre-primary education is devolved. Since there is very few government pre-primary education institutions, this is not in practice a sector that is currently being administered by district education staff. 9. District Adult Education Village polytechnics are a county function, but other higher Officer education is a national function. Function will need to be split. 10. District Roads Officer County roads devolved. If some functions of district roads officer relate to national roads, some residual functions. 11. Town Council Clerk Urban functions devolved. 12. County Council Clerk Urban and related functions devolved. 13. District Housing Officer Housing devolved as part of county planning and development. 14. District Sports Officer Sports and cultural activities and facilities devolved Promotion of sport and sports education is a national function. 15. District Culture Officer Ditto. 16. District Youth Officer Youth programs not mentioned in Fourth Schedule, therefore assumed to be national 17. District Children’s Officer Children’s programs not mentioned in Fourth Schedule, therefore assumed to be national 18. District Gender and Social Gender and social development programs not mentioned in Development Officer Fourth Schedule, therefore assumed to be national 19. District Agriculture Agriculture functions devolved. Officer 20. District Veterinary Officer Livestock functions devolved. 21. Zonal Forests Officer Forestry not mentioned in Fourth Schedule, but protection of environment is a national functions. 22. District Game Warden Ditto. 23. District Geologist Ditto. 24. District Water Officer Protection of water and securing sufficient residual water, hydraulic engineering are national functions. Water supply and sanitation are county functions. 25. District Drought Drought management not mentioned in Fourth Schedule. Management Officer Assume related to national water functions. Source: World Bank staff analysis. 136 Devolution without disruption – Pathways to a successful new Kenya Table 11-2: Departmental heads represented in a typical district (continued). 26. District Environment Environmental protection is a national function. Control of air Officer pollution is a county function. 27. District Land Registrar No land functions mentioned in Fourth Schedule other than surveying. Assume national. 28. District Surveyor Land survey and mapping are a county function. 29. District Physical Planner Planning and development is a county function. Appears to include physical planning. County Governments Act clearly includes. 30. District Land Adjudication District administration undertakes some resolution of land Officer disputes. No mention of land functions except surveying in Fourth Schedule. Assume national function. 31. District Valuer No mention of land functions except surveying in Fourth Schedule. Assume national function. 32. District Cooperatives Trade development including cooperative societies devolved. Officer 33. District Enterprise Ditto. Development Officer 34. District Trade Ditto. Development Officer 35. District Labour Officer Labour standards is a national function. 36. District Civil Registrar Not mentioned in Fourth Schedule. Assume national function. 37. District Registrar of Immigration and citizenship is a national function. Persons 38. OCPD ?? 39. District Criminal Police Services are national function. Investigation Officer 40. OIC Prisons Correctional services are a national function. 41. District Probation Officer Ditto. 42. Sub-regional Intelligence National Security presumed to be a national function. Coordinator 43. District Magistrate in Courts are a national function. Charge 44. District Kahdi Ditto. Source: World Bank staff analysis. 11.23 Until now, Kenya has followed the British colonial tradition of semi-prefectural administration in districts, but this is not the only model available. Central government is represented at subnational levels in all countries, but representation takes different forms. The Napoleonic prefectural model14 used in many European countries (France, Belgium, Netherlands, Spain and Italy) puts all central government public servants at the subnational level under the control of a single appointed prefect. A number of those countries replicated the model in the countries they colonised. In Indonesia, a former Dutch colony, the President appointed the heads of provincial governments until the decentralisation reforms introduced in the late 1990s. Germany, Switzerland and Austria do not come from the Napoleonic tradition, but have adopted a similar model, although it is not applied in all subnational jurisdictions within each country. Developed countries that follow the British model do not have prefects at the local level. Instead, a network of independent agencies reports directly to their parent agencies at the centre. The UK, Canada, Ireland, New Zealand and Australia all follow this model. Other countries, 14 The Napoleonic prefect model is a legacy of the centralist Jacobite movement born out of the 1789 French Revolution, and was the first model to influence the unitary states of continental Europe. 137 Chapter 11: Promoting intergovernmental coordination including the US and Mexico, that draw from neither tradition, also do not have any single point of control over central government officials at the subnational level. Kenya’s post-independence system of administration continued the British hybrid model typical of British colonies, under which a district commissioner with substantial powers over law and order (and in some cases limited other functions), is responsible for coordination of other agencies at district level, but has no direct control over them. Structural changes in government over the last twenty years have often affected central governments’ roles. There is now an increased emphasis on partnership, collaboration, and management approaches, focused on clear performance expectations, coupled with greater managerial authority, rather than the hierarchical supervisory approach that was prominent in the past. 11.24 Given the political resilience of provincial administration, Kenya will need to manage the risk that parallel and competing structures could undermine the effectiveness of county government. Administrative parallelism, under which staff of central government either continues to manage subnational functions, or jointly administer them alongside national staff, would weaken the accountability of county governments. Despite establishment of elected governments at commune and provincial level in Madagascar, districts continued to be the focus of de-concentrated service delivery by central government, in particular serving as the main link between service delivery facilities and central government (see Box 16-2). 11.25 Failure to give county governments sufficient autonomy in relation to staff carrying-out their functions would also undermine county accountability. To be accountable, county governments need the capacity to delegate the tasks for which they are politically responsible to agents (civil servants or contracted providers), and the finance to carry them out; citizens need to be able to measure the performance of the county government and its agents using information, and they ultimately enforce their control over outcomes through the electoral process. If county governments do not have control over public servants carrying out subnational tasks, they cannot be held to account for service delivery under-performance. Conclusions 11.26 Intergovernmental relations will be more important in Kenya than in most systems, so this aspect of the new system should receive a corresponding level of attention. Given the extent of change involved in setting up new county governments, attention is likely to be mainly focused on county-level tasks. However, in the long run the success of devolution will depend on the mechanisms of intergovernmental relations, as much as on what happens at county level. Although it will require diverting some energy away from these urgent questions, in the medium to long term Kenya’s devolution will benefit from devoting some time and resources to establishing these mechanisms from the beginning. 11.27 Successful intergovernmental relations mechanisms will have most impact in sectors. The current architecture is concentrated on apex political bodies and the budget process. But devolution will succeed or fail in the way it works within sectors. This is particularly important in Kenya, because so many sectors involve joint service delivery. Poor coordination is likely to lead to inefficient duplication, or service gaps. 11.28 Sorting out the role of provincial and district administration in the counties should be a first order priority. Conflict between the functionaries of national and county governments is highly unproductive, and will undermine effective outcomes of decentralisation. 138 Devolution without disruption – Pathways to a successful new Kenya CHAPTER 11: KEY INSIGHTS / RECOMMENDATIONS. Intergovernmental relations mechanisms are particularly important in Kenya, because the Constitution limits the national governments power in relation to county governments. Many service delivery functions are shared between levels of government, so there is added reason to focus on coordination mechanisms within sectors. Effective coordination will involve cultural change at national level, and capacity building at county level. Unless county governments feel like equal partners with national government, intergovernmental relations is doomed to failure. A substantive framework is provided for in the laws, but still needs to be operationalized. Getting intergovernmental bodies working effectively for forty-seven county members will be particularly challenging. *** Despite the many competing priorities, it is important to focus on getting intergovernmental relations bodies functioning early on. The first priority is to get Sector bodies working on issues of function assignment, service delivery standards, and performance monitoring. The second priority is to focus on intergovernmental relationships at the county level, particularly by resolving the role of provincial and district administration staff, and designing the new arrangements for remaining national staff at county level. 139 CHAP TE R T WE LV E Managing money for county development 12.1 Kenya’s Constitution fundamentally changes the way public money will be managed. Like other elements of the 2010 Constitution, the changes to Public Financial Management reflect a distrust of the centralised Executive and its opaque control over resources. The Constitution enshrines dramatic changes to the relationship between national Parliament and Executive, with profound implications for revenue-sharing and macro-economic stability. It also guarantees Kenya’s county governments a degree of independence from the national Executive in managing their finances. While this maximises opportunities for real local control over resource management, it opens up the risk that, if counties manage their money poorly, it could undermine the service delivery and democracy objectives of the Constitution. 12.2 The new PFM Act has attempted to deal with these risks. The Act lays out a sequence of steps to operationalize the annual revenue sharing arrangements, and maintain fiscal discipline by Parliament. Given the politics involved, the scope to do this within an ordinary law remains to be tested. Currently, the full PFM framework is not yet clear, since the Regulations have yet to be finalised. But the systems it establishes are complex, and counties in particular start from a very low base. Systems for sound PFM need to be put into place to ensure county funds are allocated in line with priorities, and spent in a way that supports service delivery and accountability. 12.3 Transition represents a particular risk―both for allocating funds to counties and county-level PFM systems. Implementing the new Act without integrity would send a powerful, but unfortunate signal. Neither the Constitution nor the PFM Act have addressed the complexities of how county finances should be managed during the transition period, especially given that they come into operation three months before the end of a fiscal year. To ensure respect for the law, these transition arrangements should be enshrined in law. The window of opportunity to achieve this is fast closing. Kenya’s new budget process 12.4 The Constitution completes Kenya’s transition from a Westminster to a presidential model of government. This fundamentally changes the balance of power between the Executive and the Legislature, with the latter gaining significantly enhanced powers over the budget process―especially in the vertical division of revenue between national and county governments. There will be full separation between Parliament and Executive, so that cabinet members will no longer control the budget process within Parliament. Parliament is given correspondingly greater powers to control the process, putting Kenya’s Parliament in an only slightly less powerful position than that of the United States Congress (see Figure 12-1). These constitutional changes have far reaching implications for the political economy of Kenya’s PFM process, and the incentives faced by different actors within and around it. 140 Devolution without disruption – Pathways to a successful new Kenya 12.5 The budget process will be in some Figure 12-1: Increase in parliamentary independence and power. senses turned on its head, by beginning with the spatial division 6 of revenue among counties. In United States Parliament has more power to control the past, budgeting in Kenya began 4 with allocation among sectors, South Africa France Kenya New System the executive proceeded to allocation between 2 Kenya Old Spain System ministries within each sector, and United Kingdom then among programs. Distribution 0 0 2 4 6 8 10 among different districts occurred as part of within-Ministry allocation Greater separation of powers between Parliament and Executive processes, with the result that the spatial distribution of resources is virtually invisible.1 The new budget Source: World Bank staff analysis. process will begin with the division of national revenue vertically between national and county governments, and horizontally among the forty-seven counties. The annual budget process as envisaged in the PFM Act is summarised in Figure 12-1. The budget process also involves a significant number of actors at both national and county level, including a new Intergovernmental Budget and Economic Council (IBEC), comprising the county secretaries for finance and the national Cabinet Secretary for Finance (CSF), which is intended to provide a forum for intergovernmental consultation on the budget (see Table 12-1). Table 12-1: Principal PFM actors at national and county level. National Actors County Actors Actors with equivalent at national/county level Parliament: National Assembly, Senate, and Parliamentary County Assemblies Budget Office (PBO) Cabinet County Executive Committee National Treasury County Treasuries Cabinet Secretary for Finance (CSF) County Executive Member for finance (CEMF) Accounting Officers for national government Accounting Officers for county governments Receivers and Collectors of Revenue for national Receivers and Collectors of Revenue for county government governments Actors with no equivalent at national/county level Judicial Services Commission Boards of Cities and Municipalities Parliamentary Services Commission County Budget and Economic Forum (CBEF) Controller of Budget Auditor General Commission on Revenue Allocation (CRA) Intergovernmental Budget and Economic Council (I-BEC) 1.1 Public Debt Management Office (PDMO) 1.2 Public Sector Accounting Standards Board (PSASB) Source: Adapted from Treasury (2012). 1 As is discussed below, the 2013/14 budget process will follow an accelerated timetable. 141 Chapter 12: Managing money for county development The process of revenue sharing 12.6 There are a large number of steps involved in the division of revenue. The budget process begins at national level not later than 30th August, when the CSF issues a circular to all national government entities setting out guidelines on the budget process to be followed by them, and notifies IBEC that the budget process has commenced. Next, six months before the start of the beginning of the financial year (i.e. not later than 31st December) the Commission on Revenue Allocation’s recommendations on the vertical division and horizontal allocation of revenue are submitted to Parliament, the National Executive, County Assemblies and County Executives. The IBEC also submits recommendations to the National Treasury. The National Treasury then prepares the draft Division of Revenue Bill (DoRB), which sets out the division between the levels of government in terms of the county equitable share, and the County Allocation of Revenue Bill (CARB) which divides the county amount between counties, according to a resolution made by the Senate. The revenue sharing bills must be accompanied by explanatory memoranda (in line with Art. 218(2) of the Constitution) and should take into account the recommendations from the CRA and the IBEC. 12.7 Critically, the two draft Bills must be prepared together with the Budget Policy Statement (BPS), which provides the overall ‘fiscal framework’ (i.e. total estimated national revenues and expenditures, including transfers to counties). The CRA and the IBEC are also given an opportunity to comment on the BPS before it is finalised, and the CRA must have an opportunity to comment on the draft bills under Article 205(1) of the Constitution. The Cabinet Secretary for Finance must submit the BPS, DoRB and CARB to the National Assembly jointly by February 15th each year. 12.8 The revenue sharing bills must be consistent with the fiscal framework in the BPS. The DoRB in particular has a fundamental impact on the national budget, because it determines how much is available for national spending. If the amount allocated to national level in the DoRB differs from the total of the national budget estimates, the entire and national budget process is undermined. This will also apply to the forty-seven county budget processes too. 12.9 It is also crucial that the CRA and the IBEC should make recommendations about revenue sharing that encourage fiscal responsibility.2 If these bodies make unrealistic assumptions about the fiscal space available for devolved spending, they will deprive Parliament of access to genuinely alternative advice on the fiscal framework. The National Treasury will find it easier to justify deviations from these bodies’ advice if it is unrealistic, and it will be right to do so. The risks associated with divergent approval processes for the revenue bills and the BPS are discussed further below. (see also Figure 12-1 for the timing of the different processes). Impact of Parliament’s increased power over budgeting 12.10 Parliament’s increased powers of legislative budgeting will reshape the budget process. In particular, if different parties control the Executive and Legislature, a great deal of negotiation may be required to get the budget passed. Parliament also controls access to what is sometimes called ‘supply’―an authorisation to spend money if the budget is not passed. Under Article 222(1), the National Assembly must still authorise the withdrawal of money from the Consolidated Fund, if the budget has not been passed. Under the PFM Act, county assemblies have a similar power over county governments. 2 The CRA is required to encourage fiscal responsibility in making recommendations, under Article 205(3)(c) of the Constitution. 142 Devolution without disruption – Pathways to a successful new Kenya Similar provisions in the United States have meant that the US President has on occasion been forced to shutdown the federal government for short periods, because of deadlock between Congress and President. A budgetary gridlock between the Executive and newly empowered Legislature occurred in Nigeria during the first full budget process under the new Constitution: a disagreement between the President and the National Assembly over the role of the legislature caused a four-month delay in the passing of the federal budget for FY 2000. As Wehner notes, ‘The conflict revealed that the Executive and legislative branches of government were actively testing their respective powers and roles in budgetary matters under the new Constitution.’3 The Constitution raises the possibility of a similar impasse in Kenya, as the new institutions sort out their respective roles. 12.11 There are some limitations on Parliament’s power to pass laws that authorise spending or taxing. Any member of parliament can introduce a ‘money bill’ (Art. 114) that contains spending programmes or taxes outside of the appropriations act and main revenue bills. However, the power of the Assembly to pass, or amend a money bill is limited by the constitutional requirement for the Assembly to act only on the recommendation of the relevant committee of the Assembly (taking into account the views of the Cabinet Secretary for Finance). The PFM Act seeks to expand the reach of these provisions by including a requirement that a bill originating from a member of the National Assembly proposing amendments after the passage of the budget estimates and Appropriations bill can only be passed if any proposed increase in expenditure proposed in the bill is balanced by a corresponding reduction in another item of expenditure.4 These provisions received an early test in October 2012, when Parliament passed amendments to the Finance Bill 2012 (the bill which implements the government’s planned revenue measures) that created a spending obligation of around KES 2.5 billion in gratuity payments to parliamentarians that had not been included in the budget.5 Although the bill did not receive assent by the President, the fact that the amendments passed at all suggests Parliament’s power is not yet effectively fettered by these provisions. 12.12 In any case, the Division of Revenue Bill is not defined as a money bill, so the procedures set out above do not apply. This means that the DoRB could be amended by Parliament, without the recommendation of the House Committee or the input of the CSF. Since it is free from the Constitutional restrictions placed on money bills, the DoRB is likely to be subject to significant discussion, and potentially, alteration. This raises the possibility that delays in approval of or amendments to, the DoRB could derail the budget process at both national and county level, and contribute either to gridlock between the Executive and Parliament late in the budget process, or to the passage of budgetary provisions that are fiscally unsustainable. Two-step budgeting 12.13 A number of countries have legislated for a two tier budget process to manage the risk of fiscally unsustainable budgeting by Parliament. In Kenya, this would involve a process where the legislatures at national and county level bind themselves to a fiscal framework for the duration of a fiscal year, prior to the consideration of the DoRB, CARB and budget estimates. The process of building consensus around the division of revenue is particularly important, and should begin before the fiscal framework is agreed. Although Parliament could still renege on this agreement, it would be publicly contradicting itself, and therefore it would be harder to unravel the budget at a late stage. Sweden and South Africa are two countries that have struggled with the problem of how to reconcile legislative budgeting with 3 Wehner (2001). 4 PFM Act section 39(4)(c). 5 See for example the Business Daily, Friday October 5th 2012: ‘MPs amend Bill for a KES 2 bn send-off package.’ 143 Chapter 12: Managing money for county development Figure 12-2: National budget process under the PFM Act. Constitutional Commissions National Government Parliament (National Assembly and Intergovernmental bodies and Senate) CRA submits recommendations for division & CSF issues national government budget allocation of revenue to Parliament, County circular setting out budget guidelines and Assemblies, National Executive & County process (30th August), and noti�es IBEC of Executives (6 months before beginning of FY) commencement of budget process. KEY CSF – Cabinet Secretary for Finance IBEC – Intergovernmental Budget and Economic Council National Treasury, taking into consideration CRA – Commission on Revenue Allocation IBEC submits recommendations onDoRB and CRA & IBEC recommendations, prepares DoRB CARB to National Treasury BPS – Budget Policy Statement & CARB DoRB – Division of Revenue Bill CARB – County Allocation of Revenue Bill National Treasury, prepares Budget Policy CRA and Controller of Budget provide views on Statement (BPS) taking into account views of BPS to National Treasury CRA, county governments, & Controller of Budget IBEC and CRA receive noti�cation ofDoRB and CARB from National Treasury prior to National Treasury noti�es IBEC and CRA submission to Parliament before submission of DoRB and CARB CRA provides views onDoRB and CARB to CSF submits BPS, DoRB, CARB & explanatory Parliament memoranda to Parliament (15th February) Parliament to table & discuss a report on BPS containing its recommendations & pass a resolution to adopt with or without amendments (1 March - 14 days after submission of BPS to Parliament) National Assembly budget committee to review DORB & ensure it reflects principle that �nancing follows functions for both levels of CSF submits budget estimates, supporting government. documents, & draft Bills to Cabinet (allowing time to meet other deadlines). Senate budget committee to review CARB & DoRB (30th April - 2 months before end of �nancial year). CSF submits Budget estimates & Appropriation Bill to parliament (30th April) Parliament considers DORB & CARB with a view to approving them, with or without amendments (1 month after bills are Parliamentary Services Commission introduced) accounting officer to submit to National Assembly budget estimates for Parliament Relevant National Assembly Committee to (30th April) discuss & review estimates & make recommendations to National Assembly, taking into account CSF ’s views. Chief Registrar of the Judiciary to submit to National Assembly budget estimates for Judiciary (30th April) CSF submit to National Assembly any National National Assembly to consider national Treasury comments on budgets proposed by budget estimates, including Parliament & Parliamentary Service Commission & Chief Judiciary, with a view to approving them in Registrar for Judiciary (15th May) time for the Appropriation Bill & relevant Bills to be assented to by 30th June. National Treasury to consolidate, publish & publicise budget estimates (21 days after National Assembly has approved them) N.B. National Assembly may amend budget estimates only in accordance with DoRA & resolutions on BPS, and only if any increase in expenditure is balanced by a reduction in expenditure or a reduction in expenditure is used to reduce the de�cit. Source: Adapted from Public Finance Management Act (2012) . 144 Devolution without disruption – Pathways to a successful new Kenya macroeconomic and fiscal stability, and introduced two-step institutional mechanisms in Parliament―with a clear vote on the overall fiscal framework preceding debate on individual appropriations―in order to address this challenge. In both cases, the reforms were implemented after many years of intensely political struggles over the budgetary amendment process in Parliament (see Box 12-1). Box 12-1: International solutions to the challenge of strong parliamentary powers to amend the budget: Sweden and South Africa. Literature on the role of legislatures in the budget process suggests that legislative budgeting tends to result in increases in legislative spending. Sweden and South Africa provide two examples of legislatures with relatively strong roles in the budget process, where institutional mechanisms have been found to address this challenge: • In Sweden, until the 1990s, there was a highly fragmented budget process with multiple appropriations and limited coordination among multiple spending committees resulting in a situation whereby aggregate spending and the deficit were unpredictable, right until the end of the budget process. Among the reforms introduced in order to address this, Sweden introduced a top-down budget process whereby aggregate limits were set prior to allocation decisions, with parliament voting first on the budget totals (in the Spring Fiscal Policy Bill, five months before the presentation of the draft budget) and second on individual appropriations (with presentation of draft budget estimates more than three months before the beginning of the fiscal year). Thistwo-step sequenced process helped, along with other factors, to limit and make more predictable aggregate expenditure and the deficit, and reduce ‘pork barrel’ additions to the budget. • In South Africa, public expectations were sky-high following the African National Congress’ (ANC) election victory in 1994, with huge demand for improved service delivery, but also a conflicting need to re-establish fiscal credibility. After many years of struggle between the Executive and Legislature to respond to the constitutional requirement that ‘An act of parliament must provide for a procedure to amend money Bills before parliament,’ the Money Bill Amendment Procedure and Related Matters Act became law in 2009. Under the new law, approval of the fiscal framework―which comprises aggregate spending and revenues (budgetary and extra-budgetary), borrowing, interest and debt servicing charges and contingency reserve―and the Division of Revenue Bill must take place before consideration of any amendments to the budget. It also separates macroeconomic, fiscal policy and revenue matters (responsibility of the Finance Committee), from responsibility for expenditure (assigned to appropriations committees). This separation reduces the risk of Parliament making changes to the fiscal framework in order to accommodate its own spending intentions. Source: Wehner (2010). 12.14 The PFM Act seeks to take into account lessons from other countries, by introducing a two-step budget process. The two steps, with Parliamentary votes at each stage, are as follows: • Step 1: A ‘Fiscal framework’ in the Budget Policy Statement (BPS) is submitted to Parliament by (15th February) together with the DoRB and CARB, and Parliament is required to“not later than fourteen days after the Budget Policy Statement is submitted to Parliament, table and discuss a report [on the Budget Policy Statement] containing its recommendations and pass a resolution to adopt with or without amendments�.6 • Step 2: Budget estimates are then prepared consistently with the BPS, and are submitted to Parliament by 30th April, with Parliament being required to “amend the budget estimates of the national government only in accordance with the Division of Revenue Act and the resolutions adopted with regard to the Budget Policy Statement, ensuring that: (i) an increase in expenditure in a proposed appropriation is balanced by a reduction in expenditure in another proposed appropriation; or, (ii) a proposed reduction in expenditure is used to reduce the deficit.�7 6 PFM Act section 25(7). 7 PFM Act section 39(3). 145 Chapter 12: Managing money for county development Risk of the BPS and DoRB diverging 12.15 The potential for divergence between the BPS and the DoRB arises from the fact that they are finalised according to different deadlines. While the submission of the BPS, DoRB and CARB to Parliament is done simultaneously (by the 15th February each year), their approval deadlines have not been similarly linked. All three are referred to Parliamentary Committees after submission to Parliament: the BPS and the DoRB to the committee of the National Assembly, and the CARB to the Senate Committee. However, whereas Parliament must receive the committees report on the BPS and pass a resolution by 1st March, the Senate budget committee is not required to review the CARB and the DORB until two months later, on 30th April.8 There is no corresponding deadline for the National Assembly committee to report on the DoRB. In a seemingly contradictory section, the deadline for passage of the two bills, with or without amendments, is 15th March.9 The detailed estimates for the national budget are received in Parliament on the same day as the Senate committee must review the CARB and DORB, assuming that section 8 takes precedence over section 42, which seems likely since the executive will no longer be able to directly introduce Bills into Parliament under the new dispensation. Parliament can amend the budget estimates only if the change is in accordance with the resolutions on the BPS, the effect is cost neutral (any increase is matched by a corresponding decrease elsewhere), and provided the amendment accords with the Division of Revenue Act.10 This does not allow for any difference between the DoRB and the BPS, yet there is no explicit provision requiring the DoRB to conform with the BPS, and if the senate committee’s review of the DORB differs from the BPS, there is a real risk that the DORA may diverge from the BPS. 12.16 The only way to avoid such a potential conflict would be to require that the Division of Revenue and County Allocation of Revenue Acts do not deviate from Parliament’s original resolution passed on the BPS. The PFM Act does not contain such a provision, and even if it did, it is not clear how credible such a requirement would be. The PFM Act is only an ordinary law (as opposed to having a higher legal status, as some ‘organic’ budget laws do in other countries). As such, it cannot restrict Parliament’s power to pass subsequent laws, since Parliament cannot bind itself. Respect and support for this two- step process by the actors in the budget process―especially those within parliament―will therefore be very important (see Murray and Wehner (2012) for a more detailed discussion). There cent actions of MPs in the current Parliament suggests that there will be a great deal of work to be done in building a culture of fiscal responsibility within the new Parliament, to ensure that its new powers are not abused. A medium term context may help manage expectations 12.17 Emphasising a medium-term context for revenue sharing may ease some of the pressures on Parliament to pass a DoRB that is inconsistent with the BPS. Firm floors on the county allocations should be communicated to counties early in budget process, as counties need to know their expected transfer in order to have enough time to complete their own budget processes. Since the county budget circular is required to be issued by 30th August, it is difficult to see how counties could provide firm ceilings to their spending units―since they will not know their proposed allocations until the CARB is presented in February, at the earliest. In South Africa, revenue shares are always presented in a medium term context, with the current year proposal and two out-years. When provinces begin their 8 PFM Act section 8. 9 PFM Act section 42. 10 PFM Act section 39(3). 146 Devolution without disruption – Pathways to a successful new Kenya budget process for the following year, they are authorized to assume that the indicative provincial year (the first out-year) is the minimum flow they will receive. They can budget with comfort knowing that any last minute changes to their budgets will only involve additional resources, not less. Time will tell whether the two-step budget process provided for in the PFM Act will be adequate to manage the new pressures of legislative budgeting in Kenya. It is notable that the statutory period between submission of the fiscal framework (comprising the BPS, DoRB and CARB) and the budget estimates (two and a half months) is shorter than the law requires in both South Africa (three months) and Sweden (five months). Parliament has only 14 days (from 15th February to 1st March) to consider and adopt the fiscal framework in the BPS. Depending on whether the President commands a majority in the Parliament or not, this ambitious timeline may not be adequate to manage the disparate perspectives that need to be accommodated, in order to ensure that the budget can pass through Parliament. Relationship between national and county governments 12.18 The most contentious issue in the debate over the PFM bill was the question of how much control national Treasury should have over the finances of county governments. On one hand, a huge capacity building effort is required to ensure that counties are empowered to set up systems―in many cases starting from scratch―that are in conformity with the single national PFM system envisaged under the PFM Act. Guidelines and templates need to be developed, computer equipment purchased, software installed and training provided. Logic suggests that national government is best placed to coordinate this. Moreover, under the proposed CRA formula for horizontal sharing among counties, the greatest increases in financial flows will be channelled to those with the weakest capacity to manage funds. There are also strong arguments in support of a degree of intergovernmental standardisation and oversight. For example, under what circumstances should the national government be able to step in if there are “serious material breaches� of the PFM Act by county governments, which prevent the delivery of basic services to citizens? And does this oversight role require more detailed monitoring of PFM systems to flag problems early and help cultivate less confrontational intergovernmental relations on PFM issues? 12.19 On the other hand, many stakeholders have good reason to be suspicious of too much national control. There is a widely expressed fear, based on historical experiences, that strong national government oversight of county public finances will inevitably be subject to abuse, especially where there is a split party control of national and county government. Ensuring adequate safeguards to prevent such abuse ―using innovative institutions such as the Controller of Budget and intergovernmental coordination bodies―will be a crucial determinant of the success of the new system. The PFM Act passed by Parliament in July 2012 represented the outcome of a long and detailed process of public consultation and negotiation over these big questions, and how they are manifest in the detailed provisions of the law. 12.20 Prior to the adoption of the Constitution in August 2010, the PFM legal framework in Kenya was highly fragmented. The earlier PFM laws repealed by the PFM Act included the Fiscal Management Act (2009), the Government Financial Management Act (2009), the Internal Loans Act, as well as PFM legislation which was passed in 2011 in order to meet Constitutional deadlines, namely, the Contingencies Fund and County Emergency Funds Act, and the National government Loans Guarantee Act. Combined, 147 Chapter 12: Managing money for county development these acts determine the PFM legal framework for de-concentrated government units in the Districts (health, education, and agriculture, etc.). Meanwhile, local authorities have been governed separately by the Local Government Act (Cap. 265) 1998, and the Local Authorities Transfer Fund Act (No. 8 of 1998). The urban boards that replace them will be brought under the new PFM Act once they are established, after county governments are formed. Partly as a result of this legal fragmentation, there is presently no single PFM system for all tiers of government, with local authorities operating under a different chart of accounts and financial management information system to central government. 12.21 While devolving significant powers to the county governments, the Constitution also provided the basis for a more coherent PFM legal framework. The Constitution requires that “County governments shall operate financial management systems that comply with any requirements prescribed by national legislation�.11 However, it leaves significant room for interpretation regarding how this should be translated into a PFM law or laws, and is not clear on the extent to which national government has a role in “overseeing� county PFM systems. A number of Articles of the Constitution on PFM should be highlighted in order to understand the PFM framework envisaged (see Box 12-2). Box 12-2: What the Constitution says about oversight of county Public Financial Management. The national Parliament is responsible for prescribing the financial management systems that operate at county level under Article 190(2). The power of national Parliament to regulate county Public Financial Management through legislation covers: • The structure of plans and budgets (Art. 220(2)). • Disbursement of funds out of county revenue funds (Art. 207). • Expenditure control and transparency (Art. 225(2)). • Accounting (Art. 226(1)) and procurement (Art. 227). • Auditing (Art. 226 and 229). But it is clear that the lines of accountability from public officials at the county level are to the county assembly (Art. 226(2)) and the Auditor-General reports to the county assembly in relation to county finances (Art. 229(7) & (8)). CRA is given a role to make recommendations on financial management arrangements, and financing more generally, for counties (Art. 205 and 216). The power of the national government to regulate PFM within county governments is backed up by the power to intervene (within a framework provided by a law) if a county operates a PFM system that does not meet national standards (Art. 190). It is particularly important to note that: • The Cabinet Secretary for Finance is empowered to stop transfer of funds to a county, incase of a “serious material breach or persistent material breaches of the measures established under that [national] legislation� (Art. 225(3)a). • However, there are some procedural safeguards around how the national government intervenes in county affairs, in relation to both these powers of intervention (Art. 190(5)). The Controller of Budget also plays an important role of overseeing county budget execution, overseeing the budgets of the national and county governments by authorising withdrawals from public funds under Articles 204, 206 and 207 (Art. 228(4). The Budget Controller is accountable only to the National Parliament (Art. 228(6)). Source: Constitution of Kenya. 11 Article 190(2). 148 Devolution without disruption – Pathways to a successful new Kenya 12.22 The Public Finance Management Act constitutes the national law on PFM envisaged in Chapter 12 of the Constitution. As well as replacing or consolidating a number of existing laws, the PFM Act provides a legal framework for PFM in the new county governments and urban areas and cities, that will be created under the Urban Areas and Cities Act. For the first time in Kenya there will be a single PFM legal framework for all levels of government. 12.23 Ambiguity over the role of the national Executive in overseeing county governments was a fundamental source of debate during the drafting of the PFM Bill. The Act makes limited provision for national government intervention in the cases prescribed by the Constitution, limited to circumstances in which there are serious or persistent material breaches of the PFM law by a county (see Box 12-2). For example, it may be advisable for the national government to intervene in county PFM if mismanagement threatens national macroeconomic stability, or indeed, to provide constructive support from an early stage to help to build strong county PFM systems and skills among those who operate them. At the same time, there is distrust of national government among many who fear that without strict legal protections, the centre’s power to intervene (e.g. by suspending intergovernmental transfers) will be subject to political abuse. Frameworks for support and intervention need to recognise these concerns, by being transparent and clearly communicated. 12.24 The PFM Act provisions on county PFM were subject to intense debate both within government and in the media at different times during the drafting process. Originally there were two draft PFM Bills with provisions for county government financial management: one prepared by the Task Force on Devolved Government12 (TFDG)―the County Governments Financial Management (CGFM) Bill, 2011―and the other prepared by the Treasury’s Working Committee on Fiscal Decentralisation―the Public Financial Management (PFM) Bill, 2011. The TFDG had also proposed a third draft Bill, the Intergovernmental Fiscal Relations Bill, which contained clauses relating to county PFM. The two rival PFM bills were eventually merged into a single “harmonized� PFM Bill (predominantly based on the Treasury draft) through a long process of negotiation, while the IGFR Bill was largely dropped. Box 12-3 reviews some of the major differences between the two initial draft Bills that were resolved in producing a single harmonized PFM Bill, and ultimately the approved PFM Act. Some of the differences are illustrative of the tensions regarding the appropriate role of the national government in oversight of county PFM.13 12.25 The harmonization process was important not so much because of the reduction in the number of PFM Bills―the approved PFM Act is, at over 170 pages, a cumbersome piece of legislation―but because it provided a much more coherent legal framework for PFM. As a result of this process, the major institutions involved in drafting the PFM legislation―Treasury, Ministry of Local Government, Auditor General, Controller of Budget, CRA, CIC, and TFDG―negotiated compromises on some major points of difference for county PFM. The harmoinzed PFM Bill was passed by Parliament on 27th June 2012, and assented to by the President on July 23. 12 The Task Force was established in October 2010 to make recommendations on the broad range of legislation dealing with devolved government as required by Chapter 11 of the Constitution. It reported in July 2011 and presented six bills, including the County Government Financial Management Bill and the Intergovernmental Fiscal Relations Bill, both modeled on similar South African legislation. 13 Lakin (2012). 149 Chapter 12: Managing money for county development Box 12-3: Major points of difference resolved through the harmonised PFM Act. Comparing the draft CGFM Bill dated 1st August 2011 (drafted by the TFDG), the draft PFM Bill dated 18th August 2011 (drafted by the Treasury Task Force ), and the approved PFM Act, some examples of the major differences at the time and their resolution include: • The CGFM Bill included a detailed section on the role, powers and accountability of the Governor in county financial management, while the PFM Bill only referenced the Governor in relation to their role as Chairperson of the County Budget and Economic Forum. There are relatively few explicit references (five) to the role of the Governor in county PFM in the approved PFM Act. • The CGFM Bill provided for county revenue collection (property tax, entertainment tax etc.), while the PFM Bill did not. The harmonised PFM Bill does not include provisions for collection of specific county revenues, which as the County Governments Act makes clear, must be drafted by “the body responsible for matters relating to transition� (i.e. the Transition Authority). • The CGFM Bill allowed counties to appoint the Kenya Revenue Authority (KRA) “as the collector of all tax revenues in respect of property and other taxes�. While the PFM Bill did recognise the KRA as a collector of County Government revenue, it was restricted to “collecting revenue that it is authorised to collect under the Kenya Revenue Authority Act�. Under the PFM Act, the County Executive member for finance may authorize the KRA or appoint a collection agent to be a collector of county government revenue, essentially the proposal made in the CGFM Bill. • The CGFM Bill provided for the review of county borrowing by an Intergovernmental Loans and Grants Council as provided for in the Intergovernmental Fiscal Relations Bill (this Bill, as mentioned, was largely dropped). The PFM Bill provided for an Intergovernmental Budget and Economic Council (IBEC), which was ultimately maintained in the approved PFM Act. The IBEC membership includes the Deputy President (chair), Cabinet Secretary for Finance, Parliamentary Service Commission representative, Judicial Service Commission representative, CRA Chairperson, Council of County Governors Chairperson, every County Executive member for finance, and the Cabinet Secretary responsible for intergovernmental relations. This very extensive membership may make it difficult to operate the IBEC in practice. • The CGFM Bill made no provision for single treasury accounts at county level, while the PFM Bill did require county governments to establish a single treasury account at that bank through which payments of money to and by the various County Governments should be made. This provision was included in the final PFM Act, although a clearer set of definitions was included to clearly differentiate the national single treasury account from the forty-seven county single treasury accounts, as well as the clarification that they “shall not be operated in a manner that prejudices any entity to which funds have been disbursed�. These clarifications were in part added to address the widespread perception that the requirement for national and county single treasury accounts would limit the financial autonomy of county governments from national government. • The CGFM Bill included an extensive section on financial management in urban areas and cities, while the draft PFM Bill did not include provisions for PFM at this level of government. The final PFM Act included the text from the CGFM Bill on financial management in urban areas and cities, with some alterations. Importantly, this establishes urban areas and cities as sub-units within the county government budget process. As discussed in Chapter 14, when taken together with the provisions of the Urban Areas and Cities Act, 2011, this raises some serious risks for the sustainable financing of urban services which may require conditional intergovernmental transfer arrangements to remedy. • The CGFM Bill placed the Cabinet Secretary responsible for devolved government in charge of intervening in case of a “serious financial problem� in a county and proposed the creation of a County Governments Financial Recovery Service, within the national department responsible for devolved government, in charge of developing county financial recovery plans and monitoring their implementation. It did not, however, authorise the Cabinet Secretary for devolved government to stop the transfer of funds to a county government. The draft PFM Bill included a brief section authorising the Cabinet Secretary for finance to intervene to stop the transfer of funds to county governments in cases of “serious material breach, or persistent material breaches�, but left much of the detail of this procedure to be determined in regulations. The approved PFM Act represents a compromise between the two, clearly putting the Cabinet Secretary for finance in charge of interventions in case of serious material breach or persistent material breach of the Act, and permitting the stoppage of transfer of funds to county governments, but including significant detail regarding the procedure through which this power could be exercised, including a number of mandatory checks and balances. Source: World Bank staff analysis. 150 Devolution without disruption – Pathways to a successful new Kenya Budgeting and Spending at the County level 12.26 In many respects, the biggest changes brought about by the PFM Act will be at the county level. Until now, local authorities have budgeted within the confines of a process guided and constrained by the Ministry of Local Government. Under the PFM Act, county governments will be solely be responsible for a fully-fledged budget process in which they have full discretion over resource allocation. Translating this power into improved service delivery will require effective systems, not just for budgeting, but also for budget management (execution). The following section focuses on four specific areas: (i) the steps in the county budget process; (ii) the baseline of capacity for budgeting that exists in counties at the moment; (iii) the potential for counties to have difficulty in actually spending the money they have budgeted; and, (iv) the importance of managing cash flows for achieving efficient budget execution. County budget process 12.27 County budget processes as mirror those at the national level almost completely. As at the national level, there is full separation between Executive and Legislature, and county assemblies will control the budget process through their committees. County executives are charged with the responsibility for preparing the various documents that comprise the county budget (see Table 12-2), but the county assembly will control the passage of the budget laws. By comparison with other developing countries, this is a very sophisticated budget process. 12.28 The PFM Act includes 13 provisions on the financial management of urban areas and cities. Included is a requirement for urban areas to be funded through revenue sources assigned by the county government, transfers from the county government, and borrowing negotiated by the county government on its behalf.14 In making transfers to an urban area, the county government is required to use objective criteria that reflect the service demand and responsibilities of the urban area. These criteria include; relative population, proportional land area, relative poverty levels, per capita revenue collection, cost/price factors, a minimum amount to ensure effective service delivery, and incentives for prudent fiscal management (in other words, very similar criteria to those recommended by the CRA for the horizontal formula for sharing among counties).15 Capacity for budgeting at the county level 12.29 Budgeting is likely to be the area where there will be the greatest gap between existing capacity and that which is required to comply with the PFM Act. Existing deconcentrated budget processes are highly centralised and more inclined to generate wish-lists than real prioritisation of limited resources. Local authorities do have budgets, but they cover a far more limited range of functions. This means that a massive capacity building effort is required in order to develop the capacity to implement the kind of budget process laid out in Table 12-2. 12.30 Presently, there are no district budgets. While there are district plans coordinated by District Development Officers, district level budget allocations are made on a ministry-by-ministry basis, centrally. District officials in line ministries do submit budget requests to their respective headquarters, but final allocations typically bear little relation to these bids. The details regarding budget allocations for recurrent expenditure at District level are not appropriated and are not published by the line ministries. 14 Sections 172 and 140 PFM Act. 15 PFM Act, section 173(2)). 151 Chapter 12: Managing money for county development Table 12-2: Steps in the county budget process. Timeline Step (reference to PFM Act) 30th August County Executive Member for Finance (CEMF) issues budget circular to all county entities and spending units (s.128). 1st September Integrated Development Plan submitted to County Assembly for approval by County Executive Member for Planning (s.126(3) and 125(1)(a)). 30th September County Budget Review and Outlook Paper (CBROP) submitted to County Executive Committee―contains past fiscal performance for previous year compared to appropriation, fiscal and economic forecasts, impact of actual performance on fiscal responsibility and county fiscal strategy for the previous year (118(1)-(2)). 14th October (14 days later) County Executive Committee considers CBROP and passes with or without amendments (s.118(3)). 21st October (7 days later) County Budget Review and Outlook Paper (CBROP) is tabled in Assembly (s.118(4) (a)) and published as soon as practicable after that (s.118(4)(b)). County Treasury prepares County Fiscal Strategy Paper (CFSP) taking into account views of CRA, public, interested persons and any forum established by legislation. Must be aligned to policy objectives in the national BPS. CFSP is submitted to the County Executive Committee for approval (s.117(1) -(4)). 28th February County Fiscal Strategy Paper (CFSP) is submitted to the County Assembly (s.117(1)-(2)). As soon as possible after County Debt Management Strategy is submitted to the CRA AND IBEC (S.123(3)). 28th February 14th March (within 14 days after County Assembly adopts CFSP with or without amendments (s.117(8)). CFSP being submitted) 21st March (within 7 days) CFSP published and publicised (s.117(8)). 30th April CEMF submits budget estimates. Must be in line with Assembly resolutions on the CFSP. (s.129(2). Budget documents to include detailed list of entities that will receive appropriated money, shown by economic classification and vote. Should also include a summary, statement of how the estimates comply with fiscal responsibility principles and financial objectives, and how CFSP resolutions have been taken into accounts (s.130(6), must be in a form that is clear and understood by the public (s.131(6). County Assembly Clerk submits estimates for County Assembly (i.e., its running costs) to Assembly, with copy to the CEMF (s.129(3)). As soon as practicable after County Executive Member of Finance (CEMF) publishes and publicises estimates presentation to Assembly (s.129(6)). CEMF submits comments on the estimates for the County Assembly, to the County Assembly. Relevant county assembly committee meets to consider the estimates and make recommendations to the County Assembly, and must take into account the views of the CEMF (s.131(2). County Assembly considers estimates and approves with or without amendment, in sufficient time for the county appropriation law to be passed by 30th June (s.131). may amend estimates only if any proposed increase is balanced with a reduction in another appropriation, and any proposed reduction is used to fund the deficit (s.131(3)). After estimates approved, CEMF submits county appropriation bill to the County Assembly (s.129(7)). 15th June County government must submit annual cash flow projection to the Controller of Budget (s.127(1). Source: World Bank staff analysis based on PFM Act. 152 Devolution without disruption – Pathways to a successful new Kenya Table 12-2: Steps in the county budget process (continued). 30th June Appropriation Bill must be passed by County Assembly (s.131(1). CEMF makes a public announcement of the proposed revenue measures (s.132(1)). Finance bill is tabled in the County Assembly at the same time (s.132(2)). Assembly committee considers the Finance Bill and may amend, provided total revenue is consistent with the Fiscal Framework and County Allocation of Revenue Act and taking into account recommendations of the CEMF (s.132(3) & (4)). 90 days after Appropriation law County Assembly considers the Finance Bill and approves with or without passed amendments (s.133). Source: World Bank staff analysis based on PFM Act. It is possible to identify district level development projects within the development estimates, but this usually requires a careful reading of the development estimates, as there are no summary tables by District in the development estimates either. The District therefore only receives notification of the resources allocated to it when the first quarterly Authorities to Incur Expenditure (AIEs) are received. These are also fragmented across departments, and arrive at different times for different ministries. Money cannot legally be spent without an AIE, although their arrival is sometimes significantly later then the transfer of funds (and sometimes precedes it). There is very little flexibility for district officials on how funds are spent―spending must conform rigidly to the allocation across items found in the AIE. 12.31 Local authorities do produce annual budgets. These are compiled in line with a budget circular issued by the Ministry of Local Government, and are subject to review by Ministry officials at Provincial level. Conditional grants through the Local Authority Transfer Fund (LATF)―which is managed by the Ministry―incentivise local authorities to compile budgets according to a uniform calendar, and to respond to requirements for public participation in the selection and prioritisation of projects through the development of a Local Authority Service Delivery Action Plan (LASDAP). However, for the majority of local authorities, only a relatively limited set of services is financed through these budgets (with the exception of those municipalities that are delivering health and education services). While LATF Is used to finance expenditures across a number of different sectors, these are strictly limited to development projects and the local authorities are not responsible for the recurrent costs of running the facilities. 12.32 Bringing together district and local authority budgeting systems offers the opportunity to ensure closer links between planning and budgeting. County administrations will be comprised of staff from Districts―who have very limited experience of preparing and managing budgets―and staff from former local authorities, who do have limited experience of budget management. In line with the aim of there being a single national PFM system, and in anticipation of the difficulty of establishing a budget function from scratch at county level, guidelines and associated templates to guide the formulation of county budgets should be developed. This would also provide an opportunity to integrate planning and budgeting at county level into a single process―in contrast to the highly fragmented approach currently used at central government level. The establishment of the counties also provides an opportunity for planning and budgeting to be overseen by a single institution at county level. Separating planning and budgeting across different institutions limits the relevance of plans since it tends to undermine the extent to which they are translated into budgetary allocations. Integrated county planning and budgeting guidelines, with a clear assignment of responsibility for both planning and budgeting to 153 Chapter 12: Managing money for county development the county treasury, could help to achieve this. The legal framework provided by the PFM Act does support this integrated approach, although the County Governments Act and Urban Areas and Cities Acts also have sections that relate to planning. The interpretation of the three Acts will need to an important determinant of whether the coherent approach to planning and budgeting in the PFM Act is implemented in practice. 12.33 It will also be essential to ensure that county budgets are prepared, executed and reported against using a single country-wide chart of accounts. Without such consistency, it will be impossible to properly assess aggregate county-wide spending patterns across all the counties, and the National Treasury would be hindered in managing aggregate macro fiscal risk. A single national chart of accounts (CoA) will also be essential if the Controller of Budget’s constitutional mandate is to be adequately fulfilled. Unfortunately, establishing a unified CoA will not be straightforward. At present districts and local authorities use separate charts of account. Further, until August 2011/12, budget formulation and execution had each been conducted against differing charts of account by the national government. While this has now been addressed by the production of a single harmonised national chart of accounts, the urban boards and town committees established under the urban areas and cities act (following the abolition of local authorities) will need to adopt the national CoA, which in turn will need to accommodate their unique functions. Budget management at the county level 12.34 In many countries, subnational governments experience difficulty spending the funds released to them in a given year. Experience among Kenya’s local authorities suggests that this is an issue that some counties may face in the future. As Figure 12-3 illustrates, the local authorities in 33 counties ran uncommitted surpluses in FY 2008/09. This group includes many of the historically marginalised counties which are likely to see significant increases in funding under the CRA formula, including Tana River (45 percent surplus), Wajir (43 percent), Marsabit (25 percent), Taita Taveta (17 percent), and Garissa (12 percent). While this may in part relate to the fact that local authorities like to enter the new financial year with a cash surplus in case their first quarter LATF transfers are delayed (which is often the case), it also underlines the low volumes of annual spending in many local authorities at present, especially in areas that stand to receive substantial funding increases under the CRA’s proposed formula. The capacity of PFM, HR and service delivery systems to spend dramatically increased amounts of funding will be an important determinant of the success of the devolution project. In Uganda perceived failures by district governments to manage and spend unconditional decentralised funds (among other shortcomings) effectively led to a reintroduction of tighter national controls, including far greater levels of conditionality than had been the case when decentralisation was first introduced. 12.35 Even after cash is released to subnational spending agencies, there is the additional challenge that cash may not readily reach frontline service delivery units, or that it may be provided with too little flexibility. While this is also a challenge for deconcentrated service delivery, it is even more difficult in decentralized settings because―as is illustrated by the decentralization of education and primary health care to municipalities in Chile16―resources are very often diverted from the intended use by intermediaries. In many cases this is only partially attributable to corruption―it is also simply a result of chronic under funding of local governments.17 In Kenya this will mean developing procedures for 16 Letelier (2006). 17 Kaiser and Kai (2006). 154 Devolution without disruption – Pathways to a successful new Kenya Figure 12-3: Selected Local Authority budget surpluses by county as a % of total revenue, 2009/10. 50 45 40 35 30 25 Percent 20 15 10 5 0 Tana River Wajir Marsabit Tharaka Nithi Kajiado Taita Taveta Kiambu Trans Nzoia Uasin Gishu Meru Garissa Kericho Kwale Mandera Migori Busia Nyandarua Kakamega Nyamira Laikipia Bomet Nakuru Turkana Kisumu Pokot Murang'a Nandi Kitui Bungoma Samburu Nairobi Narok Kili Source: World Bank staff analysis. resource allocation, cash management, and budgetary oversight through the Office of the Controller of Budget, that prioritize and protect funding for devolved service delivery. A crucial part of this control architecture should be budget lines that ring fence funds for service delivery facilities separately from the budget for their administrative supervisors. In health, this is essentially the systemic weaknesses that the Health Systems Support Fund (HSSF)―a direct facility level grant circumventing district treasuries―was set up to address, by providing small grants directly to facilities and circumventing intermediaries. In the longer term it would be better if the PFM system as a whole supported this objective at county level, rather than requiring the establishment of parallel funding flows like HSSF. Importance of cash management 12.36 Within counties cash flow management will be undertaken in line with regulations to be approved by the county assembly. Every county government entity is required to submit an annual cash flow plan to the County Treasury, with a copy for the Controller of Budget. It may help counties to establish good practices from the outset if model or template guidelines are developed in support of county cash management, in line with the principles of a treasury single account. 12.37 The Controller of Budget will also have a very important role to play in oversight of budget implementation in the counties. The Constitution empowers the Controller to “oversee the implementation of the budgets of the national and county governments by authorizing withdrawals from public funds�.18 At county level, where the disciplines of strategic prioritization, budgeting and budget execution will be very new, it is quite likely that there will be significant pressures―technical, financial and political―for counties to deviate from their approved budgets during execution. However, they also need to strike an appropriate balance with budgetary flexibility, providing more room for in- year alterations, than the currently very rigid framework of District Authority to Incur Expenditure (AIE) allows. This raises a risk that even if adequate allocations are made for devolved service delivery in county budgets, the money may not be released to county departments as budgeted during execution. 18 Article 228(4). 155 Chapter 12: Managing money for county development However, the Constitution, PFM Act, and Controller of Budget Act are somewhat ambiguous on exactly what the role of the Controller of Budget should be at county level. While it seems likely that this will involve the authorization of cash releases to spending agencies, this would need to be carefully linked to county cash flow management regulations mentioned above to ensure that it does not become an additional obstacle to counties spending money efficiently. 12.38 One way many countries seek to address these budgetary challenges is through programme based budgeting. This involves moving away from input based budgetary controls (i.e. budgeting by line item) with limited room for in-year reallocation between items, towards a system where budgets are formulated around programmes (e.g. pre-primary education). If the legal structure of county budgets, over a number of years, is amended to legally protect programmes rather than line items, this may be an important way of safeguarding funds for frontline service delivery. Transitioning to the new systems 12.39 Counties will not take on their newly assigned functions and funding overnight―instead there will be a managed transition process. The Transition to Devolved Government Act created the Transition Authority (TA), which is responsible for facilitating the phased transfer of functions to county governments. It will do this by determining the readiness criteria that counties must fulfil in order to take on functions, and assessing whether or not counties have reached a state of readiness (see more detailed discussion of the function transfer process in Chapter 4, and its implications for the overall transition process in Chapter 16). 12.40 The transition period raises considerable challenges for the annual budget process, in particular, on how to budget for devolved functions being temporarily executed at national level. With the elections taking place in mid-March 2013, there will only be around three months before the next fiscal year starts. This is not sufficient time for counties to prepare budgets for the 2013/14 financial year if they are going to follow the process set out the PFM Act (see Table 2). It is unrealistic to expect them to complete a full budget preparation process in that time. Further, many devolved functions will remain the responsibility of the national government during the transition period, before the Transition Authority has determined that they should be fully transferred. The 2013/14 national budget will still need to clearly demonstrate that funds are set aside for the counties, however, as discussed below. The functions and funds that are fully transferred to counties in 2013/14 could be budgeted for centrally through a Transitional national law―including forty-seven county budget schedules showing county allocations disaggregated by function―coordinated by the Transition Authority and Treasury. This would mean that the first full budget preparation process by counties themselves would be in preparation for the 2014/15 financial year. 12.41 Showing all the funding for devolved funds in a single law could also address the potential confusion resulting from asymmetric transfer of multiple devolved functions. Asymmetric transfer of the 14 devolved functions in Schedule 4 to the forty-seven counties (i.e. separate transfer of potentially 658 functions) would create confusion and undermine the accountability of the county governments at a crucial early stage. Furthermore, there is the problem posed by functions that are transferred partway through a budget year. In Chapter 4 a phased approach to function transfer is proposed, under which functions would be ‘bundled’ into groups for transfer in three phases. Each phase would include exactly the same functions for each county―although counties might enter the different phases at 156 Devolution without disruption – Pathways to a successful new Kenya different times. This will help to ensure that citizens know which level of government is responsible for which function during the transition period, thereby improving accountability (counties should not be criticized for under-performance in a function they are not yet responsible for). By setting a clear and transparent transfer process, and perhaps also by asking candidates for the position of Governor to agree to this process prior to the elections, the Transition Authority may help to reduce what will likely be very intense political pressure to immediately transfer functions and funds after the elections, irrespective of county preparedness. Helping Counties Manage Finances Accountably Streamline funding flows to counties 12.42 The predictability of intergovernmental cash releases is extremely important for county level financial management, accountability and service delivery. The problem of subnational governments in many countries being unable to fully spend funds was mentioned above. Late release of transfers from national government is a key contributor to this problem. International experience suggests that where transfers are highly unpredictable, subnational budgets become almost useless, and service delivery suffers (some examples are provided in Box 12-4). Box 12-4: Predictability of intergovernmental transfers: International experience. International experience provides examples of the adverse effects of unpredictable intergovernmental transfers and possible institutional mechanisms to address them: • In Indonesia and the Philippines, poor estimates and unpredictable release of transfers over the budget year distort budget execution. Cambodia is an extreme case: grant disbursements to provinces are so consistently late and so divergent from initially budgeted amounts that budgets themselves have become virtually meaningless. But similar difficulties with the timely release of intergovernmental transfers are fairly widespread in the East Asia region (White and Smoke, 2005). • In Nigeria, local government overdependence on intergovernmental transfer revenues, coupled withuncertainty about the amount and timing, “facilitates local evasion of responsibility under the guise offiscal powerlessness� (Khemani, 2006: p. 292). • In Uganda, the Poverty Action Fund (PAF) established a virtual fund within which budget releases were more predictable than for non-PAF budget lines, and this included non-wage conditional grants. However, the very high degree of earmarking of these conditional grants undermined local government discretion and drew local governments’ focus away from the budget overall (Williamson, 2010). Source: World Bank staff analysis. 12.43 Deviations from approved county transfer amounts in budget execution should be avoided. It will therefore be extremely important to look at the how the flow of funds to county governments will work under the newly emerging legal framework: although funding for counties in the Division of Revenue and County Allocation of Revenue Acts may be more equitable and transparent than in the past, deviations in budget execution (by national government, of these transfers) could significantly undermine this. Article 219 of the Constitution states that ‘A county’s share of revenue raised by the national government shall be transferred to the county without undue delay and without deduction, except when the transfer has been stopped under Article 225’. The PFM Act provides for a more detailed framework for the implementation of Article 219, based on cash management, rather than automatic transfer of one twelfth of the county equitable share each month: 157 Chapter 12: Managing money for county development • By 15th June, every county government shall prepare an annual cash flow projection and submit it to the Controller of Budget, with a copy for the IBEC and the National Treasury. • The National Treasury shall, at the beginning of every quarter, and in any event, not later than the fifteenth day from the commencement of the quarter, disburse monies to county governments. • The disbursement shall be done in accordance with a schedule prepared by the National Treasury in consultation with the Intergovernmental Budget and Economic Council, with the approval of the Senate, and published in the Gazette, as approved, not later than 30th May in every year. 12.44 The PFM Act provisions will help to ensure that the county equitable share and conditional grants are transferred in a timely and predictable manner. There may however still be some tension with the requirement that the National Treasury must not allow the National Exchequer Account to be overdrawn at any time. In case of a conflict between these provisions of the PFM Act, it is not clear which would take precedence. It is also unclear whether the deadline for publication of the disbursement schedule should be interpreted as by 30th May prior to the financial year, or during the financial year. The former would be too soon to include conditional grants to counties made in the national budget, which is approved by 30th June, and would only work if conditional grants are somehow agreed prior to the approval of the national budget estimates. The latter would mean publication 11 months into the year. 12.45 The Controller of Budget―an almost unique financial management role internationally―will have an important role to play in ensuring that the National Treasury disburses funds to counties in line with the gazetted disbursement schedule. A further innovation could be for the Controller to publish in national and county media outlets, the monthly or quarterly cash releases to counties in comparison to the amounts envisaged in the disbursement schedule. This practice has added considerable transparency to intergovernmental flows in Uganda, for example, where releases have been published by the Ministry of Finance since 1997. In Nigeria the Federal Ministry of Finance developed a practice of publishing the federal, state, and local government shares of revenue from the country’s federal account on a monthly basis. This would help to ensure adherence to the disbursement schedule, and assuming this is adhered to, would transfer public scrutiny to how effectively the counties are using their transfers―Governors would not be able to conceal the fact that funds had been released or the amounts.19 Make use of performance information 12.46 As well as setting policy, a key role for national government will be setting standards and monitoring performance. This should not be seen purely as a top down control process―instead this process should take full advantage of intergovernmental coordination arrangements. It also needs to take into account ways of unleashing the competitive potential between counties. Experience from other countries suggests that use of competitive pressures through league tables and rankings can provide stronger incentives than the financial carrots and sticks provided by conditional grants. Such mechanisms can also provide an important basis for comparative lesson learning and experience sharing between counties. This could help both in tracking advances in building essential systems―especially for Public Financial Management―as well as providing early warning of service delivery problems in devolved functions, bringing this to the attention of the governor and national government. This is especially important in basic service delivery functions, especially primary health care. 19 World Bank (2008). 158 Devolution without disruption – Pathways to a successful new Kenya 12.47 One way of promoting this would be through a county performance assessment tool. This would create incentives for good performance by: (i) focusing county leaders’ attention on key indicators of basic service delivery; (ii) signalling important gaps in capacity which the national government is responsible to address; (iii) generating a constructive atmosphere of competition between counties; (iv). providing a publicly accessible information base for county citizens to judge the relative performance of their counties; and, (v) providing a framework to potentially link some county funding to performance achievements in order to create further incentives for improvement. 12.48 The performance monitoring/assessment tool could include a small, but broad set of county-level indicators. They could cover service delivery results, key public administration systems including financial management and procurement, and implementation of social accountability mechanisms, that all counties will collect and report data on. The institutional and administrative arrangements could include: (i) detailed roles and responsibilities of stakeholders at national and international levels; (ii) horizontal and vertical reporting arrangements across government; (iii) data validation mechanisms and methods to ensure adequate popular participation; and, (iv) dissemination and use of progress reports and results data. Capacity building 12.49 A massive capacity building task faces Kenya, to empower its new county governments to manage funds accountably. Nevertheless, some comfort can be gained from the experience of Uganda. Starting from a much lower base of capacity in the mid-1990s than Kenya has today, Uganda has successfully implemented major improvements in district government’s PFM capacity over the last 15-18 years (see Box 12-6). Box 12-5: Success factors for subnational performance monitoring systems. Dumas and Kaiser (2009) review international experience in the development of subnational performance monitoring systems, and distil a number of success factors that should inform system design processes : • Strong partnership between central and subnational levels of government is essential. Rather than creating a compliance tool, active participation and use of information can result in a more legitimate system in the eyes of subnational governments. • A well designed system needs to consider: coupling of indicators and performance objectives; setting realistic targets; adequate use of incentives, and; emphasis on satisfying demand for performance information (what needs to be measured? why is it important? who wants it? • Think big, start small. Make a virtue of simplicity and avoid over engineered systems. Initial diagnostic work should avoid duplication or overlapping efforts. Costs should be kept at manageable levels. Ultimately, usage of the information produced is the measure of success. The system should be inherently flexible, allowing for permanent, progressive and incremental improvements • Country ownership is crucial! Donor driven efforts are usually not sustainable, and a strong champion is usually needed. Source: Dumas and Kaiser (2009). 159 Chapter 12: Managing money for county development Box 12-6: Decentralization and local capacity: Ugandan Experience. Uganda’s district local governments were introduced in the mid-1990s building on the existing participatory democracy structures of the resistance councils. Administrative and fiscal decentralization was introduced over a three-year period. Financial procedures were specified in the 1998 Local Government Financial and Accounting Regulations. At the outset, planning, budgeting and financial management capacity was particularly weak. Early development plans were ‘wish-lists’ and budgets were unrealistic, late, and bore little relationship to development plans. An early stimulus to greater transparency of local government budgeting came in the form of a requirement from 1997 to publish in newspapers national transfer releases to district governments, and to require schools to put up notices showing what they received from the district government. This approach was remarkably successful in increasing the amount of funds that districts passed on to service delivery facilities. In 1998/99 a Local Government Budget Framework Paper process was introduced that mirrored a national process that had been introduced two years earlier. Local governments were invited to a national budget conference, and workshops were held with local government staff. In their individual local gover5nment budget framework papers, the districts and municipalities reviewed revenue performance and made medium-term revenue projections, reviewed sector performance and made medium term spending plans. With guidance from the national Ministry of Finance these local medium-term budgeting processes improved over time. In 2000/01 a more structured basis was introduced for accounting for Poverty Action Fund (PAF) conditional grants, funded mainly by development partners. While these conditional grants have proliferated over time and impose greatrigidity on local government budgets, the framework did help introduce a more structured relationship of accountability. In the second year of PAF reporting the Ministry of Finance introduced a requirement for bank reconciliations to be reported with grant conditions. Although reality did not always follow the legal framework for planning, budgeting and accounting, the Local Government Development Programme helped to increase compliance, through a combination of incentives in the form of performance-linked unconditional grants, and capacity building grants to address capacity weaknesses. The program started in the mid-1990s and was rolled out to the whole country in 2003. Source: World Bank staff analysis. 160 Devolution without disruption – Pathways to a successful new Kenya CHAPTER 12: KEY INSIGHTS / RECOMMENDATIONS. The new Constitution has provided a welcome opportunity to rationalize the PFM legal framework, as well as replacing or consolidating existing laws, the PFM Act provides a legal framework for PFM in counties and urban areas. Budgeting for transition will be particularly challenging because counties will only have very limited time to prepare budgets for 2013/14 and because they will take on their assigned functions gradually, following an asymmetric transfer process. Under the new dispensation, the national legislature will have vastly increased power in the budget; this could potentially derail the budget process at national and county levels if mechanisms are not put in place to manage the impact of legislative budgeting. The predictability of funding flows to counties will be crucially important for county level financial management, accountability and service delivery―the Controller of Budget could focus on monitoring and publicizing the timeliness of intergovernmental transfers. The budget process at county level will have to be built up from scratch as there is currently only limited experience and capacity at the local level, especially within districts. This will require central coordination and support for county PFM system building. Budget implementation and cash flow management at county level will be challenging. *** During the transition period, the national budget could include county votes showing total county allocations even if a portion remains executed by the national government on behalf of counties. This could include a transitional budget law, passed at national level for 2013/14 with forty-seven county budgets schedules showing county allocations disaggregrated by function. To mitigate the risk of Executive-Legislative gridlock during the budget process a two-step budget process could be introduced with parliamentary votes at each stage to generate consensus around the fiscal framework. The controller of budget should be empowered to oversee the timely and predictable release of county transfers. Publicizing the amount and timing of these transfers would be an important component of this oversight. Building county PFM systems from scratch will require central support to capacity building, monitoring of county progress against clear benchmarks and standardized guidelines, and templates developed nationally (for example to help counties develop an integrated planning and budget process). 161 CHAP TE R THIR TE E N Social accountability: Transparency, accountability and participation 13.1 There are three requirements to prevent accountability failures in decentralization. A sustained effort to: (i) make information transparent; (ii) enable citizens to participate in local government; and, (iii) hold local leaders to account. These three elements form the basis for a system of social accountability at the local level. Devolution presents a particular challenge to service delivery, breaking apart more traditional, centralized accountability relationships, and requiring new ones to be established. 13.2 The challenge is compounded where the historical precedent is not well established. Contrary to the common expectation, that decentralization enhances accountability and efficiency in service delivery, empirical research illustrates multiple examples of accountability failures, leading to corruption and substandard service delivery.1 Typical challenges to successful devolution include limited capacity, poor information systems, weak checks and balances, an often poorly organized civil society at the local level, and an absence of local media. Such weaknesses at the local level are understandable, given the common history of centralized government, which leads to a tendency to focus these things in the center, rather than the local level. These various elements of local accountability systems will not emerge overnight, but can develop with consistent efforts. 13.3 But there are significant potential payoffs for service delivery and citizen empowerment if challenges are tackled from the get-go. This particularly, as the traditional top-down accountability relationships are weakened with devolution. Top down relationships become more difficult due to county government autonomy, and the distance from the central government. The potential outcomes include improved targeting of resources, reduced corruption, increased efficiency and overall better governance. 13.4 Transparency, participation and accountability are also clear requirements in the new Constitution. The Constitution includes provisions for clear fiscal reporting and citizen access to timely, accurate information, participation in Public Financial Management and policy formulation, citizen budget monitoring, and independent commissions to handle citizen complaints (Articles 35, 201, 232). The Constitution tasked Parliament with translating these principles into a legal framework for PFM with establishing mechanisms to ensure implementation (Article 225). 13.5 Embedding these elements in the legal and regulatory framework is important, but developing effective county systems and capacities will be critical for enabling participation and accountability. Current laws can institute the principles of participation, and require the public availability of information, which are both important. However, developing functioning systems will be paramount, including a monitoring and evaluation system, that assesses performance and needs across counties, a PFM system that reliably tracks budgets and expenditures against outputs and produces high quality information, the production of financial and performance information that is relevant to citizens, 1 Bardhan and Mookherjee (2006). 162 Devolution without disruption – Pathways to a successful new Kenya a tailored set of participatory institutions to the local context, and efficient mechanisms to gather citizen feedback. A key need is to put in place performance management systems to assess central and subnational government on a range of indicators.2 The basic elements of a social accountability system 13.6 In addition to functioning systems to track finances and performance, social accountability in local government requires: (i) Transparency of public finances and performance is ensured through rules, mechanisms, and capacities for sharing information on government programs, budgets, expenditures, and results with citizens. (ii) Participation mechanisms to enable Figure 13-1: Elements of social accountability systems. citizens to participate in setting budget priorities and monitor expenditures, and assessing Government service delivery performance; they also include feedback systems, Transparency: Participation which provide citizens with the information Accountability and Feedback: for citizens information opportunity to provide comments from citizens and grievances. (iii) Accountability mechanisms include both direct Citizens and indirect relationships, where service providers are sanctioned if they fail to meet an established Source: World Bank. standard (see Figure 13-1). 13.7 Transparency on government programs, rules, finances, and performance is critical. To participate effectively, citizens need to access financial management and performance information on public service delivery, in formats that are readily available, clearly presented, timely and relevant. Online access to information is the easiest means of making this information accessible to a national audience, but paper documents should also be available in public offices on request. Transparency of fiscal information at the facility and project level, including schools, health centers and Community Driven Projects (CDD) projects, best ensures that this information is relevant for citizens. Transparency can be safeguarded through legal provisions that mandate the publication of information on programs, finances, and performance; these legal provisions may be found in a PFM law, sectoral legislation such as a Water Act, or in regulations. The experience of Uganda shows how powerful transparency can be in affecting service delivery outcomes (see Box 13-1). 13.8 Civil society can play a critical role in advocating for national and subnational government systems, that enable transparency and effective citizen participation, in providing feedback on service delivery and policy priorities, and in communicating and verifying information to citizens. Civil society can mobilize citizens to provide feedback on policy and budget priorities, as well as to monitor the results of these programs. Civil society can also check the accuracy of, and demystify information 2 “Indicators contribute to enhancing the efficiency and effectiveness of sub-central service delivery by: (i) revealing information and reducing information asymmetries which exist among levels of government; and (ii) encouraging performance improvements by altering the incentives which sub-central governments face. The choice of the objectives the systems will serve is critical. The objectives contribute to subsequent decisions regarding design, link directly to decisions regarding the use of the indicators, and ultimately form the basis for assessing if the system has been effective. A predominately top-down approach to design and use of indicators by the central government can be perceived as an ex-post substitute for ex-ante control of sub-central public services. Intergovernmental collaboration is essential for ameliorating this tension and producing relevant and useful information�. 163 Chapter 13: Social accountability: Transparency, accountability and participation Box 13-1: Tracking expenditures in Uganda: Promoting efficiency. Once revealed that less than 30 percent of allocated capitation money was reaching the schools on average at the end of 1995, the government acted immediately to improve the flow of information, and make budget allocations transparent by: (i) publishing amounts transferred to the districts in newspapers and radio broadcasts; (ii) requiring schools to maintain public notice boards to post monthly transfer of funds; (iii) legally provisioning for accountability and information dissemination in the 1997 Local Governance Act; and, (iv) requiring districts to deposit all grants to schools in their own accounts, and delegating authority for procurement from the center to the schools. By 1999, capitation grants received by the schools had almost reached 100 percent. Source: World Bank staff analysis. presented by government. Government may offer information to citizens, but fail to present it clearly or communicate effectively to citizens. Kenya’s civil society has extensive experience in monitoring decentralized government programs, such as the Constituency Development Fund (CDF) and the Local Authority Transfer Fund (LATF). Many Kenya civil society groups have grassroots networks that allows them to communicate directly with and mobilize citizens. Such communication is critical for presenting and explaining information to citizens. Local media, likewise, is essential for communication between government and citizens. Newspapers, television and radio are critical media outlets for citizens to access information. Like CSOs, media can play a critical oversight role for local governments, uncovering relevant information for citizens who often build on pieces of information made public. Moreover, they are clearly capable of spreading this information. Local media tends to be weaker in less populated areas, however, leading to the need for special focus on communication with more isolated populations. 13.9 Transparency of national government disbursements to local governments is key, but often overlooked. It is common in devolved systems for lower-level governments to blame poor service delivery on the national government’s failure, or delay in release of funds. In Kenya, the Controller of Budget is already constitutionally mandated to oversee the implementation of the budgets of the national and county governments. In addition to this oversight role s/he could be required in PFM legislation to publish a simple monthly report on intergovernmental transfers to show the public the volume and frequency of transfers. This approach was utilized in Nigeria’s decentralized system, and helped to diffuse public mistrust of central government, focusing public scrutiny and pressure on the use of funds. Of course, information alone is not a sufficient condition to improve service delivery, but it is an important ingredient in a broader set of institutional reforms. 13.10 Toward an accountability model that empowers clients. Many developing countries rely primarily on the traditional top-down means of accountability, but a number of other potential routes exist.3 For citizens, the three main routes of direct accountability include: (i) political accountability, in which citizens hold leaders accountable through the ballot box; (ii) public accountability,4 which includes popular mobilization, the media and a general form of social status and embarrassment; and, (iii) a market mechanism, where, if there are competing service providers, the citizen can choose another. These forms of direct accountability form the foundation of client power. 3 Other formal accountability mechanisms include horizontal accountability from the legislative and judicial branches of government and diagonal accountability from supreme audit institutions such as the ombudsman or an anti-corruption commission. 4 Political accountability is generally not a form of client power directly toward service providers, but it is a form of direct accountability in PFM of local governments, particularly as relates to the budget. 164 Devolution without disruption – Pathways to a successful new Kenya 13.11 Citizens also hold an indirect Figure 13-2: Accountability framework for service delivery. relationship of accountability The State with service providers, providing Co Politicians Policymakers m information to the state, who can pa ct then act on this information. Figure ice ro ute of accou nt Long abili Vo ty 13-2 illustrates that citizens are able to influence service providers through Citizens/Clients Short route Providers a two-step process, the long route of Coalitions/Inclusions Management Client power accountability. First, citizens express Non-poor Poor Frontline Organizations their preferences or grievances to the state (voice), to which, because of elections, politicians and policy- Services makers listen; then, service providers Source: World Bank (2004). respond to the authority of the politicians and policy-makers, because of the top-down accountability relationship between them (compact). 13.12 Moreover, transparency efforts should adapt to the accountability context. Transparency means providing information to citizens, but information in and of itself will not change the behavior of service providers; rather, that information should inspire some form of reward or sanction from citizens. An active media, organized citizen groups, and even a functioning judicial system, are all vehicles for citizens to mobilize around issues, and will allow transparency to lead to sanctions from citizens, and thus lead to improvements in service delivery. 13.13 Participation mechanisms take different forms, but the same principles must be met to be effective. Some participation mechanisms enable citizens to express their views on development priorities, expressing their needs and preferences for action from government. Others are feedback systems, such as complaint mechanisms, customer satisfaction surveys, etc., which enable citizens to directly provide input on service delivery performance, and the use of funds. To be effective and fair, these mechanisms must be open to the public, and especially to marginalized groups, well designed to allow substantive input, and relevant to citizens so that their participation is meaningful. Too often, participation mechanisms become a means of checking a box, with no substantive discussion, and little willingness of government to listen. Or, at times, participatory bodies become a vehicle for only wealthy citizens to participate, while the poor, and especially marginalized groups, are not invited or fail to participate effectively. As described below, the Local Authority Service Delivery Action Plan (LASDAP) and CDF projects are often accused of these limitations. 13.14 Participation and feedback mechanisms rely heavily on the responsiveness of powerful government officials. Only in extremely rare cases do citizens have the power to actually execute their decisions, or ensure their feedback is taken into consideration. Instead, the recipient of the information must be willing to act on the information, sanctioning substandard behavior if necessary. For example, in Kenya, Water Action Groups provide feedback on the quality of water services in Nairobi, Mombasa and Kisumu. If there are breaks in water pipes or problems with billing, these groups channel complaints to the respective water service provider, who are responsible for taking action.5 5 Such a system works because of the responsiveness of the WSP, though a threat of elevating the complaint to the Water Service Board, an action that is almost never done. 165 Chapter 13: Social accountability: Transparency, accountability and participation 13.15 Effective citizen participation requires information sharing and participatory feedback mechanisms ―a piecemeal approach is unlikely to be successful. Uninformed citizens cannot participate effectively, meaning that key fiscal and performance reports, especially the budget, must be made public in a timely and accessible manner. Citizens who participate substantially in the identification and design of a project, are much more likely to provide feedback on it, such that participation promotes accountability. These elements of social accountability can reinforce one another. Learning from past experiences with devolved funds 13.16 Kenya can draw on its extensive experience with devolved funds to reinforce both upward and downward accountability by local authorities. The Local Authority Transfer Fund system seeks to reinforce both citizen participation and timely submission of plans and financial reports, by linking a proportion of funding to compliance with these requirements. A Local Authority Service Delivery Action Plan (LASDAP) must be prepared annually with input from citizens. Similarly, the Constituency Development Fund offers lessons for transparency, participation and accountability. 13.17 However, a number of reports document areas for improvement for public participation in Kenya’s devolved funds, such as CDF and LATF. For instance, one report6 states, “Limited participation during the consultations and consensus meetings on the projects, implies that LASDAP does not always reflect the priorities of the communities.� Another report7 observes, “… Figure 13-3: Perceptions of Local Authorities in terms of participation and transparency. there are concerns that CDF 90 monies are not managed in a 80 Access to Budget transparent manner; that many 70 Participation in Council CDF projects are not useful to 60 Transparency of Council local communities; and that local Consultation of Public by Council 50 Percent Complaint Mechanisms communities are not sufficiently 40 Don't Use Funds for Private Gain involved in its management.� 30 These negative claims, particularly 20 in reference to local authorities, 10 appear to be supported by data 0 from the Afrobarometer survey of Very/ Fairly Badly Very/ Fairly Well Don't Know 2009 (see Figure 13-3). Source: Afrobarometer survey (2009). 13.18 Still, there is room for improving participatory practices based on the experiences of these decentralized funds. For example, the LATF system was largely successful in ensuring reporting by local authorities, even though the information they produced is not usually made public. The LATF regulations specify in detail the penalties for late submission of required information, but there is no requirement in the Act for the reports to be made public, so the information has typically not been made available to citizens. The Public Expenditure Review of 2010 also observed that the quality of reports was quite substandard, reinforcing the need for enhanced capacity and transparency. The LASDAP process is intended to include citizens in the local authority planning process, through public meetings, but several reports8 suggest that attendance is often limited and budgets are not well aligned with the priorities identified in the LASDAP. 6 Kenya Local Government Reform Programme (2007). “Study on the Impact of the Local Authority Service Delivery Action Plan.� Ministry of Local Government. p. x. 7 Gikonyo (2008). 8 Kenya Local Government Reform Programme (2007). 166 Devolution without disruption – Pathways to a successful new Kenya 13.19 CDF is fairly transparent with its finances, leading to greater accountability for projects. The CDF website places its financial year proposals on its website, for public scrutiny, although various sources note that the website has been down or not updated for significant periods of time.9 The CDF allocations have their own separate account, and are dedicated to just a few projects. This stands in contrast to the LATF funds, which are transferred to the local authorities, mixed with the local revenues in their accounts and are dedicated to various types of spending. 13.20 The CDF experience also shows that there is an important role for civil society in auditing projects. A number of CSOs perform audits of CDF, including the National Taxpayers Association and Kenya Public Funds Monitoring Network. These organizations evaluate the proper execution of the project, identifying inconsistencies in spending, and engage with citizens to hold their local CDF committees accountable. Map Kibera recently developed a community tracking and mapping website, to provide citizens with access to official details, and current information regarding CDF projects.10 13.21 Drawing on the LATF system, the production of financial and performance reports could be tied to transfers, and these documents should be publicized, especially on the internet. The production of these documents could be among the performance conditions attached to performance grants discussed earlier. A more robust system of performance funding would build on the LATF, by also requiring publication of the reports on the internet, and providing a system of independent verification to confirm their accuracy. Transparency of this information will better inform citizens, and will also aid in creating pressure, to improve the quality of the reports. 13.22 The LATF system also illustrates that, if staff costs are not controlled, there is little financing to support projects preferred by citizens. According to the Public Expenditure Review of 2010, no local authority successfully spent 50 percent of their budget on capital expenditures, as is intended under LATF regulations. Instead, much of these funds are used to finance administrative costs, primarily salaries, often leaving relatively small amount of financing for services and projects. While staff can contribute to the delivery of services, hiring practices are often driven by the desire to provide employment to select individuals. If staff costs use up much of the budget, it leaves little financing for those projects and services that citizens often demand from government, which reduces the value of citizen participation at the local level. 13.23 The LATF experience further underscores the need to strengthen the financial reporting system, primarily in terms of capacity and verification of reports. Information can be made transparent, but that information may fail to meet a minimum level of quality, reducing its usefulness. Local PFM capacity building is necessary for county governments, along with a reporting system that seeks to provide accountability for the quality of financial reports, particularly with audits. A strong financial reporting system, which includes the timely production of accurate financial information, may then support citizen participation by making quality information public. 13.24 Building on the LASDAP system, devolution laws could require county governments to involve citizens in the budget and planning process, give the county administration responsibility for facilitating this, and integrate planning and budgeting more closely. Responsibility for designing opportunities for 9 See http://www.cdf.go.ke/project-monitoring. 10 See http://cdf.org 167 Chapter 13: Social accountability: Transparency, accountability and participation participation is better vested in the county administration, so that it remains political. The Governor and assembly members should be invited to participate, but should not control who attends. The county administration should also be responsible for upholding certain good practices of participation, such as openness, inclusion of minorities and marginalized groups, and clear communication of events to the general public. Planning and budgeting units should be under a single county department, to help make sure that planning actually feeds into budget development (and vice versa). Supporting institutions for a functioning social accountability system 13.25 Social accountability systems for PFM do not operate in isolation of the larger PFM systems; rather they depend critically on them. In general, a better functioning PFM system leads to a more effective social accountability system. This is a further reason why one cannot expect to achieve a fully functioning social accountability system overnight, but rather, it improves hand-in-hand with the larger PFM system. 13.26 The goal is a system to produce financial and performance information that is dependable and accurate. As highlighted above, an enforceable accountability mechanism is essential for this information system to function properly, with clear penalties for failure to produce accurate financial and performance information in a timely manner. Capacity building for local officials responsible for producing this information may be necessary as well. 13.27 A second priority is to strengthen the relationship between planning, in which citizen participation is commonly included, and the budget process. Participatory processes are quite costly in terms of both expenses and staff time. Planning processes tend to utilize citizen participation processes more than budgetary processes, because planning exercises are conducted only once every several years, as opposed to the annual budget process. Moreover, more infrequent participatory processes make sense, when they are more intensive, because citizen preferences tend not to change from year to year, but rather every several years. To ensure that the participation that is focused in the planning process is meaningful, there must be a clear connection between the annual budget and the multi-year planning process; an administrative unification of planning and budgets is the best way to ensure this. 13.28 The legal and regulatory Figure 13-4: An example of a toolbox for county governments. frameworks are not the only means of supporting social Budget Formulation accountability: county governors All Fed. Budgets-Canada Municipal Budget, Brazil and administrators should have easy access to social accountability toolbox to use Performance Budget Review Monitoring Civic & Analysis voluntarily, for improved service India & Filipine Engagement DISHA India, delivery. A toolbox is a set of Report Card DASA, S. Africa social accountability interventions that target specific problems, Budget/Expenditure and function in certain contexts. Tracking PETS-Uganda, Popular tools include participatory G-Watch-Phil budgeting, scorecards, social audits Source: World Bank Demand for Good Governance Website (2012). 168 Devolution without disruption – Pathways to a successful new Kenya and procurement oversight committees. Such a toolbox is useful because legislation and regulation tend to provide a basic floor for social accountability only. Numerous social accountability tools do not appear in legislation. Government along with civil society groups such as the Association for Local Governments in Kenya, should work towards providing a social accountability toolbox for leaders to use, depending on the local context, and the particular results desired. Such a toolbox should build upon successful local experiences with social accountability. Conclusions 13.29 Instituting effective social accountability systems in county governments will be a multi-faceted, long-term process. A complete transformation of citizens’ relationship to a county government clearly cannot happen overnight, even with a perfect legislative backdrop. However, legislation forms a first necessary step toward establishing these systems, and effective legislation provides fertile ground for these systems to grow upon. 13.30 Kenya’s constitution and supporting legal framework emphasize transparency and Public Participation (PP). Multiple articles in the Constitution provide for participation of the people as part of the national values and principles of governance, and require national and county government to involve the public in the legislative and Public Financial Management processes of government.11 The Public Financial Management law provides for public participation in form of consultation in the budget process, disclosure of public financial report and statements, public membership and involvement in PFM created institutions, and sanction as enforcement measures. The Parliamentary Budget Office is instructed to allow for Public Participation in the exercise of its mandate (sec. 10). In the Preparation of the Budget Policy Statement, the views of the Public are to be taken into consideration in terms of section 25(5)(f), as aided by 25(6). There is also a requirement for the Budget Policy Statement to be published and publicized. 13.31 The emerging legal framework also provides for citizen participation in the budget process at national and county levels. The Cabinet Secretary responsible for PFM is required to come up with regulations specifying how, where and when the public is to be involved in the budget process, the public is to participate in the preparation of the budget estimates of the Judiciary and Parliamentary Service Commission, and the relevant National Assembly Committee is to factor in the views of the public, in their considerations of the estimates laid before it, prior to submission to the national assembly. At the county level, It is required that the public be consulted in preparation of the County Fiscal Strategy Paper, and the County Budget circular is to prescribe the manner in which the public will participate. The relevant committee of the County Assembly is required to take into account public views in considering budget estimates, and the accounting officer of an urban area or city is to ensure that the public participates in the preparation of the annual budget estimates/strategic plan. There are also requirements for various budget documents (e.g., Annual Reports, Quarterly Report, Pre and Post Election reports) to be published and publicised within laid out timelines. Public involvement is also specified in the County Budget and Economic Forum (counties are expected to form county budgets and economic forums for purposes of providing means of consultations over the budget process by the county inhabitants). 11 Article 69(1)(d) – PP in the management, conversation and preservation of the environment. Article 118 - parliament to facilitate PP and involvement in the legislative and other business of Parliament and its committees. Article 196 (1) (b) –It requires that the county assembly facilitates PP in the legislative and other business of the assembly. Article 201(a) also outlines PP as one of the principles of public finance alongside openness and accountability. 169 Chapter 13: Social accountability: Transparency, accountability and participation 13.32 Kenya has the opportunity to be a leader in Africa with respect to transparency and local government accountability, but it will not be easy. A strong legal framework is important, but building effective participation and accountability mechanisms will also require substantial capacity at both levels of government. Capacity will be needed at the national level to develop systems (including for soliciting citizen feedback and registering complaints), and at the local level to implement them (including transforming complex financial information into formats that are accessible by citizens). This is an important component of a project of capacity building, to support the effectiveness of devolved government. It is critical that these long-term goals should not be overlooked in the preoccupation with immediate and urgent concerns, that will inevitably accompany the early transition period. Translating new legal provisions into national and county systems that enable and benefit from public participation will require careful design of county systems and extensive capacity building. Key elements for consideration as this process unfolds include: 13.33 Action 1: Making public financial information publicly available, and linked to results, in formats that citizens can understand. Information and transparency form the foundation of accountability. Three types of information together make government most accountable: budget information, expenditure information and results/outputs. Expenditure information should eventually be linked to the results achieved, so that a meaningful assessment can be made of whether the spending achieved value for money. While governments sometimes share this information, its effect is often minimal because the information is difficult to acquire or is in a format that is incomprehensible. As a result, this information should be readily available, clearly presented, and timely, to maximize its use by citizens. 13.34 Action 2: Creating space for citizen participation. Around the world, citizen participation in the local government budget process is becoming increasingly common, a well known example being participatory budgets in Brazil. This participation commonly focuses on both establishing citizen priorities and scrutinizing the budget proposal by local governments. Participation is most effective when multi-year plans are linked to budgets in local governments, given that participation is generally focused on this planning stage. However, for participation to be effective, citizens must be provided with timely and accessible information on options and trade-offs, and be asked to comment on budget areas that are meaningful to them. Participation may be facilitated by a mandate in legislation, as citizens have a legal hook by which to demand a substantive process. At the same time, a mandate should not dictate the exact form of participation, which is best left to the county governments. 13.35 Action 3: Ensure accounting systems can track spending on individual projects and by service delivery units. Strong accounting systems are the bedrock of good accountability. The international framework for measuring effective financial management systems, Public Expenditure and Financial Accountability (PEFA), recognizes the importance of being able to track funding to service delivery units. Charts of accounts should also be able to track what has been spent on individual projects. Information at this level of detail is essential for tracking diversion of funds to less important administration expenses. 13.36 Action 4: Tailor particular social accountability systems to the local context via a participatory process, with scope for adaptation in the future. Social accountability systems must be tailored in a manner that responds to citizen needs. A simplistic, rules-based approach to social accountability will fail to reach the necessary outcomes. Legislation is particularly critical to make information transparent in a meaningful way, and to provide the general principles for participation and transparency. However, 170 Devolution without disruption – Pathways to a successful new Kenya with this legislative backdrop, the social accountability systems must then be adapted to the local context, which includes the size of the county, level of urbanization, capacity of the county government, number of marginalized groups and others. Such a process of adaptation is best done, not in a top down manner, but in consultation with the county government, along with civil society. In other words, participatory processes should be designed using a process of participation. 13.37 Action 5: Develop a toolbox to target the special needs of county governments. Legislation can attempt to provide a minimal benchmark for a social accountability system, but many local leaders may seek to go beyond the minimum, and use special social accountability tools. As mentioned, such tools include social audits, scorecards, participatory budgeting and procurement oversight committees. Government and civil society play a critical role in developing and disseminating such a toolbox. CHAPTER 13: KEY INSIGHTS / RECOMMENDATIONS. Increased accountability of government is at the heart of the Constitution, but nothing guarantees that decentralization increases accountability and efficiency in service provision. At the local level weaknesses include limited capacity, poor information systems, weak checks and balances, and often poorly organized civil society. Setting up a system of social accountability implies: fiscal transparency, participation mechanisms, and accountability mechanisms. The experience with devolved funds provides useful lessons on how to strengthen both upward and downward accountability. *** Public financial information (including results) should be made public in a way that allows citizens to assess the efficiency and effectiveness of national and county spending. Citizens should be actively involved in planning and budgeting at local level,and equipped with the tools to make significant contributions. Accounting systems at county level should allow track spending on individual projects and by service delivery unit. Social accountability toolkits should be developed for the benefit of local administrators. 171 C HAP TE R FO U R TE E N Financing and management of urban areas 14.1 Unless corrective action is taken, Kenya’s cities may well be the big losers of the devolution process. This would be dramatic as Kenya’s cities (just like other cities worldwide) are growth engines for the entire country, and the main source of own fiscal revenues for local administrations. Kenya’s devolution is unique in that it involves simultaneous decentralization of key services and resources from the central to the county level, but also recentralization of urban management, currently provided by the local authorities, to the new county administrations. But unless arrangements are made to adequately resource and manage Kenya’s cities, the transition to devolved government could well result in failing urban service delivery, which would hurt residents and businesses, and so compromise the country’s growth prospects. 14.2 Urban governance arrangements still need to be finalised. In the new dispensation, cities will be managed by unelected boards, directly appointed by the county executive. They will also have limited powers and discretion. Moreover, all but three of Kenya’s current local authorities will meet the conditions required to become cities or municipal boards. Instead many major urban centers will be governed by town committees with weak and unclearly specified management arrangements. Urban development is good for economic growth 14.3 Kenya is experiencing a demographic Figure 14-1: Kenya's demographic transformation. and geographic transformation of which urbanization is a central feature. The 90 80 Total country’s population is growing, but its 70 Population Population millions demographic profile is also changing 60 50 Working Ages (see Figure 14-1). First, there are more 40 15 - 64 people who are of working age, which 30 0 - 14 means that per capita incomes will 20 10 Older 65+ increase. Second, people are moving 0 to urban areas. Although Kenya is still 1950 1960 1970 1989 1990 2000 2010 2020 2030 2040 2050 a predominantly rural country, the 80+ population of urban areas is growing Male 2010 70 - 74 Male 2030 faster than rural areas, largely on account Female 2010 of migration. By 2030, almost 50 percent 60 - 64 Female 2030 of Kenyans are predicted to live in urban 50 - 54 Age group areas. Most of them will have come 40 - 44 from other parts of Kenya, seeking the 30 - 34 opportunities that urban areas offer. 20 - 24 10 - 14 14.4 Urbanisation and economic growth go 0 - 14 6 4 2 0 2 4 6 hand-in-hand. As countries grow, they Population (thousands) also urbanise and vice versa (see Figure Source: World Bank (2011c). 172 Devolution without disruption – Pathways to a successful new Kenya Figure 14-2: Richer countries are more urbanized. 100 Population of urban residents in 2009 (percent) 80 60 40 Kenya 20 Uganda Tanzania 0 4 6 8 10 12 Log of GDP per capita in 2009 Source: World Bank (2011c). 14-2). In China, 120 cities account for three-quarters of the country’s GDP. The economic activity that takes place in urban areas is on average more productive. This is because firms and people benefit from being located close to one another, generating the “agglomeration economies�, that are beneficial to economic growth. Firms find easier access to the skills and services they need, while individuals can tap into a larger employment market. 14.5 Urban and rural development are not competing but complementary objectives. Fostering the economic development potential of urban centers need not disadvantage rural residents. Urban areas are the main source of fiscal own revenues for local governments, which also provide services to rural residents. With the right policies, all citizens can share in the benefits of economic growth, through better access to higher quality social services, and basic infrastructure. Likewise providing better access to services in rural areas fosters better economic growth in urban areas. It helps to ensure that people move for the right reasons (because they have marketable skills), rather than for the wrong reasons (because they cannot access social services). Kenya’s devolution arrangements have the potential to improve access by rural residents to health and other important social services, provided that county governments have sufficient resources, and the capacity to translate them into quality services. 14.6. The economic and social benefits of urbanization will only materialize if urban areas are well managed. Urban areas need policies and services that are business-friendly, and keep cities liveable. These policies include flexible land-use planning, and zoning that facilitates conversion of land for different purposes including affordable housing; connective infrastructure like transport, water and sewage; and social and other basic services that reach all residents. Conversely, the negative impacts of poor urban management―traffic congestion, pollution, slumification and crime (colloquially described as the problems of “time, grime and crime�)―harm economic growth. The right kind of urban governance is critical for good urban management. The remainder of this chapter considers the implications of Kenya’s devolution policy choices, for the management of Kenya’s increasingly important urban areas. Urban management under devolution 14.7 Existing arrangements for urban management will be profoundly affected in the new dispensation. The current system of local authorities, with elected councillors and significant discretion over staff, resources and functions, will be replaced by a new system that makes urban managers clearly subordinate to county executives, including being dependant on county governments for financing. 173 Chapter 14: Financing and management of urban areas 14.8 Kenya’s existing system of local government has existed since before independence. The 175 local authorities created under the Local Government Act, govern both rural (county councils) and urban areas (city, municipal and town councils). City, municipal and town councils provide urban services and in a few cases, a limited range of social services. The Constitution provides for these functions to be assumed by county governments (see Box 14-2). Although the system of local authorities is not abolished by the Constitution, the national government’s powers to legislate for urban government are limited (see Box 14-1). Box 14-1: Constitutional provisions on urban government. The Constitution provides for a national law on cities and urban areas. Article 184 envisages that the national law will cover the classification of urban areas, principles for their governance and management, and participation of citizens. However, the Fourth Schedule of the Constitution has assigned virtually all urban service delivery functions to county governments, as well as the main sources of revenue that are currently used to finance urban services. In light of this, it is unclear how urban governments would be empowered (given functional responsibilities), and how those would be financed. Interpretation 1: The national government’s powers allow it to establish urban governments. To be meaningful, this power must include assignment of powers, functions, revenues and funding to those governments. The first interpretation would argue that the Constitution clearly intends the national government to establish a system of governance for urban areas. It could be argued that: - it makes no sense for the national government to be able to establish a system of urban government, but not give it functional responsibilities. - the capacity of the national government to assign urban functions to urban governments is supported by Article 186(4), which provides that the national government may legislate on any matter. Interpretation 2: The national government’s powers extend only to establishing the framework for urban governance, but county governments must choose to establish urban governments, empower and resource them. The second interpretation would argue that the national government’s power in relation to urban and city government is limited to establishing a framework of governance, criteria for classification, and determining how citizens should be represented. The actual establishment of individual urban governments would be a county decision, but counties could be required to establish urban governments in circumstances where certain conditions are met. In addition, the national government could delegate further revenue-raising powers to urban governments that meet the requirements of the national model. This could provide an incentive to counties to establish conforming urban governments, by giving them access to additional revenue sources. The Urban Areas and Cities Act is framed on the basis of a compromise between these two interpretations. City and municipal boards and town committees are established under the Act, although county executive committees will make individual board appointments. The management of cities and municipalities is formally vested in the county government (Section 12). The Act recognizes a principal-agent relationship between the county government and the board of an urban area or city, and specifies that the boards are financially accountable to the county government (Section 11). Although county governments will delegate most powers to city and municipal boards, the Act itself also delegates some important powers. These include integrated development planning, land use control and development, and facilitation and regulation of public transport (Section 20). Source: World Bank staff analysis of Kenya Constitution. 174 Devolution without disruption – Pathways to a successful new Kenya Box 14-2: Urban functions assigned to county governments under Kenya’s Constitution. The Fourth Schedule of the Constitution assigns powers and functions to national and county governments respectively. The functions assigned to county governments in Part 2 of the Schedule include: • Refuse removal, refuse dumps and solid waste disposal. • Licensing and control of undertakings that sell food to the public. • Control of outdoor advertising and public nuisances. • Libraries, museums, cinemas, video shows and hiring, and sports and cultural facilities. • County parks, beaches and recreation facilities. • Street lighting, traffic and parking. • Public road transport. • Animal control and welfare including dog licensing. • Regulation of markets. • Planning and development including land survey and mapping, boundaries and fencing. • Housing. • Electricity and gas reticulation and energy reticulation. • Storm water management systems in built-up areas. • Water and sanitation services. • Firefighting. Source: World Bank staff analysis of Kenya Constitution. 14.9 In the new dispensation, Kenya’s cities will be managed by appointed boards. Kenyan citizens’ poor opinion of their local authorities has contributed to a strong push for non-elected management of urban areas. A 2010 survey found most citizens are skeptical of local governments.¹ More than 60 percent of respondents considered their local government as dong a bad job of their core functions―maintaining roads and market places, removing rubbish and maintaining health standards in restaurants. Despite procedures that require citizen involvement in planning, 80 percent of respondents considered the councils as being particularly poor at involving local people in their decisions. The only area where councils were considered to do a reasonable job, was in the collection of license fees and property rates. 14.10 The UACA establishes a system of urban management without elected representation. The new urban management bodies will have far less power and autonomy, than the local authorities they replace. Most board functions will be delegated by county governments; they have no guaranteed funding, and they will be appointed at the discretion of the county executive committee (see Table 14-1). Comparing Kenya with international experience 14.11 One of the “twelve golden rules of fiscal decentralization� is that urban and rural areas should not be treated in the same way. Many systems of local government distinguish between urban local governments, and rural local governments. Indian rural local governments (panchyats) are relatively disempowered compared with the larger cities, and in the Philippines, cities and municipalities have special status and access to substantial own sources of revenue. In Vietnam, urban areas have more independence from higher levels of government, than other subnational tiers, and they have local ¹ Mitullah (2010). 175 Table 14-1: Comparison of cities, municipalities and towns under the Urban Areas and Cities Act. Classification Population threshold Governance Powers and functions Staffing Finance City Over 500,000 ▪ Status conferred by ▪ Body corporate with ▪ Manager appointed by ▪ Budget presented to President on resolution of power to sue and be sued, county Public Service Board county treasury for Senate (s.7). acquire property, enter into (s.28). inclusion in County ▪ Board of 11 part-time contracts, borrow money ▪ Other staff determined by Appropriation Act (s.20). members appointed and make investments county public service (s.12). ▪ Funds of a board consist by county executive (s.12). of transfers from county committee, 6 of whom ▪ Functions include assembly, money accruing appointed through developing integrated the board in the exercise competitive process, 5 development plan, control of its functions, grants and nominated by stakeholder and sub-division of land, donations. (s.43). bodies (s.13). monitor and regulate ▪ Requirement to produce ▪ Chair and vice-chair services provided by other budget estimates to elected by members (s.17). service providers, facilitate be approved by county and regulate public assembly (s.45). Municipality Over 250,000 ▪ Status conferred by transport (s.20). ▪ Audited accounts to be governor on resolution of ▪ Make by-laws (s.21) . submitted to county county assembly (s.9). ▪ Other functions and executive committee (s.46). ▪ Board of 9 part-time executive powers delegated members appointed by county government, by county executive including collection of rates committee, 5 of whom (ss 20, 21). appointed through ▪ Deliver services including competitive process, 4 through partnerships, joint nominated by stakeholder venture (31-35). bodies (s.13). ▪ Prepare integrated ▪ Chair and vice-chair development plan and elected by members (s.17). submit to county executive for approval (ss 39-42). Town Over 10,000 ▪ Status conferred by ▪ Not a body corporate ▪ Administrator appointed in ▪ Requirement to produce governor in consultation (s.31). same way as city/municipal budget estimates to with town committee ▪ Deliver services as manager(s.31). be approved by county (s.10). for municipality ‘with assembly (s.45). necessary modifications’ ▪ Audited accounts to be (ss 31-35). submitted to county ▪ Provisions on “Functions executive committee (s.46). of a board� (s.20) do not apply. Prepare integrated development plan and submit to county executive for approval (ss 39-42). Source: World Bank staff analysis. Chapter 14: Financing and management of urban areas 176 Devolution without disruption – Pathways to a successful new Kenya elections. Even in countries where the powers and functions of local government make no distinction between urban and rural jurisdictions, the jurisdictional boundaries incorporate rural and urban areas distinctly. Indonesian cities (kotamadya) are separate jurisdictions to rural local governments (kabupaten). In Africa, clear distinctions between urban and rural local governance arrangements are the norm. 14.12 This urban-rural divide is reflected in the capacity and performance of the respective local governments. Urban local governments tend to have higher tax collection capacity, because their tax base is more affluent, endowing them with greater financial resources. They have correspondingly higher costs because they are often responsible for a greater range and scale of infrastructure that is more expensive to construct and maintain. They have a greater need for capital for investment in new infrastructure, and use more sophisticated systems to manage their assets. Higher levels of education may lead to greater effectiveness of participatory mechanisms, and bottom-up accountability in urban local governments. 14.13 Although Kenya’s county governments are elected Table 14-2: Average size of African local governments. and have some fiscal autonomy, they are not a Country Average pop. substitute for local governments. To begin with, Algeria 21,000 all but two county governments include a mix of rural and urban areas. At an average population Benin 120,000 of 810,000, they are also more than six times the Bukina Faso 43,000 average size of local governments in other countries Cote d'Ivoire 20,000 (see Table 14-2). Whereas the main aim of the Egypt 15,000 Constitution is decentralization, consolidation from Ethiopia 153,000 175 Local Authorities (LAs) to forty-seven county Gabon 29,000 governments actually means that urban service Ghana 169,000 delivery responsibilities have been recentralized. Kenya Local Authorities 218,000 The effectiveness of urban management will thus Kenya County Governments 820,000 depend to a large extent on county governments Mali 18,000 decisions, concerning delegation of powers and revenues to urban boards, and how well they hold Morocco 21,000 them accountable. Mauritania 14,000 Mozambique 477,000 14.14 In large federal countries urban management Namibia 33,000 has traditionally been left to second-tier Niger 55,000 governments. Kenya’s Constitution takes a similar Nigeria 115,000 approach. Yet there are important differences. In Rwanda 310,000 large federal countries the state governments have Senegal 29,000 large populations and well-established systems of South Africa 172,000 public administration. Many of Kenya’s counties are small in size, and are likely to experience Togo 19,000 capacity weaknesses, particularly as they are being Tunisia 39,000 established. Most importantly, almost all Kenya’s Uganda 30,000 counties will be fiscally weak―that is, they will be Zambia 35,000 heavily dependent on the national government for Zimbabwe 129,000 transfers. Source: UCLGA (2010). 177 Chapter 14: Financing and management of urban areas 14.15 Even in federal countries, there is a growing trend for national governments to play a strong role in governance and financing of urban areas. This reflects an increased recognition of the importance of urban management for national economic development. In Mexico and India, for example, successive reforms have made local governments increasingly connected, financially and in other ways, to national rather than state or provincial government. In Australia, the Commonwealth Government pays financial assistance grants directly to local governments. The Commonwealth has recently released a national policy on the development of urban areas, as well as a raft of budgetary measures, to improve urban infrastructure, notwithstanding the lack of any constitutional mandate to do so. The policy includes a set of criteria for urban planning, against which the strategic plans of major urban areas will be periodically reviewed. Gaps in The Urban Areas and Cities Act 14.16 The main difference between the Kenyan model of urban governance and that of most other countries, is the absence of democratic representation on the boards and committees of urban areas. However, there are a number of other issues of concern. These issues may mean that it is very challenging to put the right accountability arrangements in place, and to ensure that urban service delivery is adequately provided. 14.17 The powers and accountabilities of city/municipal boards and those of town committees will be very different. The UACA provides for corporate (city or municipal) boards to manage the largest urban areas, while town committees are provided for the remaining urban areas, many of which are still quite large. It is not entirely clear how town committees will function. They are clearly deemed not to be corporate bodies. There is no provision for them to have staff, other than an administrator. They are not assigned executive functions, and there is no provision for the county to assign functions to them. The UACA does not deal with town committee funding (although it does include provisions requiring them to prepare plans and budgets). On one interpretation, their lack of corporate status means they will function in an advisory capacity only, with the administrative aspects of urban management effectively integrated into the county administration. See Table 14-1 for a more detailed exposition of the differences between boards and committees. 14.18 The distinction between boards and committees is significant because only three of the existing local authorities qualify to become city or municipal boards. The Urban Areas and Cities Act sets the threshold for a city at 500,000, and for a municipality at 250,000. Only cities and municipalities are entitled to boards; the remaining urban areas are classified as towns, and the management arrangements are limited to town committees. There are five urban areas in Kenya with a population of more than 250,000. Two urban centres, Nairobi and Mombasa, qualify as cities on the basis of their population. A third, Kisumu is deemed to be a city under the Act. The Act also prescribes that “city-counties� (defined as “a city which is also a county�) should be governed as county governments. Nairobi is deemed to be the capital city, and is to be governed and managed in the same manner as a county government. Since the 2009 census defines Mombasa county as having a population that is 100 percent urban, it is assumed that that neither Nairobi nor Mombasa will have city boards. 14.19 Only three urban centres qualify to have have municipal or city boards. Kisumu is Kenya’s third largest urban area and will have a city board. The only other urban areas with populations of more than 250,000 are Nakuru and Eldoret. The remaining urban centres with populations less than 250,000 qualify only for town status. In these urban centres, unless the Act is amended, the rather uncertain 178 Devolution without disruption – Pathways to a successful new Kenya management arrangements for town committees Table 14-3: Twenty-one urban centres with more than will apply. A number of substantial urban centres 80,000 residents will not have municipal boards. with viable existing local governments, including Urban Centre Population 21 urban centres each with more than 80,000 Ruiru 240,000 residents (see Table 14-3) may thus be effectively Kikuyu 230,000 recentralised into the county administrations, as a Kangundo-Tala 220,000 result of the operation of the UACA. Naivasha 170,000 14.20 There are three ways in which the management Machakos 150,000 autonomy of urban areas could be expanded. Mavoko 137,000 One is by lowering the threshold for the definition Thika 135,000 of municipalities. However, since the threshold Vihiga 120,000 was raised (from the original 75,000 proposed by Nyeri 120,000 the Task Force on Devolved Government)(TFDG) Malindi 120,000 during Parliamentary consideration of the bill, this Ngong 110,000 seems an unlikely option. Alternatively, the Act Kitui 109,000 could deem all county capitals to be municipalities Karuri 107,000 in the same way that Kisumu is deemed to be a city. This would mean that all county capitals Mumias 100,000 would have municipal boards to manage them. Kitale 100,000 This would still leave important towns like Ruiru, Kericho 100,000 Naivasha, Ngong, Thika, and Mumias managed as Kimilili 95,000 towns. A third alternative would be to enhance Awasi 93,000 the managerial autonomy of town committees, to Kakamega 90,000 make them function in the same way as municipal Kiambu 85,000 boards. Kisii 80,000 Source: Government of Kenya (2009). Delegation of functions and responsibilities 14.21 To be accountable, city and municipal boards need clear responsibilities, adequate resources to carry them out, and appropriate and transparent reporting. The UACA assigns some specific functions to city and municipal boards (see Box 14-3), but it leaves it to county governments to delegate both executive authority, as well as the power to levy charges and collect other revenues. There are no provisions on powers, functions or delegation of revenues for town committees. The provisions on service delivery apply to town committees “with the necessary modifications�. 14.22 There is no clear process for counties to delegate functions to municipal or city boards. Whereas the TTDG Act 2012 provides a clear process for the national government to transfer functions to county governments, there is no requirement for transparency, concerning the delegations given by county governments to city and municipal boards. This potentially undermines the accountability of both urban boards and county governments. Lack of transparency around delegations to urban boards, also undermines the principle that funding should follow function. If it is not clear what functions the urban boards are expected to carry out, it will be difficult to make an assessment, as to whether they are adequately funded, or whether or not they are performing. 179 Chapter 14: Financing and management of urban areas Box 14-3: Urban Areas and Cities Act provisions on powers and functions of city and municipal boards. Functions (Section 20): • Oversee affairs of the city or municipality. • Formulate an integrated development plan. • Control land use and development, within framework of spatial plans as delegated by the county government. • Promote infrastructure development and services as delegated by county government. • Maintain an information system. • Monitor and regulate city and municipal services provided by service providers other than the city or municipal board. • Prepare budget for approval by county executive committee. • Collect rates, taxes, levies, fees and surcharges as delegated by county government. • Facilitate and regulate public transport. • Such other functions as may be delegated by the county government. Powers (Section 21): • Exercise executive authority as delegated by the county government. • Ensure provision of services to residents. • Impose fees and charges as authorized by the county government for provision of services. • Make bye-laws. • Such other powers as are delegated by the county executive committee. Service Delivery (Sections 31-33): • Deliver services on behalf of the county government, as specified in national or county laws. • Establish service delivery entities with the approval of the county executive committee to carry out its functions. • Enter into a partnership with a utility for the provision of social infrastructure services, in consultation with the governor and with the approval of the county assembly. • Contract a private entity to deliver a service with the approval of the county assembly. Source: World Bank staff analysis. 14.23 If county governments were required to publish the functions and revenue streams assigned to urban boards, this would strengthen the accountability framework for urban service delivery. Guidelines on optimal function assignment, as well as a recommended approach to gazetting delegations to urban boards, as well as publishing them in the newspaper, would make it easier for citizens to understand what they are entitled to receive as services, and from which level of authority. Financing Urban Areas 14.24 Kenya’s new devolution arrangements make county governments responsible for financing urban service delivery. For over sixty years, Kenya’s local governments have been relatively fiscally autonomous, raising taxes from their residents and using these to provide services. Transfers to local governments were introduced just over a decade ago. Once urban service functions are taken over by county governments, along with the tax bases that previously sustained municipal and town councils, the landscape of local government service financing will change profoundly. The urban boards and committees will depend on the county government for funding―either through transfers from the county budget, or delegation of power to collect revenues (either as charge for services or by collecting and retaining other county taxes). Since the new bodies are management bodies, not governments, they cannot impose taxes and charges, unless authorized by a national or county law. 180 Devolution without disruption – Pathways to a successful new Kenya 14.25 There is a risk that urban services may be inadequately funded under the new arrangements. Because rural residents will dominate most counties (see Figure 14-3), county governments may choose to preference rural services, and redistribute the revenues raised from urban residents away from recurrent costs of urban services and towards other purposes. This may jeopardize the economic development potential of Kenya’s urban areas. However, it should be noted, that this risk arises only in relation to those counties with predominantly rural populations. It can be assumed that the mainly urban populations of Nairobi, Mombasa and Kiambu will elect representatives whose preference is to ensure urban services are well provided. Figure 14-3: Despite rapid urbanization, most counties are still predominantly rural. 100 90 80 70 60 Percent 50 40 30 20 10 0 Tharaka Mombasa Nairobi Kiambu Kisumu Machakos Nakuru Isiolo Kajiado Uasin Gishu Migori Vihiga Kericho Laikipia Nyeri Garissa Taita Taveta Marsabit Bungoma Kisii Trans Nzoia Lamu Nyandarua Bomet Kwale Mandera Samburu Busia Murang'a Embu Kirinyaga Kakamega Tana River Wajir Elgeyo-Marakwet Homa Bay Turkana Nyamira Kitui Nandi Meru Makueni Baringo Siaya West Pokot Narok Kili Urban Rural Source: World Bank (2011c). 14.26 The main source of financing to county governments is an unconditional equitable share transfer. The county equitable share is guaranteed to be not less than 15 percent of national revenue. The national government can also make additional conditional grants from its share of revenue, provided there is fiscal space to do so. If the national government shifts more than 15 percent of its expenditure responsibilities to the county governments, while keeping the equitable share at―or close to―the Constitutional minimum, there will be scope for conditional transfers. 14.27 County governments are likely to raise most of their own revenue from property tax and single business permits. These are the two largest tax bases assigned to counties. Fiscal federalism theory suggests that at subnational level, revenues and expenditure responsibilities should be aligned―in other words, that local governments should finance service provision from revenues raised from the beneficiaries of those services. Under the Kenyan model of devolution, revenues raised primarily from businesses and urban residents, are at the disposal of county governments, which service wider communities. 14.28 Most of the revenues raised by county governments will come from urban areas. Some estimate of how much counties will generate from their urban centres can be made on the basis of the revenues collected by the existing urban authorities―the city, municipal and town councils. Table 14-4 shows 181 Chapter 14: Financing and management of urban areas Table 14-4: Shares of local revenues collected from rural and urban areas, 2008/09. Municipal Councils & Town Councils County County Councils (Incl. Nairobi City Council) Property rates SBP Total Property rates SBP Total Baringo 3,471,896 7,060,065 10,531,961 - 7,121,072 7,121,072 Bomet 8,857,396 7,294,305 16,151,701 395,000 9,472,250 9,867,250 Bungoma 17,022,412 21,900,083 38,922,495 - 22,007,631 22,007,631 Busia 6,232,147 17,030,166 23,262,313 - 6,816,009 6,816,009 Elgeyo Marakwet 685,430 3,123,050 3,808,480 1,829,708 5,032,390 6,862,098 Embu 6,589,049 18,139,291 24,728,340 2,253,255 19,224,717 21,477,972 Garissa - 5,574,857 5,574,857 - 2,131,404 2,131,404 Homa Bay 4,938,713 5,461,005 10,399,718 - 3,385,833 3,385,833 Isiolo - 1,588,837 1,548,208 3,137,045 Kajiado 1,952,076 3,062,698 5,014,774 24,062,054 35,152,990 59,215,044 Kakamega 21,164,865 17,368,264 38,533,129 2,397,988 19,501,173 21,899,161 Kericho 60,576,850 11,787,484 72,364,334 21,174,785 9,844,420 31,019,205 Kiambu 88,121,000 109,215,409 197,336,409 1,983,874 53,010,689 54,994,563 Kilifi 50,788,675 33,571,477 84,360,152 9,949,698 9,791,395 19,741,093 Kirinyaga 3,693,090 14,033,723 17,726,813 4,253,270 31,862,944 36,116,214 Kisii 12,097,290 23,192,110 35,289,400 - 8,200,925 8,200,925 Kisumu 78,124,827 51,941,564 130,066,391 2,351,893 9,050,707 11,402,600 Kitui 4,785,093 15,634,298 20,419,391 - 27,116,634 27,116,634 Kwale 663,174 1,352,599 2,015,773 51,476,016 11,741,703 63,217,719 Laikipia 32,598,728 21,626,030 54,224,758 4,003,281 12,771,573 16,774,854 Lamu - 8,604,545 3,941,634 12,546,179 Machakos 105,142,864 45,057,931 150,200,795 150,144 29,718,242 29,868,386 Makueni 311,278 7,021,997 7,333,275 167,288 21,571,126 21,738,414 Mandera - 1,717,960 1,717,960 - 728,470 728,470 Marsabit - - 6,086,948 6,086,948 Meru 10,609,106 31,105,949 41,715,055 716,288 41,943,152 42,659,440 Migori 1,037,983 17,935,747 18,973,730 188,350 4,152,000 4,340,350 Mombasa 457,182,230 263,133,595 720,315,825 - Murang'a 9,566,776 17,902,989 27,469,765 42,752 34,129,460 34,172,212 Nairobi 1,773,629,923 832,202,732 2,605,832,655 - Nakuru 100,895,597 119,980,615 220,876,212 13,030,292 34,875,179 47,905,471 Nandi 8,923,601 8,627,723 17,551,324 313,811 6,124,137 6,437,948 Narok - 5,853,910 5,853,910 - 3,709,645 3,709,645 Nyamira 475,782 4,257,110 4,732,892 - 4,330,326 4,330,326 Nyandarua 2,796,871 6,254,560 9,051,431 3,892,227 35,345,594 39,237,821 Nyeri 21,667,977 37,143,877 58,811,854 347,584 21,284,936 21,632,520 Source: World Bank based on LAFT annual report (2008/09). 182 Devolution without disruption – Pathways to a successful new Kenya Table 14-4: Shares of local revenues collected from rural and urban areas, 2008/09 (continued). Municipal Councils & Town Councils County County Councils (Incl. Nairobi City Council) Property rates SBP Total Property rates SBP Total Samburu - 3,556,924 3,556,924 - 1,570,895 1,570,895 Siaya 2,569,444 11,126,254 13,695,698 - 13,904,825 13,904,825 Taita Taveta 4,658,141 11,209,948 15,868,089 837,326 5,890,901 6,728,227 Tana River - - - 4,095,554 1,978,520 6,074,074 Tharaka-Nithi - - - - 2,947,066 2,947,066 Trans Nzoia 13,935,503 18,135,566 32,071,069 1,285,138 9,986,597 11,271,735 Turkana - 4,253,700 4,253,700 - 4,031,680 4,031,680 Uasin Gishu 128,951,632 47,233,890 176,185,522 2,043,334 8,972,026 11,015,360 Vihiga 1,395,163 8,200,461 9,595,624 53,279 6,762,202 6,815,481 Wajir - - - - 1,719,050 1,719,050 West Pokot 438,970 5,070,079 5,509,049 9,108,900 1,514,617 10,623,517 TOTAL 3,046,551,552 1,895,351,995 4,941,903,547 172,596,471 612,003,895 784,600,366 Source: World Bank based on LAFT annual report (2008/09). the amount of revenue collected by urban councils in form of property tax and single business permit fees in the 2008/09 year, compared to the amount collected by rural councils. 14.29 Urban service delivery will depend on the priority which county assemblies give these expenditures in their budgets. The UACA envisages that county governments will provide transfers to city and municipal boards, but provides no guidance as to how these amounts should be calculated. The PFM Act includes a set of criteria, to guide county governments in designing transfers, and requires them to seek recommendations of the CRA in arriving at a formula for within-county distribution. 14.30 The national government could help to ensure that urban services are adequately provided through earmarked urban service grants. Grants might be paid either to the county governments or to the bodies that manage them (urban boards). However, most county revenues are fungible, and can be allocated to any purpose. An earmarked grant might have the effect of allowing counties to reduce their own funding to urban services. County governments might abandon their own responsibility for taking care of urban services, and blame the national government for inadequate service delivery in urban areas. To ensure that counties maintain their own level of funding for urban services, transfers could include an additionality clause, which binds the counties to maintain a certain level of funding to urban services. A county that does not allocate an adequate amount, would not receive the earmarked grant. 14.31 For now, it seems likely that the national government will continue the Local Authorities Transfer Fund. LATF was the vehicle for financing the former local authorities. However, some changes will be needed. The LATF Act will need to be revised to make way for financing urban boards or committees that are constitutionally part of county governments. As clarified under the recent amendment to the UACA that were included in the PFM Act, urban board budgets form part of the county appropriation law. If all urban areas that qualify as towns are classified as such, there will be many more boards and 183 Chapter 14: Financing and management of urban areas committees than there are local authorities today. The national government should exercise caution not to try to use LATF grants to make urban boards report directly to the national government (as is the case now); as this might make county governments abandon responsibility for financing them properly. 14.32 Finally, transfers may still be insufficient to finance the type of infrastructure development that Kenya’s largest cities need. While counties may be able to borrow, this will be subject to national government review and guarantee. This guarantee is helpful to prevent the risk of uncontrolled borrowing at the subnational level, and of resulting macro-imbalances. However, the flipside may well be insufficient access to capital for infrastructure, particularly in Kenya’s larger cities. As the national government is already struggling to bring back the debt-to-GDP ratio to a target of 45 percent it may well be reluctant to encourage subnational borrowing, that will end up on its balance sheet. CHAPTER 14: KEY INSIGHTS / RECOMMENDATIONS. There is great uncertainty surrounding the fate of cities under the new dispensation, and Kenya may be a unique case in the world where municipal services are actually being re-centralized. Kenya’s cities will no longer have an elected leadership (instead appointed boards) and only a handful will have a corporate body to manage them. All of the revenues currently enjoyed by local authorities will become county own revenues, and urban service resourcing will be at the discretion of mostly rural counties. Specific arrangements will be needed to ensure that large cities are appropriately governed and it will also be key that urban functions are also appropriately resourced. The current legislation has important gaps including the role of town committees, or the process whereby counties might delegate functions to cities. Leaving urban services and governance un-attended would be dangerous, as urban areas are worldwide the main engines of growth in a country. *** More urban centers should have corporate bodies to manage them. This could be achieved either by lowering the threshold for ‘municipality’ status, or by amending the powers and functions of towns under the current law. Functions of city and municipal boards should be clarified as well as the formal process for counties to delegate additional functions to them. Set standards for urban service delivery and require boards to report against these to the county assembly. City and municipal boards should be given own revenue powers, through delegation of additional national taxing power. Establish cost benchmarks to monitor future resourcing of urban functions. Calculate current cost of urban services, based on past spending, and monitor to ensure that funding is adequate to at least maintain resourcing of services at current levels. Extend the role and functions of local authorities during the transition, until their staff, functions and assets can be accommodated by the new counties. The national government could help to ensure that urban services are adequately provided through earmarked urban service grants. To ensure that counties maintain their own level of funding for urban services, transfers could include an additionality clause, which binds the counties to maintain a certain level of funding to urban services. 184 CHAP TE R FIFTE E N Decentralising public services 15.1 Public service transition is a crucial dimension of decentralization, but the challenge of implementing such a transition has been underemphasized. Decentralisation disperses power both geographically and institutionally. It creates new responsibilities for inexperienced actors, and diffuses scarce skill sets across a larger institutional landscape. At the same time, it offers the potential for public servants to be more accountable to the citizens they serve, more amenable to being monitored by the politicians who direct them, and gives them greater access to the processes that decide priorities in policy and budgetary allocation. Realizing major changes in the way the public service operates is difficult in any context, but all the more so, when the scale and scope of the change are as broad as will be the case in Kenya. 15.2 The outcomes of devolution depend on how effectively public servants are motivated, directed, supervised, supported and held accountable. The vision of devolution will be translated into reality through the work of public servants in county government administrations. Many of them will be former national public servants. If devolution helps them perform to higher standards, public service delivery will improve. Conversely, the public service can also be a major source of inefficiency and ineffectiveness. Whether devolution makes public servants more responsive and accountable, or more inefficient and ineffective, depends very much on the framework for decentralized public service management. Clarifying exactly what the future hold for public servants who currently work in counties, is crucial. Anxiety among government staff almost always translates into demotivation and lack of productivity. Existing administrative arrangements 15.3 Kenya’s existing public service structure is currently highly centralized, so devolution will bring about major changes. Forty-seven new county civil services will be established, creating two immediate sets of challenges: firstly, to define the overall governance framework for country civil service, and secondly, to manage the HR transition implied by their establishment. This chapter will discuss the first of these challenges. The second one will be discussed in the subsequent chapter on transition. 15.4 Services are currently delivered via deconcentrated structures with limited local autonomy. Most line ministries have representation in many, or all of the 288 districts and manage staff and funding through organizational silos. Financing for services from the line ministry budgets is channeled through district treasuries operated by the Ministry of Finance. Almost no human resource functions are carried out at district level, and a very small proportion of ministry budgets (estimated at between 2-5 percent) are managed there. Officials of the ministries of Provincial Administration and Planning play coordinating roles. Even in Mumias district, which became a separate district only five years ago, no less than twenty ministries are represented. The informal network of government actors at district level also includes the staff of Constituency Development Fund (CDF)1 committees, (nominally staff of the Planning Ministry)who are responsible for overseeing project spending worth over KES 20 billion 1 The CDF is a grant program for small infrastructure grants allocated to projects chosen by a committee chaired by the member of Parliament for that constituency. 186 Devolution without disruption – Pathways to a successful new Kenya annually, and the staff of elected municipal, town and county councils. Around 90 percent of local authority staff are not civil servants, and are recruited and managed locally under delegation from the Public Service Commission. The negative impact of these vertically siloed arrangements on service delivery at the local level has been much commented upon in the past.2 15.5 In parallel to sectoral ministries, the central administration is also represented locally by provincial and district administration; staff of the Ministry of Internal Security and Provincial Administration, reporting to the office of the President. Provincial and District Commissioners are political appointees. They exercise control over a section of the police force known as the ‘Administration Police’, but do not control staff of line ministries working in districts. However, despite their lack of direct control, they wield considerable defacto authority because of their role in maintenance of internal security. Below the level of districts and sub-districts the system of provincial administration includes appointed officials and chiefs, who play an important role in informal dispute resolution and maintenance of order. Over 12,000 staff are employed in the different cadres of the provincial administration structure,3 and it is not yet clear how they will fit within the new county government arrangements. The Constitution mandates that the structure of provincial administration be reviewed to align it with devolution within five years, but does not require its abolition.4 All indications at this stage are that the system will remain in place for the time being.5 Although the system of provincial administration was widely criticized in the past as an instrument of repressive central governments,6 during the country-wide consultations conducted by the Task Force on Devolved Government (TFDG), the importance of the role currently played by chiefs and other officials of the district administration was frequently expressed.7 15.6 The new devolution arrangements offer the opportunity to rationalize these highly fragmented administrative arrangements, improve coordination, and rationalize service delivery gaps and overlaps. However, some of the civil servants at the district level will remain attached to national functions, and so there remains the challenge of redesigning local administrative arrangements to facilitate continued coordination between national and county levels of government. Legal framework for managing county public services 15.7 Kenya’s new counties will have considerable autonomy, including over public service management. The Constitution provides for county governments to be responsible, within a framework of uniform norms and standards, for establishing and abolishing offices, appointing public servants, and exercising disciplinary control over them.8 The TFDG considered that this role was best achieved by having forty seven separate public services. A range of new or reformed institutions will feature in the landscape for the management of devolved public services (see Box 15-1). Key among them will be the forty seven new county Public Service Boards, which will exercise the power to create and abolish offices, fill them, and discipline their occupants, on behalf of county governments. Under amendments introduced in Parliament, the CGA now requires each county to have a separate assembly services board, chaired by the speaker, responsible for managing staff of the county assembly. 2 World Bank (2002), JICA (2007), Transparency International (2011) 3 Task Force on Devolved Government (2011). 4 Constitution of Kenya, Sixth Schedule, Section 17. 5 In early 2012 the President refused to assent to the County Government Bill because of provisions requiring provincial and district administration staff to report to county government, and for Governors to chair the county security committee. Amendments were introduced to the Bill in June and it was passed without these provisions and later received Presidential assent. 6 Branch and Cheesman (2007). 7 This may have been partly due to the fact that district administration staff organized community consultations in most districts and were well represented at the Task Force hearings. 8 Constitution of Kenya, Article 235. 187 Chapter 15: Decentralising public services Box 15-1: New institutions and new roles. Previously responsible for all public servants, the Public Service Commission is reformed and constitutionally entrenched. But its role in relation to county public servants will be limited to hearing appeals from decisions of county Public Service Boards (Constitution Article 234). Forty-seven county Public Service Boards will be established under the County Governments Act. Board members will be appointed by the governor of each county with the approval of the county assembly, and enjoy the same protection from political dismissal as the members of a constitutional commission do, under the Constitution. The role of a board is to establish and abolish offices, appoint public servants, and exercise discipline over them (County Governments Act, sections 57-59). Forty-seven county Assembly Service Boards will be established under the County Governments Act. The Boards have four members, one of whom is from outside the assembly. They are responsible for running the assembly, establishing offices and recruiting staff, and preparing the budget for the county assembly (County Governments Act, section 12) A constitutionally mandated Salaries and Remuneration Commission (SRC) has replaced the former Permanent Public Service Remuneration Review Board established in 2003. The role of the SRC is to set and review the salaries of state officers, and to advise the national and county governments, on the remuneration and benefits of all other public officers (Constitution Article 230(5)). State officers are defined to include the President, members of the executive, governors and members of county executive committees, members of Parliament and county assemblies, judges and magistrates, members of constitutional commissions, and holders of various offices established under the Constitution. For all other public officers, the SRC will not determine salary levels, but will merely advise on them (Constitution Article 260). A Transition Authority has been established under the Transition to Devolved Government Act. Its functions include carrying out a human resource audit of government and local authorities, assessing capacity needs, and recommending measures to ensure county governments have adequate capacity during transition (Transition to Devolved Government Act Section 7(2)(h)(iii), (j) and (k)). County governments are also explicitly subject to the values and principles of the public service set out in the Constitution. These include professional ethics, transparency, accountability, responsive and impartial delivery of services, efficient uses of resources, and equal opportunity of employment (Constitution Articles 10 and 232). These principles need to be operationalized. For national public servants, the operational application of public service values and principles is under the Public Officers Ethics Act. It is not clear if this will apply to county public servants or not. There is no mention of these principles and values in the County Public Service Act. Source: World Bank staff analysis. 15.8 The CGA provides part of the legal framework for the management of county public servants. It includes provision for various aspects of transition (including secondment of national public servants to counties) and it also provides the regulatory framework for counties to engage their own public servants. It is envisaged that the human resource needs of counties will be addressed initially through secondment, after which county governments will recruit their own staff. These issues are further 188 Devolution without disruption – Pathways to a successful new Kenya discussed in the next chapter. 15.9 However, the CGA leaves a number of important gaps in the policy framework (see Box 15-3). Presently, Kenya’s laws provide no answers to the following five crucial questions regarding county public servants: i) Who sets county public servants’ salary and other benefits? The SRC is mandated to provide advice under Article 230(4) of the Constitution, but the county government is not obliged to accept it. ii) What discipline procedures are county public servants subject to? The CGA does not deal with discipline procedures. iii) Who sets the grading of different positions in the county civil services? Grading is not covered by the County Governments Act. iv) How is the code of ethics applied to county public servants? v) Who decides the organizational structure of the county administration? This could be an extension of the county Public Service Board’s power to create offices in the county public service, but is it also responsible for deciding organizational structure―or is this the role of the county executive committee? 15.10 It is important that these regulatory gaps are filled before counties begin to employ their own staff. Article 235 of the Constitution provides clear authority for the national government to provide a framework of national standards to be set out in legislation. Whereas employment of individuals in private companies is governed by employment contracts or the common law, the employment terms and conditions of civil servants are set out in laws, regulations and codes. Kenya’s current civil service regulatory framework is mainly in regulations. In contrast, the framework of uniform norms and standards for county public servants envisaged by Article 235 must be in an Act of Parliament. Since the CGA provides for automatic secondment of public servants as soon as the county governments are established, that regulatory framework needs to be presented to Parliament before it rises at the end of 2012. 15.11 The first question is how much national government should regulate, and how much should be left to counties. Even if it is decided to leave some matters to counties to legislate in the long run, some interim provisions will be needed to provide an adequate framework before counties develop the capacity to make their own laws. The Transition Authority could play a role in proposing an approach to government, in conjunction with the Ministry of State for Public Service. 15.12 The second question to answer is which national agency should fill these gaps. It is not clear who is responsible for national policy on employment of county public servants. National public service laws will not apply to county public servants unless they are expressed to do so in a national law. The CGA has covered some aspects of county public service regulations, but did not mention important ones. The Task Force on Devolved Government proposed the establishment of a County Public Service Advisory Authority that could have taken on this role, but it was decided not to include this body in the final version of the law when it was presented to Parliament. The question of what national authority should fill these gaps has been left unanswered in the CGA. 15.13 The missing elements of the regulatory framework could be provided in regulations under the CGA, or they could be in separate legislation under the auspices of the Ministry of State for Public Service. This policy choice is about more than just legal questions. If the first approach is adopted, two ministries 189 Chapter 15: Decentralising public services will regulate public service matters, at different levels of the government. The Minister9 responsible for intergovernmental relations will regulate county public services, while the ministry responsible for the public service will regulate the national public service. If the second approach is adopted instead, then the regulation of county public services will be partly under the Ministry of Intergovernmental Relations (by virtue of the provisions in the CGA) and partly under the Minister for Public Service. This poses a risk of overlap and contradiction. Neither scenario is ideal. International experience suggests that generalist ministries responsible for lower levels of government, like ministries of home affairs, provincial affairs and the like, do not have the technical capacity to regulate public service matters effectively. Balancing central control and local autonomy 15.14 Devolving public service management involves balancing the risks of too much central control, against the risks of too much local autonomy. Like other areas of design of devolution arrangements, design of public service management arrangements involves trade-offs. The risk of too much centralization is that county governments will not have sufficient control of their staff to deliver services the way their citizens prefer, and to be fully accountable for their performance. But if there is not enough central oversight local controls could be ineffective, undermining accountability and service delivery. This is one of the paradoxes of decentralization―that effective devolution actually requires a strong central government.10 Ineffective local controls generally manifest in three different types of risks being realized. Elite capture and politicisation 15.15 The capture of local public services by local elites is a common dysfunction of decentralization. Weaker systems at the subnational level create more opportunities for public sector recruitment and promotion, and makes the process to be non-meritorious and less transparent (see the example of Indonesia in Box 15-2). Box 15-2: Indonesia: Worsening corruption in Human Resource Management after decentralization. Patronage was a defining feature of the highly centralized Indonesian bureaucratic state presided over by President Soeharto for 31 years. Patronage was not a new phenomenon, and had adapted to different regimes, since well before colonisation. Contrary to some of the predictions of theory, after Indonesia’s “big bang� decentralization in 1999 these networks strengthened rather than dissipated. As control of resources was shifted to regional governments, factions at the centre ‘colonised local legislatures and administrations.’ Since levels of accountability, transparency and efficiency were generally lower at the local level, opposition to local corruption was weaker and more diffuse. The fact that corruption in personnel practices had increased after decentralization was evidenced by a sharp rise in the prices paid for civil service jobs. The benign conditions for fostering corruption were compounded by poor human resource practices, including: little or no formal human resource planning; non-transparent appointments; widespread engagement of casual staff; inadequate examinations for entry to the public service; and appointments based on ethnicity and kinship. A recent study found regulatory provisions designed to ensure merit-based recruitment are reported to be merely decorative, and that it cost between US$ 5,500 to US$ 8,000 for a diploma nurse to secure a job in a desirable location. Source: Blunt et al. (2012). 9 The structure of executive government will change after the 2012 election. Whereas ministries are now headed by Ministers who are chosen from among the members of the National Assembly, after the election ministries will be called departments, and the President will choose Cabinet Secretaries who are not members of Parliament to head them. 10 World Bank (2000). 190 Devolution without disruption – Pathways to a successful new Kenya 15.16 Establishing independent county Public Service Boards has the potential to address some of these problems, but there is a risk they will have too much power. Recruitment should be sensitive to local needs, and reflect local preferences, but this also carries a risk of elite capture. Empowering independent county Public Service Boards is intended to ensure depoliticisation, but this may either open up different avenues for corruption, or alternatively come at the cost of managerial efficiency. The CGA provides for the county Public Service Boards to be all-powerful in the appointment and management of county public servants. The concentration of human resource management authority in the hands of county Public Service Boards creates potential for fiscally unsustainable recruitment, and excludes direct supervisors and other stakeholders from an appropriate role in recruitment (see Box 15-4). 15.17 Politicisation is also a potential risk of county Public Service Boards, because they are political appointees. Uganda’s Constitution provides for each district to have its own district service commission (DSC), much like Kenya’s county Public Service Boards. DSC members are appointed by politicians, which led to a situation whereby DSCs were used to influence public service recruitment for political needs. Where independently appointed public servants resisted attempts by politicians to engage in corrupt conduct, or to interfere in procurement, local leaders sometimes demanded to have these officers dismissed.11 The integrity of Kenya’s county Public Service Boards will depend on the quality of the process used to recruit their members. Members of county Public Service Boards are required to have university degrees, not to have breached a professional code of conduct, and to satisfy the leadership and integrity provisions of the Constitution. However, they are appointed by the governor with approval of the county assembly. Although the appointment process is required to be ‘competitive,’ this is not specified. There is no requirement for members to have any expertise in public sector human resource management. The independence and competence of county Public Service Board members will depend on how committed individual governors are in selecting those: (i) with appropriate skills and experience; and, (ii) who are committed to check the political interests of the governor in the way they exercise their powers. 15.18 Matching sub county structures to political boundaries could exacerbate the risk of a different kind of political capture. The CGA specifies that sub county administrative structures should align with the boundaries of constituencies, and structures below the sub county should align to the boundaries of wards.12 Constituencies are the electorates for members of the National Assembly, while wards are the electorates for members of County Assemblies. If the Constituency Development Fund (CDF) is retained―as seems likely from the 2012/13 Budget Policy Statement―the alignment of constituencies and sub-counties may carry the unintended risk of increasing the allegiances between sub county administrators and local politicians, although with a focus on national parliamentarians rather than county ones. National MPs may see the county governor as a competitor rather than a collaborator. If the MPs are empowered with resources (assuming CDF continues and is controlled by national parliamentarians), sub county administrators may feel more allegiance to national MPs (since they are the source of most discretionary funding available to spend in their area) than they do to the county government. In Papua New Guinea, the alignment of sub-provincial (district) administrative jurisdictions with the political boundaries of open electorates generated alliances between national members of parliament and district administrators who undermined the authority of the provincial governments in areas of their functional responsibilities. Since MPs also had available considerable 11 Smoke et al. (2010). 12 County Governments Act, Clause 48. 191 Chapter 15: Decentralising public services Box 15-3: Key gaps in the regulatory framework for devolved civil service management. The Public Service Revised Code of Regulation sets out the terms and conditions of employment of Kenya’s national civil servants. In addition, it covers the conduct of staff appraisals, as well as the procedures for appointment and dismissal of public servants, and for wage negotiations with the Kenya Union of Civil Servants. When county governments come into effect, this regulation will no longer apply to county civil servants, unless a national law prescribes its continued application to county governments. A regulatory framework for county public services should include: Rules governing organizational structures, functions and staff responsibilities―principles, establishment of departments, responsibilities of different authorities and senior staff, including heads of departments. It should be clear how departments are established, and how their functional responsibilities are specified. This could also provide a framework for assigning specific functions to members of county executive committees (in effect, county cabinet secretaries). None of these topics is covered in the County Governments Act. Governance—including application of the code of conduct to county public servants, how values and principles will be operationalized, and links to other governance bodies like the Anti-Corruption Commission. Management framework—Rules governing management practices―creation of offices and schemes of service, assignment of duties, redeployment, delegation, discipline, appeals and management improvement. There should be specific provisions relating to officer performance management, including for departmental heads. It should clarify whether there is to be a senior executive service performing the most important management tasks, with special conditions of employment and performance expectations. Departmental heads would normally have special appointment procedures and special duties, and reporting obligations. There should also be provisions for organizational performance management, that clarify who is responsible for monitoring and improving organisational performance, and what mechanisms or arrangements support this. There should be a specific authority for the body responsible for organizational performance management to conduct performance reviews, either on its own motion or at the request of the governor or a member of a county executive committee. The County Governments Act covers some of these matters, but is silent on many of them. It is not clear who has authority to make administrative orders, fix grades, assign grades to individual jobs, set schemes of service and determine the qualifications for employment at each level. There is no provision for officer performance management, and the provisions on organizational performance lack mechanisms to operationalize them. There is no provision for a senior service, nor specification of the special role and duties of departmental heads. It is also unclear who the formal employer of county public servants will be, who they will be accountable to, and who will determine their base salary and other allowances. It is not clear who will be responsible for dealing with employment claims, either about pay or conditions. Special provisions—applying to specialist officers like health workers, scientists, lawyers and teachers (county governments will presumably employ some nursery school teachers). It should also be clear the extent to which these provisions apply to quasi-government bodies, like the municipal boards established for urban areas. It is not clear whether a national body will approve special provisions relating to specialist public servants, or individual counties will set these themselves. It is almost certain that the capacity of the forty-seven county Public Service Boards will be shallow in the early years, and if counties are left to regulate these matters themselves, it is likely that the result will be highly fragmented rules and processes at county level. An obvious example is discipline procedures, which are not specified in the County Governments Act, leaving each county to develop its own discipline procedures would leave the Public Service Commission adjudicating appeals from forty-seven different sets of procedures. Source: World Bank staff analysis. 192 Devolution without disruption – Pathways to a successful new Kenya Box 15-4: Potential problems with concentration of power in the hands of county Public Service Boards. • Fiscal sustainability. If agents of a county government can make decisions with fiscal implications (like appointments and promotions), the county treasurer or executive committee member for finance should have the power to veto the decision, if it is not fiscally sustainable. The County Governments Act allows county department heads to apply directly to the county Public Service Board, for additional positions (Section 60(3)). • Supervisors should have a role in recruitment. Decisions about recruitment, promotion, transfers and discipline should include a voice of the direct managers of the affected officer. This is typically accomplished through a three- stage recruitment and selection process, in which the direct supervisor has a role in the final selection, if not also in the short-listing. The County Governments Act makes the county Public Service Boards almost solely responsible for these decisions. In the absence of national regulations about recruitment procedures, individual county Public Service Boards will presumably decide their own procedure. • Involvement of other relevant agents in recruitment. Procedures should spell out the involvement of other relevant agents in recruitment. These might include third party representatives such as a technically qualified person from the relevant national ministry (in the case of recruitment of a senior health official, for example), or a representative of a professional association or body (for example when recruiting accountants or engineers). • Checks and balances provide the best safeguards to integrity of public sector recruitment. The requirement for these procedures should be specified in the national framework of uniform norms and standards for county public services. The recruitment process should be designed to ensure that no single agent is likely to control the recruitment process from beginning to end. Source: World Bank staff analysis. discretionary funds, the district administrators had more to gain from aligning themselves to national MPs, than from practicing loyalty to the provincial government that had employed them.13 15.19 Aligning administrative and political boundaries is inefficient in other ways. Whereas electoral boundaries are designed mainly to distribute the population as evenly as possible among the constituencies, administrative boundaries should be designed for managerial efficiency. Matching administrative boundaries to political ones can result in nonsensical administrative units―for example, with residents on one side of a road in one district and those on the other side in another one. The CGA has directed that county governments establish “village units� below the level of sub county and ward. These layers of decentralization may involve substantial additional cost and in some settings (like urban Nairobi and Mombasa) they may not be a useful locus for service delivery or administration. 15.20 Many national public servants in counties are worried that they may be discriminated against, if they do not come from the county where they are currently working. The CGA does attempt to deal with this by directing the county Public Service Boards to be guided by the requirement that; “at least thirty percent of the posts at entry level are filled by candidates who are not from the dominant ethnic community in the county�.14 However, since the criteria for appointment decisions also do not include a requirement for appointment on merit,15 it is not clear how effectively this direction will work. 13 Whimp (2005). 14 County Governments Acts, section 65 (1)(e). 15 Article 232(1)(g) of the Constitution says that fair competition and merit should be the basis for appointments and promotions in the public service, subject to the principle of representing Kenya’s diverse community. 193 Chapter 15: Decentralising public services Uncontrolled spending on personnel expenditure 15.21 Uncontrolled spending on personnel costs is a common feature of weak subnational governments in many countries. In the Philippines, national-level regulations limit the number and cost of staff that can be engaged by local governments to between 45-55 percent of the previous years’ income. The center also determines pay scales and benefit entitlements. Despite this, between 1992-2001 aggregate, local spending on personnel averaged 56.8 percent of the previous years’ income, suggesting that these controls were not particularly effective.16 The reasons for excessive spending on wages are complex, and may lie partly in political pressure to satisfy patronage demands by awarding jobs; but it is also easier to spend money on salaries than on any other type of expenditure. Many subnational governments in developing countries have large unspent financial balances and struggle to fully spend their revenues. Chapter 12 notes that Kenya’s local governments also regularly report unspent balances. 15.22 Allocation formulas that do not take account of local cost differences may exacerbate the risk of excessive spending on personnel. Again using the example of the Philippines, a relatively crude revenue sharing formula did not follow the assignment of functions as closely as it should have, leaving some local government units with insufficient funding, and flooding others with excess funds that created incentives for overspending and overstaffing.17 Spending on goods and capital projects requires complex financial management procedures, whereas spending on personnel merely requires adding someone to the payroll. But spending on personnel alone does not contribute to better services. Unless personnel have adequate materials and funding for operational costs like travel, and adequate facilities from which to work, they are unlikely to deliver quality services. Furthermore, spending on salaries is “sticky�―meaning that it can be hard to reallocate quickly to other purposes. 15.23 There is anecdotal evidence that uncontrolled salary spending is already rife in local authorities, despite national government controls. The Ministry of Local Government (MoLG) in theory has power to control staffing levels within local authorities,18 but does not appear to have worked successfully. A recent report by PWC for the board of Nairobi City Council found that the Council was over-staffed, had a large number of ghost workers, and majority of staff were not adequately qualified for their positions.19 It would be insightful to understand why this situation has developed despite the existence of central oversight controls. If the reason is because of political interference, there is a risk that similar oversight mechanisms at the county level will be even more vulnerable to the same risks. 15.24 The public service provisions in the County Governments Act emphasize control over recruitment of public servants, but leave the employment of non-public-service staff relatively unregulated. Hiring casual workers is a common strategy to avoid the strict procedures applied to permanent public service staff. For example in the Philippines, over one-third of local employees in the early 1990s were non-career staff on short-term contracts, and in Indonesia more than ten percent of the government workforce is on contracts.20 The CGA provides a detailed framework requiring county public servants to be recruited and appointed by the county Public Service Board. In contrast, the only limitations on hiring of casual staff will be left to individual county Public Service Boards to decide, but the Act provides no guidance as to how they should do this (see Box 15-5). Casual hiring also circumvents qualification standards and pension requirements.21 16 Green (2005). 17 Ibid. 18 Menon (2008). 19 http://www.allkenyanews.com/2012/04/92pc-of-city-hall-workers-incompetent-pwc/ 20 Green (2005). 21 Ibid. 194 Devolution without disruption – Pathways to a successful new Kenya Box 15-5: Regulating casual or non-civil service employment. Many public service regulatory frameworks distinguish between engagement of civil servants and engagement of non- civil service staff. The latter may be employed as casual workers (employed on a week-to-week or month-to-month basis without a right to ongoing employment), or as contract workers on a fixed term, or in non-professional positions like drivers, labourers and the like. The regulations setting out terms and conditions for public service engagement should ensure that these cadres of workers are provided for separately. Where the framework for regulating civil service employment is tight, employment of non-civil service workers can act as a loophole that politicians and managers exploit to circumvent these controls. For example, a governor may engage persons on short-term contract without any requirement to justify the position, conduct a competitive recruitment process, ensure professional qualifications standards are met, or comply with principles for merit and ethnic diversity in recruitment. It is not uncommon for these contract arrangements to be rolled over repeatedly, so that the individual is in effect in long-term, full-time employment, but not subject to public service controls. The County Governments Act provides very tight controls on the engagement of civil servants. In contrast, the Act says relatively little about the engagement of non-civil servants. It does say that that the county Public Service Boards shall “regulate the engagement of persons on contract, casual workers, volunteers and interns in its public bodies and offices� (Section 74), but it is not clear what this means. Are county Public Service Boards empowered to make subsidiary regulations that bind county governments? If so, what sanctions can be imposed if the county government disregards the regulation? It would be better to either: (i) give individual county Public Service Boards clear authority to make and enforce rules on engagement of non-public servants; or, (ii) empower a national authority to set policy on engagement of non-civil servants by county governments, and adopt or propose regulations to enforce that national policy. Source: World Bank staff analysis. Widening capacity gaps 15.25 There is already gross inequity in the distribution of public service skills across Kenya. The extent of this inequity can be most readily gauged in the health sector, where the ratio of doctors to population varies by a factor of 90, from 4,100 people per doctor in Uasin Gishu to 378,000 /doctor in Kisii. The counties which have the lowest levels of existing public service skills are likely to be those who will have the hardest time attracting skilled personnel, given their overall remoteness, lack of mobility prospects they offer, and weak systems for incentivizing staff. For instance, there is some evidence that the impact of the recent economic stimulus package component for engaging health staff actually widened this gap; as existing public service personnel in remote areas seized the opportunity to apply for employment in the more desirable urban locations. 15.26 Having forty-seven individual county public services poses a significant disadvantage for smaller counties. The limited career growth opportunities are likely to deter many skilled candidates from applying for jobs within their services. This in turn will create pressure to enhance other terms and conditions of employment, most notably remuneration. While counties most affected by this problem have the highest percentage of revenues not already tied to existing expenditures, the pressure does potentially crowd out other important types of expenditures, particularly development spending. 195 Chapter 15: Decentralising public services 15.27 Attracting and retaining skilled staff in remote areas works best if incentives are more than just financial. However, these approaches work best within the framework of a single public service where staff can be offered a greater access to training and promotion, if they serve for a certain period of time in a remote area, or are barred from accessing these benefits until they have served there. This may be best achieved by maintaining an open labour market across county governments, allowing mobility and career advancement opportunities. Other strategies include: • Targeting young, well-educated, adaptable candidates through a new specialized recruitment effort. • Interim hiring on a contract basis to maintain flexibility to make terms and conditions more attractive, pending review and more permanent adjustment of pay policy. • Establishing a cadre of shared specialized staff to work across smaller counties, comprising of staff like valuers, specialist doctors, internal auditors and specialized procurement staff, possibly with the help of donor assistance initially, to create a unit to employ them.22 A robust framework for managing decentralized staff 15.28 The framework for managing decentralized staff should aim to achieve five objectives. First, county government functions must be clear, so that staff can know what is expected of them. Second, county governments must be able to hold their staff accountable for their performance. They must be able to direct and supervise staff, review their performance, and reward or discipline good or poor performance. Third, they need to be able to allocate staff where they are needed. Even if lower-level governments do not allocate staff to specific functions, they should be able to influence this through establishment controls.23 Fourth, county governments should have control over their financial resources flexibly, including through controlling both wage rates and numbers of staff, subject to hard budget constraints. Fifth, county governments need to be able to attract and retain skilled staff. To achieve these objectives, managerial controls need to be shifted to lower levels of government across six dimensions of managerial control: payroll processing, budget and establishment control, recruitment, career management, performance management and pay policy.24 Nevertheless, there are some important areas where national standards and national oversight can be important (see Box 15-6). Box 15-6: Five reasons for retaining some national involvement in local civil service management. (1) Standardization of working conditions across the country can help poorer regions. If working conditions are not standardized, local governments in poorer regions will find it difficult to compete against richer areas. (2) Standard terms and conditions also broaden civil servants career paths. This encourages civil servants with skills to stay at the local level. (3) Central government limits on hiring and pay protect subnational governments from local political pressure to overspend on wages and salaries. (4) National standards may be needed to reinforce minimum professional qualifications in sectors like health. (5) Central government can use the civil service as a tool for national integration, in countries where ethnic or other tensions threaten stability. Source: Green (2005). 22 Adapted from Reid (Unpublished). 23 Establishment controls are mechanisms that limit the overall number of staff that can be engaged by providing an overall ceiling usually framed in numerical but sometimes also in fiscal terms. 24 Evans and Manning (2004). 196 Devolution without disruption – Pathways to a successful new Kenya Box 15-7: National wage policy can compromise subnational governments. Colombia In the early 1990s Colombia began a process of decentralizing education funding and management to departments, districts and municipalities over a certain size. The additional responsibilities (including payment of teacher salaries) were funded from an intergovernmental transfer, the situado fiscal, fixed at 24.5 percent of national revenue, and specified in law to be spent on the key service delivery responsibilities of subnational governments. The spending responsibilities of subnational governments also increased rapidly. However, a rigid formula and conditions that the funds could only be spent on teachers’ salaries provided incentives for uncontrolled expansion of staff numbers. Over time, some subnational governments realized that they could not meet teacher salaries from the transfer, and so a special additional transfer had to be created, to bridge the gap. Philippines The 1989 Compensation and Position Classification Act set salaries in all but first-class local governments at lower than those at the center. However, since the salaries of the national public servants who had been transferred to local governments stayed the same, a wide gap was created between local and national staff working within the same local government. This influx of better paid staff caused a fiscal stress to local governments. It was subsequently exacerbated by the 1993 Salary Standardisation Law, which raised salary levels across local governments. The law also compressed salary bands, thus lessening the incentives for civil servants, to remain working for local governments, since the rewards of promotion had been reduced. China Civil servants in China are paid according to a national pay scale. Some poorer localities cannot afford the nationally mandated salary increases, and have been forced to either ignore these instructions or supplement personnel budgets from other sources, in particular, funds intended for capital expenditure. These pressures reduce the scope for local managers to have real control over their budgets. Source: Green (2005) and Reyes (2001). Pay policy 15.29 Pay policy is one area where national policy has both advantages and disadvantages. Transferability of public servants (something that will be quite important for smaller and more rural counties) is enhanced if pay and benefits are similar, for similar job levels across the whole public service. However, there are also many examples of central governments striking fiscally unsustainable wage bargains that have compromised the fiscal sustainability of subnational governments, and undermined their capacity to manage staff effectively (see Box 15-7). 15.30 An alternative approach is provided by the example of Brazil, which has imposed subnational personnel spending rules, as part of a fiscal responsibility framework. The Fiscal Responsibility law of 2000 imposes fiscal targets on subnational governments that personnel spending should not exceed 60 percent of net fiscal revenue, with ceilings for each branch of government.25 Conclusions 15.31 Kenya’s approach to devolving public service management is bold and transformative, but also potentially risky. Most countries decentralizing to small-medium subnational units provide a national- level service commission (like the Sierra Leone Local Government Service Commission) to manage the 25 Liu and Webb (2011). 197 Chapter 15: Decentralising public services technical aspects of human resource management. Without such a body, Kenya’s county Public Service Boards will need a great deal of help in developing the systems and practices for managing county public services effectively, and resisting the potential for politicization. 15.32 The regulatory gap concerning management of county public services is the most significant issue in the overall regime for county public services. A national law should make this clear, and provide clear authority for a specific body to set policy, fix or propose regulations or laws where necessary to implement that policy, and provide overall support and national leadership to counties in managing their public services. The framework should set national standards, and guide county governments in their day-to-day management of public servants. It should: (i) set the general framework for management; (ii) specify who is responsible for regulating the detailed rules about how county public servants should be managed; and, (iii) provide for statutory instruments (regulations, codes or the equivalent) to specify the details of the management arrangements. The framework should specifically address the potential loophole of non-public service employment. 15.33 As remaining policy is developed, decisions will need to be taken regarding the balancing of benefits of a national framework, with the political and responsiveness benefits of local control. Areas which may benefit from a national framework include providing career development pathways to encourage public servants to serve in smaller or more remote areas and imposing limits on salary spending. 15.34 County Public Service Boards will need a great deal of support, especially at the start, to function effectively. The independence and integrity of county Public Service Boards will be critical for safeguarding against political interference in civil service management. At the same time, there will be substantial pressures to put the boards in place early, since no recruitment to county public services can occur without them. The CGA requires governors to appoint board members through a competitive process, but does not specify what it should involve. A regulation under the Act (adopted before county governments become operational) should set out the process in detail, to ensure that there is a wide opportunity for applications, minimum qualifications for appointment, technical scrutiny of qualifications, and input from county citizens. The national government body responsible for policy on management of county public services should provide support to develop county Public Service Boards, in form of training, mentoring, standard procedures, peer review and opportunities for the forty-seven county Public Service Boards to share experiences and learn from each other. 15.35 Providing a pathway to increased autonomy for counties which demonstrate capacity could safeguard counties’ aspirations to control their own future. Even if a national regulatory framework is the starting point, counties that demonstrate capacity should be able to explore different options for more effective management of their public servants. The national regulatory framework should apply only until a county has been approved to take on the function of regulating its own public service, under its own laws and procedures. 198 Devolution without disruption – Pathways to a successful new Kenya CHAPTER 15: KEY INSIGHTS / RECOMMENDATIONS. 47 new separate public services will be created and Kenya’s new counties will have considerable autonomy, including over public service management. Introducing a single county public service structure provides an opportunity to streamline and rationalize a highly fragmented administrative structure at district level, but the future role of provincial administration remains to be resolved. The County Governments Act provides part of the legal framework for management of county public servants, but it leaves many gaps, and it is not obvious which national agency is responsible to fill them. There are important choices to be made about how far national standards should apply, and how much should be left to county governments. Too much flexibility in counties’ ability to make their own public service arrangements would create risks of: (i) elite capture; (ii) uncontrolled personnel spending; and, (iii) widening capacity gapes across counties. County Public Service Boards have been given too much power, which may put county fiscal sustainability at risk. A better approach is to give more stakeholders a formal role in HR decisions, and in doing so, minimize the risk of a single agent having too much control. *** National policy on county public services should be developed but first a national law should make it clear who is responsible for it. At the very least, the framework of regulation should provide uniform procedures and a comprehensive regulatory framework, to apply until the counties pass their own laws. National regulation of some aspects of county public service management is needed, in order to maximize career progression opportunities, and encourage public service mobility. Some national standardization of county pay policy would be beneficial, but care should be taken against imposing unaffordable fiscal burdens on county governments. The national government should support the establishment of county Public Service Boards to ensure that they are fully independent and competent, and a regulation under the County Governments Act should set out the process and criteria for appointing Board members. County Public Service Boards need support and capacity building to develop systems, practices and procedures for managing staff. 199 CHAP TE R SIX TE E N Managing a complex transition 16.1 Kenya’s transition to devolved government will be as complex as it is ambitious. Implementing devolution will involve simultaneous changes to both political and administrative arrangements, in a context where other key government institutions (like Judiciary, Parliament and oversight bodies) are also being reformed. An entirely new layer of elected and executive government is being created, that will assume major service delivery functions. With so much change occurring concurrently, complexity is inevitable. 16.2 Preparation and management for such a dramatic transition will be key, since whether it is managed well or poorly will set devolution forward or backward for years to come. There are six specific challenges which this chapter reviews, drawing on earlier analysis in the report: (i) management and coordination arrangements for transition; (ii) a complex asymmetric transfer of functions; (iii) the movement of staff from line ministries and local authorities to county governments; (iv) setting up sound PFM systems in county governments; (v) financing county governments during their first fifteen months; and, (vi) ensuring continuity in the delivery of urban services. Finally, transition will involve establishing systems in the forty-seven counties, to enable them to operate quickly after the county governments are elected. Putting Kenya’s devolution in context 16.3 Kenya’s devolution is ambitious by global standards. It is ambitious not so much because of the range of the functions being devolved, or the proportion of revenues being shared―in other countries decentralization has often involved a wider range of functions usually including education, and sometimes including courts and police. Rather, it is ambitious because of the scale of administrative change. Kenya’s current system of government is highly centralized by comparison with other countries in the continent. During the 1990s African nations embarked on an unprecedented transfer of power, resources and responsibilities to subnational governments. But among 24 African countries surveyed in 2002, Kenya was the only one not contemplating or implementing decentralization reforms.1 Today, the reverse is true. Kenya is one of the few unitary African states where the structure of subnational government can only be amended by a constitutional referendum. Transitioning from a highly centralized to a relatively autonomous devolved system will involve major changes in political culture, and also will require a significant effort to build the capacity of a new generation of county leaders. 16.4 Transition will be especially complicated because it involves simultaneous changes to political and administrative arrangements, at the same time as most institutions of government are also being reformed. In terms of political structures, the existing system of 175 elected local councils is being consolidated into forty-seven county assemblies. At the same time, an entirely new layer of administration is being created. Vertically, the new county administrations will operate partway 1 Brosio (2002). 200 Devolution without disruption – Pathways to a successful new Kenya between the 8 existing provincial administrations and more than 280 district administrations, and they will involve horizontal consolidation of local authority administrations and district administrations that have until now been operated largely independently of each other. There are very few global precedents for this kind of simultaneous political and administrative reorganization. Even the ‘big bang’ decentralization undertaken in Indonesia in the late 1990s does not compare, since the changes to administrative arrangements were relatively understated by comparison with those that Kenya is embarking on (see Box 16-1). Box 16-1: Transitioning to Indonesia’s ‘big bang’ decentralization. Until the late 1990’s the Indonesian state had been politically highly centralized. Dramatic changes followed the resignation of President Soeharto in 1998, two months after he had been relected. His successor, President BJ Habibie, oversaw a dramatic transition to popular democratic government that involved giving power, including control of staff and funding, to elected bodies at provincial and city (kota) or regency (kabupaten) levels (together referred to as regions). The transition was effected over about three years, and involved moving responsibility for around 2.1 million staff to the regional level. Although the transition was not without its problems, including the scrapping of the central coordination agency halfway through the process, the predicted disruption of service delivery did not occur. Despite its political centralization, the Indonesian public sector had long been deconcentrated. During the colonial period, the Dutch had established subnational government through sultantates and regencies. They had considerable local autonomy but reported to the centre. In 1979 regional administration had been restructured along uniform lines incorporating administrative divisions at province, regency/city and community levels (kekamatan). Regional peoples’ assemblies were established, but they had relatively little power. Administrative heads presided over each level, although their span of direct control was relatively limited and they reported to the centre. At decentralization, the power and control at the local level increased, but via empowerment of the existing electoral and administrative units. The new political bodies took over the functions of the existing assemblies, and they were given a wider set of functions and greater power to control resources. After devolution, the governors (heads of provinces), bupati (heads of kabupaten) and walikota (city mayors) who had previously been appointed by central government now were elected by and reported to the assemblies. On paper, the changes seemed more dramatic than they were on the ground. The units of de-concentrated administration remained the same, although reporting lines changed. Most of the public servants being transferred to the regions were already working there, reporting to their parent ministries in Jakarta. Following decentralization, they reported locally instead, and were in theory, at least more harmonized by being under the umbrella of a locally elected body. In practice, many of the old administrative divisions persist. Twelve years after decentralization the treasuries in some kabupaten continue to be comprised of separate siloed bureaus (corresponding to the different departments of the national ministry they had originally reported to) with no single administrative head. Source: Turner et al. (2003). Managing the risks of major change 16.5 Implementing the administrative elements of devolution will involve radical transformation of Kenya’s public sector. International experience suggests that change on this scale is likely to entail three main risks: (i) the risk that service delivery is disrupted; (ii) the risk of leaving an unsustainable fiscal burden at the centre; and, (iii) the risk that nothing really changes. 201 Chapter 16: Managing a complex transition 16.6 A golden rule of implementing decentralization is ‘do not disrupt service delivery’. Deterioration of services can rapidly erode popular support for decentralization, providing ammunition for recentralisation. In some cases, the end result is an unsatisfactory situation of ‘partial decentralization’, in which subnational governments have in sufficient power to determine service delivery outcomes, and citizens are therefore unable to hold them truly accountable.2 Uganda’s current decentralization arrangements are a legacy of this problem (see Box 16-2). To avoid this outcome, it is important to ensure that county governments are not put at risk of over-reaching themselves too early. The risks of service delivery deterioration are significant in the Kenyan situation, because of the scale of the changes to systems of public administration and, the short time-frame within which these are to be undertaken. 16.7 It is much easier to devolve funding than to devolve functions, and there is a risk that the centre is left with an unsustainable fiscal burden. A number of countries, particularly in Latin America, found that they had focused too much on devolving funding, while not paying sufficient attention to simultaneously devolving functions. Uncontrolled expansion of the total public wage bill can also be an outcome of poorly managed decentralization. In both cases the result is pressure on the central government to borrow unsustainably to meet its expenditure needs, and the outcome is to the detriment of both devolution and fiscal sustainability. 16.8 The scale and speed of the change raise the possibility that, in the end, nothing really changes. With so many different aspects of public administration that require concentrated effort to reform, it may be easier to just keep on doing the same thing. In some countries, decentralization has simply added an additional layer of political governance, without actually changing the way services are delivered. Unrealistic design may prove too difficult to implement, resulting in duplication, rather than devolution. The existence of two (central and subnational) administrative structures side-by-side creates unnecessary cost without improving service delivery or accountability, creating a problem of ‘persistent parallelism’. Madagascar, for instance has experienced this problem (see Box 16-2). 16.9 Five aspects of Kenya’s transition process warrant close attention and are dealt with in this chapter. The first is how such a complex transition will be managed and coordinated. Second, how a potentially complex and cumbersome asymmetric transfer of functions can be made simpler. Third, the single biggest transition issue will be to manage the movement of public servants from line ministries to county governments, in line with the (staff-intensive) service delivery functions. Four, the sudden transition from local authorities to new county governments raises serious risks of disrupting urban service delivery. Five, county governments will need sufficient resources to operate from day one, and the legal authority to spend funds, and they will need systems to manage funds accountably. Institutional arrangements for managing transition 16.10 Kenya now has a clear legal framework for managing transition, but a great deal of work remains to be done. In February 2012, Parliament enacted three laws to implement Chapter 11 of the Constitution (on devolved government). Together with the provisions of the Fifth Schedule of the Constitution, these laws provide a set of institutional arrangements for managing transition through a Transition Authority (TA), an independent body with broad membership and powers to coordinate implementation by the various organs of the government (see Box 16-3). The TA was appointed in July, and its members have a great deal to accomplish within eight months, until county governments are established. 2 Shantayanan et al. (2009). 202 Devolution without disruption – Pathways to a successful new Kenya Box 16-2: Two risks of poorly managed transition. Service delivery disruption undermines support for decentralization. Uganda’s 1995 decentralization was implemented very quickly. The design of the system gave local governments a level of autonomy and responsibility that in retrospect was clearly out of step with their capacity. Capacity building at the local level did not keep pace with the rapid devolution of power and resources. Line ministries resisted decentralization, and it was not in their interests to actively support capacity building in districts. District governments in some cases performed poorly. In the early 1990s a survey of the use of education funds revealed that very little of the funding was trickling down to the school level, with most funding being redirected to district administrative costs. International donors investing in social services put pressure on the government to tighten the conditions attached to district funds. Over time, the transfers to districts have become a complex web that is highly opaque and removes most discretion from local governments over spending. Although this lack of discretion makes it difficult to hold district governments accountable for poor outcomes, citizens in Uganda have a lower regard for their local governments than in many other Africa countries. A 2008 Afrobarometer survey recorded more than 60 percent dissatisfaction on most of the questions asked in the survey about local government responsiveness. Problems of persistent parallelism In some countries, political (and sometimes even fiscal) decentralization precedes administrative decentralization. While staff reporting to national government continue to be responsible for service provision, local governments are disempowered. Madagascar underwent a democratic transition in the early 1990s that provided the context for decentralizing power first to almost 1400 communes in 1995, and the establishment of six “autonomous� provinces through 1998 amendments to the Constitution. But the actual transfer of power and resources was never fully implemented. At all levels, central government continued to carry out many of the functions of the decentralized levels of government. A political crisis in 2002 sparked by tensions between autonomous provinces and the central government led to a reversal of decentralization policy, but without any official strategy being announced. Two causes for the persistent parallelism in Madagascar were been identified: first, the lack of clarity in roles and responsibilities at both levels of government which left lower levels of government unsure about what their functions were, and national government unsure as to who should resolve the problems. As a result, both central administration and national service delivery departments maintained deconcentrated authority at all levels. Second, the national government stalled on empowering the provinces, with which it was in conflict. Lack of political will to implement decentralization translated into maintenance of the status quo. Source: Kuteesa (2010). Empowering the Transition Authority 16.11 The TA will need funding to be effective. The most urgent priority is to make sure that the TA has sufficient resourcing, both in the current financial year, and under the 2013/14 budget, so that it can function effectively, including at the county level. The Commission on Revenue Allocation (CRA) and the Commission on Implementation of the Constitution (CIC) were established in early 2011, and realized that their operations were constrained by not having their own budget. 16.12 The TA will need to command respect across the government. International experience suggests that implementation is rarely completely successful (see Box 16-4), but having an independent body dedicated to the task of piloting it is a good start. However, the TA cannot make devolution happen alone. Devolution will occur through the actions of line ministries in each sector, and the TA’s job will 203 Chapter 16: Managing a complex transition Box 16-3: Institutional arrangements for transition. The Transition to Devolved Government (TTDG) Act specifies two transition phases: Phase One, from the commencement of the Act until the first election, and Phase Two, for three years after the first elections. A 15 member Transition Authority (TA) is established under the TTDG Act, to remain for a period of three years following the first elections. Six of the members are Principal Secretaries of relevant state departments (Ministries). The other nine members including the Chairman are fulltime. Its function is to facilitate and coordinate transition to devolved government, including: (i) determining the phased transfer of functions (including developing criteria for transfer); (ii) preparing an inventory of assets and providing a mechanism for their transfer; (iii) auditing human resources of government and local authorities and advising on rationalization and redeployment; and, (iv).assessing county capacity needs and recommending and facilitating measures to address them (TTDG Act, Section 7). It is also responsible to carry out a program of civic education, facilitate the development of county financial management systems and initial preparation of county budgets (TTDG Act, Fourth Schedule, Section 1). The TA will issue guidelines to state entities on the preparation of transition implementation plans (TTDG Act, Section 16). All entities are required to submit plans within the timeframe specified by the CIC (which oversees development of legislative and administrative procedures to implement the Constitution and reports to the Constitutional Oversight Implementation Committee of Parliament (CIOC)), which will then monitor the progress of implementing them, and may require a state entity to submit a progress report to it. In carrying out its functions, the Authority has broad powers to collect information and conduct investigations (TTDG Act, Section 8), and it can also make regulations (Section 37). The TA will report monthly to the CIC, and the CRA on the progress in implementing transition (Section 7(2)(n)) and every three months it will submit a progress report to the President, Parliament, the CIC and the CRA. The CIC will monitor and oversee the transition process. Source: World Bank staff analysis. be to ensure that they all do their part. The downside of creating a new body is that it must find a way to make other organs of the government to follow the direction it sets. Although the Transition to Devolved Government Act (TTDG Act) gives the TA power to make regulations, there is no clear sanction if they are not followed. The best way the TA can ensure that the government follows its lead is by: (i) ensuring that it gets Cabinet to sign off on its strategic direction; (ii) meaningful involvement of the implementing ministries; and, (iii) using transparent reporting of expectations and progress in meeting them, in order to create an environment where line ministries feel public pressure to meet their planned obligations. 16.13 Driving implementation through a planning process makes good sense. The process of developing individual ministry implementation plans will require close coordination between the CIC and the TA, as well as a concerted emphasis on having them implemented (see Box 16-5). Under the TTDG Act, the CIC is responsible for requiring the plans to be submitted, and monitoring their implementation, but the TA issues the guidelines about what the plans should contain. The guidelines issued by the TA could play a crucial role in setting the agenda of issues that line ministries should address, but ideally, it should be determined to dialogue with line ministries. This is why it would also be useful to develop an overarching strategy as well. The strategy should guide the decision about what to do first. It should highlight the respective roles of the TA and line ministries, and the relationship between the bureaucratic machinery, and the political leadership. 204 Devolution without disruption – Pathways to a successful new Kenya Box 16-4: Lessons from international experience of managing transition. First, coordination mechanisms are often too informal, providing little incentive or pressure for participation on the various agencies from which cooperation is required for decentralization to succeed. Second, there are not uncommonly multiple coordination mechanisms with poorly specified and sometimes overlapping functions, and they often have inadequate interaction with each other. Third, coordination mechanisms typically focus on the design of decentralization reforms; it is rare to have a mechanism that focuses effectively on implementation; if such mechanisms exist, there is often a weak interface between design and implementation. Fourth, there is typically a lack of appropriately structured and balanced responsibility among the various key parties, with the all too common problem of assigning leadership to a single interested agency or an inappropriate one, potentially undermining the willingness of the other partners to cooperate. Fifth, there is rarely sufficient enforcement authority given to decentralization coordinating mechanisms. Even if sound decisions are made, there is no reason to expect effective coordination, if there is no way to hold the concerned agencies to the terms of these agreements. Source: World Bank staff analysis. Box 16-5: Four ways to ensure plans get implemented. • Incorporate clear timelines, action officers and completion of indicators for each action. • Be absolutely realistic about how long it will take to achieve a milestone. • Monitor progress both formally and informally, on a very regular basis. • Use the opportunity of formal monitoring to review plans and adjust them so that they remain up to date. • Publish progress regularly and highlight the milestones each agency was expected to meet, and whether or not they were met. Source: World Bank staff analysis. Implementing transfer of functions 16.14 Transfer of functions, staff and funding lie at the heart of devolution. Delays in transferring any one of these three can profoundly affect the effectiveness of devolution. If the definition of functions being assigned is not clarified, there is a risk that line ministries will choose to interpret function assignments, in ways that effectively keep functions centralised. Confusion about function assignment is also likely to generate uncertainty among county public servants, who will not know whether they are meant to report to the new county government or not. Transition issues in relation to function assignment are among the most important to resolve as quickly as possible, because they form the starting point for settling questions about which staff will be transferred, and how much funding county governments will need. 205 Chapter 16: Managing a complex transition 16.15 The Constitution foreshadows that transfer of functions will take place over the three year period of transition following the first county government elections. The Sixth Schedule of the Constitution calls for the transfer to happen ‘no more than three years from the date of the first election of county assemblies’. Instead of having all county functions transferred on day one of the operations of county governments, the Constitution recognizes the need for a gradual transfer, to ensure that services are not disrupted. The main challenges in such an asymmetric process include: (i) avoiding an overly complex process; (ii) ensuring counties really are capable of carrying functions out; and, (iii) managing political expectations. 16.16 The Transition to Devolved Government Act approach to asymmetric function assignment may prove practically unworkable. It calls for each county to apply to the TA for function one by one and for the TA to evaluate whether or not the county has the capacity to run the function before it is transferred (see Box 16-6). This is likely to be a long, drawn out, overly contentious and expensive process, for both county and national governments. It will require that clear benchmarks are established to determining whether a county is capable of performing an assigned function, and that these benchmarks are reflected in the criteria that the TA approves as the basis for functions to be transferred. Box 16-6: Legal architecture for asymmetric transfer of functions. The Constitution provides a law to deal with phased transfer of functions assigned to counties, over a period of not more than three years, including the establishment of a criteria that must be met before functions are devolved, to ensure that counties are not given functions they cannot perform. Under the TTDG Act, the transfer of functions happens in two different ways. First, the TA identifies functions that should be transferred to county governments immediately after the first elections (Section 23(1)). Second, individual counties can make requests for transfer of the remaining functions (Section 23(2)). Applications for additional functions must be dealt with by the TA within 60 days, and should be based on the criteria that the TA has approved. Where a county does not meet the criteria, the TA should identify practical measures to build the capacity of the county government to enable it to perform the functions. The criteria can relate to the existence of appropriate legislation, establishment of administrative units, conduct of a capacity assessment by the county, presence of appropriate infrastructure and systems, including financial management systems, and whether the county has an approved plan in place in relation to the function (Section 24). To support the identification of devolved functions, the TA is required to ‘carry out an analysis of functions and competency assignment’ and develop a ‘plan for distribution of functions and competency’ including amendment of necessary Acts during Phase One (TTDG Act, Fourth Schedule, Section 1(j)). Source: World Bank staff analysis. 16.17 A phased approach to function transfer could be considered. Chapter 4 discussed these issues in more detail, and proposes an approach under which functions are grouped together to be transferred in three steps. The first step would involve those functions that the county must clearly be responsible for from day one (under the TTDG Act, these functions must be identified 30 days before the next election). 206 Devolution without disruption – Pathways to a successful new Kenya • The first set of functions to be transferred immediately on the establishment of the county government, would include functions that are obviously incidental to an independent level of government (like operating the county assembly and executive), and those which do not naturally belong to any national ministry (like urban service functions that were previously performed by local authorities). • The second set of functions to be transferred would be those that only require basic systems, and plans are in place (like day-to-day operation and maintenance of health facilities, and provision of agriculture extension services). • The third set of functions to be transferred would be those which require specialised systems or skills to be in place before the county is able to manage the functions. For example, transferring the responsibility for management of pharmaceutical procurement might require engagement of staff with skills to manage specialised procurement, training of staff in facilities, a stock management system, and a well-costed budget. Another area where skills may need to be developed is the capacity to manage large projects requiring feasibility assessment, design and public tendering. Line ministries should be involved in developing criteria for transfer, particularly of more complex technical functions. 16.18 In practice, transferring functions in a phased and asymmetric may involve thinking about functions in new and different ways. Generally, functions are thought of as sectoral components (for example, heath might be divided into clinical services, primary health care, or even into individual programs), but in fact, it might make sense to think of them as different public administration elements, not all of which can be transferred together. Transfer of a function involves the county government taking over several components of public administration in relation to the function, including: (i) responsibility for planning; (ii) supervising staff; (iii) managing of payroll; (iv) assuming control of assets and records; (v) control of operational and maintenance funding for the function; and, (vi) the responsibility for capital funding and major procurement. It may be more realistic to consider transferring these “function elements� in stages, rather than transferring a function like “health,� or “agriculture�. 16.19 The County Governments Act has effectively presaged this kind of separation between the different elements of devolved functions, by providing that all staff are seconded immediately. Under section 138 of the CGA, all officers appointed by the Public Service Commission and are serving in a county on the date of constitution of that county government, are deemed to be in the service of the county government. In other words, responsibility for supervision of staff in respect of all functions will be legally transferred to county governments straight away, regardless of whether the TA has formally transferred the specific functions to which they are attached to the respective county governments. 16.20 Giving county governments immediate responsibility for day-to-day supervision may help to manage political expectations. Although the Constitution clearly suggests that functions should be transferred in line with capacity, many Kenyans believe this is politically unrealistic. It is believed that counties that will be least able to take over functions, will be precisely the one which have historically been marginalised, and that Governors will not accept that there might be good reasons why their county is held back compared to a neighbour. Providing counties with immediate responsibility for holding public servants accountable for performance might be a useful way to manage their expectations about functions being transferred immediately after they take office. 207 Chapter 16: Managing a complex transition Transfer of public servants 16.21 Anxiety among public servants at the county level poses the possibly of greatest risk to implementing devolution. Many, but not all, of the public servants working in districts are likely to be seconded to the county government, as soon as it is formed. If public servants do not know whether to report to the county government or not, it is highly likely that service delivery will be negatively affected. In some cases, it is not clear what terms and conditions these public servants are employed under. A basic rule of thumb for minimising the risks of service delivery disruption, is to leave people doing the jobs they do now. 16.22 Initially, national public servants will be seconded to county governments. This will allow time for the counties to make their own permanent appointments to the county public services, including from among the seconded public servants. The CGA provides for all public servants (appointed by the Public Service Commission) who are working in a county at the date when the county government is formed, to be seconded to the county on the same terms and conditions.3 This provision is very broad, and might be interpreted to include public servants from all ministries, not just those which are devolved― for example, staff of the Ministry of State for Internal Security and Provincial Administration, Ministry of Education and so on.4 16.23 The first issue to resolve is which public servants Table 16-1: Authorized positions identified as attached to devolved functions, 2012/13 budget. will be seconded. This is why it is important to Country Average pop. resolve questions of function assignment that affect Local Authorities—Scales 9-1 2,000(1) large numbers of staff (for example, relating to Local Authorities—Scales 20-10 30,000(2) provincial hospitals) as soon possible. A December Medical Services 17,800 2010 report of the Public Service Commission Public Health 15,300 estimated that 190,000 officers at job groups A-N Agriculture 6,000 (those deemed to be at the “implementation� Livestock 12,700 levels) were working in counties. These figures included police, administration police and prison Youth 4,200 service members who will not be seconded. The Gender & Children 1,200 2012/13 budget identifies the budget lines that Lands 2,300 are expected to be devolved, and the number of Public Works 1,300 positions funded against them. Including the local Total 93,800 authority staff (who are not provided for in the Source: Figure (1) and (2) from Public Service Commission of Kenya (2010), all other figures from 2012/13 Budget Estimates, national budget), the total number of staff involved Volume 1 and 2. may be over 90,000. 16.24 It is not completely clear whether staff of local authorities are included in the secondment provisions. Under the CGA local authorities will be abolished (by the repeal of the Act) once final election results are known. While local authority staff at scales 9-1 who are appointed by the Public Service Commission (PSC) clearly fit this definition, by far the largest cadre of local authority staff are on scales 20-10. Section 138(6) expands the definition of officers who were “appointed by the Public Service Commission�, to include officers appointed on powers delegated by the Commission. That would appear to be intended 3 Section 138, County Governments Act 2012. 4 This is assumed to be a drafting error, and that only those public servants attached to devolved functions will be seconded. 208 Devolution without disruption – Pathways to a successful new Kenya to capture the local authority staff on scales 20-10 whose appointment is covered by a regulation under which PSC powers are delegated to local councils.5 However, the arrangements in Section 138 do not entirely fit the circumstances of local authority staff.6 16.25 The second issue to be resolved is their terms and conditions. The CGA provides that public servants are seconded on their current terms and conditions. The staff of local authorities are employed on salaries that are lower. They may have expectations that their salary levels will be brought in line with those of other seconded staff. If they are, this implies a considerable extra cost burden for county governments. 16.26 It is also not clear how the salaries of seconded staff will be managed. Payrolls for ministry staff are presently managed centrally. Ultimately, they should be transferred to county governments, but this will require establishing a computerised payroll system at the county level. In the interim, one option is to leave payroll being processed centrally. Alternatively, the responsibility for managing variation advices could be managed locally, and advised to the central processing authority. In either case, there needs to be capacity to link this spending to county budgets, in order to ensure that counties pay for their own staff, and only their own staff. 16.27 The legal framework for managing county public services should be clear before county governments are formed. As discussed in Chapter 15, there are still some substantial gaps in the legal framework for managing public servants, including the question of who decides their remuneration. These questions need to be resolved urgently. The respective roles of the Minister of Local Government and the Minster of Public Service in regulating county public service matters should also be clarified. 16.28 Setting up county Public Service Boards is a crucial priority. Chapter 15 also discussed the risks associated with politicisation of appointments to county Public Service Boards. Similar problems occurred in Uganda soon after decentralization, because district governments were empowered to appoint the members of the District Service Commissions responsible for hiring and firing district staff. Eventually, the power to hire the more senior staff of district governments was removed to the national level – very similar to the arrangements that are in place for appointment of senior local authority staff in Kenya. The integrity of the county Public Service Boards will hinge on the “competitive process� through which their members are selected. Since the establishment of the county Public Service Board is one of the first actions that any county government will take, it is important that the process for appointing members to be specified before the election. The TA should make it a priority to propose a regulation under the CGA specifying the process that governors should follow in appointing county Public Service Boards. Fiscal risks of the secondment approach 16.29 Ultimately, county governments will not be obliged to take over line ministries staff who are seconded to them. A secondment ends when the officer transfers to the county government, or the county government releases the officer back to the national government.7 It is possible, and indeed 5 Public Service Commission (Local Authority Officers) Regulations 2007, Regulation 10 provides that the PSC’s powers of appointment, promotion, transfer . 6 In particular, section 138(1) refers to officers being deemed to be “on secondment from national government� and in section 138(5) the secondment comes to an end when the officer is “released by the county government to the national government�. 7 County Governments Act, section 138(5). 209 Chapter 16: Managing a complex transition likely, that some national public servants doing devolved jobs may not be absorbed into county public services―either because the individual public servants prefer the career prospects offered in the national public service, or because the county does not offer them a position. If public servants believe they have a right of return to the national government, it is possible that they may opt not to apply for a position in the county public service―particularly if the county in which they are working is not one in which they want to spend the rest of their career life or do not see the best prospects for professional development. 16.30 The ‘zero-based’ approach to staffing county public services has been justified on political grounds. It is argued that county public services should not be obliged to accept corrupt or underperforming national public servants that will undermine their capacity to achieve real change. However, the President also gave a public undertaking in June 2011 that no person would lose her or his job as the result of devolution.8 However, it is a very unusual approach, except in situations where the number of staff being seconded is very small (as for example, when Sierra Leone’s local councils were first established, and secondment was used to set up the core staff of councils, numbering around 25 per council). Most commonly, countries that are establishing a new level of government simply transfer existing staff performing devolved functions to the new subnational governments. The Task Force on Devolved Government decided not to take this approach, because it interpreted Article 235 of the Constitution to mean that counties should be able to make their own choices about hiring public servants. 16.31 Giving public servants a right of return has potentially major fiscal implications that need to be assessed. If a large number of former secondees are returned to the national government, it may not have either the jobs or the funding for these staff. If even 10 percent of the staff identified by the 2012/132 budget as being attached to devolved functions return to the national government, the additional wage cost to the national government could be in excess of KES 5 billion.9 The combined wage expenditure of the local authorities in 2008/09 (the last year for which aggregated data is available) was almost KES 9.5 billion. There may be a greater risk of local authority staff not being retained by county governments, since they are perceived to be overstaffed. These potential fiscal risks, and options for ameliorating them, should be assessed before staff are seconded. If secondments need to be subject to certain terms and conditions in order to manage these risks, these should be imposed when the staff are first seconded immediately after the election. 16.32 Uncontrolled expansion of the national wage bill is a common outcome of decentralization, even where national staff are automatically transferred to the new subnational governments. In Kenya’s case, allowing counties to begin constructing their public services from a clean slate, exacerbates this risk. While the intention of allowing counties to be fully responsible for the performance of their public services is laudable, it may come at the cost of overall fiscal stability. This is just one of the trade-offs that must be managed, as the detailed design of devolution arrangements is developed. A related risk is the possibility that recruitment could be biased in favor of particular ethnic or tribal groups at the local level, at the expense of efficiency and possibly of social peace (replacing one system of patronage―real or perceived―by another). 8 But 22 Feb statement by Planning Minister that half the local authority staff would lose their jobs. 9 Personnel emoluments budgeted against devolved functions in 2012/13 are around KES 40 billion, with an additional KES 4 billion provided in the development budget for additional health staff financed by the economic stimulus package. 210 Devolution without disruption – Pathways to a successful new Kenya 16.33 Uncertainty among the public servants is likely to be the greatest threat to devolution. If public servants are uncertain about their employment future, they are unlikely to have much motivation to work. The better skilled staff may look for other openings. So far, there has been virtually no public discussion or debate about what will happen to the public servants who are currently working on devolved functions. The Transition Authority, working with the Ministry of State for Public Service, should ensure that a clear policy is made on how public service secondments, and transfers will be handled. Establishing Public Financial Management systems 16.34 There will be two major PFM challenges around transition. First, it is not clear how counties will be provided with resources during their early months, not least of all because there is still much uncertainty over which counties will take over what functions, as they come into existence within the fiscal year. Secondly, setting up sound PFM systems at county level will be a massive undertaking. 16.35 An approach to financing counties for the three months of 2012/13 and for the 2013/14 year should be developed well in advance. As discussed in Chapter 12, If the elections take place in mid-March 2013, there will only be 3 months before the 2013/14 financial year commences. It is unrealistic to expect county governments to complete a full budget preparation process within that time. Indeed, encouraging a truncated budget process for counties preparing their own budget for the first time could lead them to adopt poor procedures on a longer term basis. It would be better to support counties to begin their first full budget process in August 2013, in accordance with the PFM Act timetable, so that the 2014/15 budgets are prepared properly. 16.36 In the meantime, county governments will need legal authority to spend from day one, and this could be provided in a national law. The TA is authorised under the Transition to Devolved Government Act to assist with the preparation of county budgets. Beginning with the known costs of the functions that counties have inherited, these budgets could be prepared to cater for all devolved functions― including those managed centrally. This approach would facilitate easy transfer of control over funds to counties, if the TA authorises the transfer of functions part-way through a budget year. Total resources for counties could be presented as ‘county votes’ clearly set out in a national law. These would identify the total funding allocated to each county, and specify how much is to be spent directly by the county and how much will be spent on behalf of the county through national government systems. From 2014/15, the counties would allocate their resources through their own budgeting systems. 16.37 In order to manage the funds they control, counties will need Public Financial Management systems to be established, and staff trained to use them. Chapter 12 sets out a more detailed analysis of the existing state of Public Financial Management systems in counties. Key areas to focus on include: • Establishing bank accounts and cash management systems. • Developing a country-wide chart of accounts for counties that addresses the full range of county functions, including the functions of former local authorities. • Bridging the several local authority PFM systems that counties will inherit (some of which will be computerised and others manual). • Streamlining revenue collection which could continue to be managed by local authorities, in the interim. 211 Chapter 16: Managing a complex transition • Building capacity for revenue forecasting. • Training staff of district service functions in budgeting. Many of these staff are likely to have little or no experience of real-world budgeting, within firm resource constraints. • Developing model procedures for the 2014/15 budget process, including a budget circular, procedures for sector planning, and templates for the county fiscal strategy paper, budget estimates, appropriation bill and finance bill. Transitioning from local authorities to county governments 16.38 Local authorities will be legally abolished upon the establishment of county governments. Under the CGA, the Act under which the local authorities are created, the Local Government Act, is repealed on the final announcement of all the results of the first election held under the Constitution.10 This means that the local authorities as elective bodies, and as employing entities, will disappear. Since the law currently authorises the collection of revenues by local authorities, this too will cease, unless transitional arrangements are put into place. 16.39 A number of legal loose ends remain to be tidied up in order to ensure smooth transition from local authorities to county governments. The CGA empowers the Transition Authority to deal with all issues that arise as a consequence of the repeal of the Local Government Act, but a number of these issues can only be resolved by legislation. This is recognised in the Urban Areas and Cities Act (UACA), which provides that staff of the former local authorities shall be “seconded or otherwise deployed as may be provided by law�.11 The UACA refers to a number of other consequences of abolition of local authorities being dealt with by law: • “Rights, assets and liabilities accrued in respect of the properties vested in local authorities�… “shall be dealt with as provided by law� (Section 55 UACA). • Directions, resolutions, orders, authorisations, licenses and permits issued by local authorities are presumed to have been issued by boards under the Urban Areas Act (however, since there are many more local authorities than there are boards, and many counties will not have an urban board, this provision does not entirely cover the question) (Section 56, UACA). • Contracts and other similar arrangements entered into by local authorities continue to be in force, and are vested in a body established by law (Section 58, UACA). • Legal rights and causes of action against a local authorities will be sustainable against a body established by law (Section 59 UACA). 16.40 It may be prudent to allow the administrative structures of local authorities to remain for a period after county governments are established. While it would be inconsistent with the new model of devolution to allow elected local councils to remain in operation after the election, it would make good sense to keep the administrative structures of local authorities intact. The approach adopted for transition during the local government amalgamations in South Africa was to enact legal provision for the staff of a former council to form an administrative unit, to function until the new municipality had adopted a staff structure, reporting to the manager of the new municipality to which they were attached.12 10 County Governments Act, section 134. 11 Section 57. 12 See Section 33, South African Principal Structures Act 2000. 212 Devolution without disruption – Pathways to a successful new Kenya 16.41 County powers to collect revenues should be legally assured before the election. The Constitution assigns the entertainment and property taxes to counties as exclusive taxing powers. In addition, most counties are likely to want to raise revenue from other sources already mobilised by local authorities, such as trade licensing. Citizens of counties are already used to paying these taxes to local authorities (or at least, being liable to pay them). Counties can capitalise on this existing level of revenue mobilisation, but to do so, they will need legal authority―either from their own laws, or under laws of the national parliament. But it will take some time to put into place the capacity for counties to pass their own laws. It would be unfortunate if there were a legal lacuna after county governments are formed, during which time no taxes will be collected by anyone. 16.42 Parliament could empower county governments to collect taxes as soon as they are formed, but legislation should be enacted before the election. County taxing powers are the only subject matter on which the national government cannot pass laws. To minimise the possibility of a national law being declared invalid, the national Parliament should pass these enabling laws before the county governments are established. Ideally, the law should be framed so that it operates on an “opt out� basis―meaning that when individual county assemblies develop the capacity to pass their own laws, and wish to do so, the law will not apply. The window for Parliament to enact laws before it rises prior to the election, is rapidly closing. 16.43 Legislation could also provide a streamlined legal procedure for transfer of assets. The TTDG Act authorises the TA to deal with the transfer of assets, including immoveable property. Transfer of title often involves quite lengthy procedures. It would help to speed up the process of transferring land titles to county governments, if there were a procedure that allows this to be effected at the direction of an appropriate official, such as the Cabinet Secretary for Intergovernmental Relations. Engaging where it matters Dealing with line ministries 16.44 The process of developing the devolution arrangements has not been well integrated with line ministries. It is line ministries who will be devolving their functions, seconding their staff, and reorienting their own role to reflect a new focus on policy, standards and service delivery performance, and supporting the development of county capacity. Ministries are most likely to undertake effective change if they are actively responsible for planning it. Additional ways in which the TA could enhance coordination includes: • Seconding an official from each main ministry to the TA, to provide organisation-to-organisation links, and jump-start the development of the TAs networks within the mainstream bureaucracy. • Providing line ministries with standard toolkits that they can use (for example, techniques and templates for unbundling functions to be devolved). • Holding regular structured discussions with the Principal Secretaries (Permanent Secretaries) of the key ministries in devolved sectors to identify contentious issues and agree on a process for resolving them. • Establishing networks between officials in different line ministries, so that they can learn from each other, and to encourage a coordinated approach to resolving common issues and problems. 213 Chapter 16: Managing a complex transition • Convening focused discussions among interested ministries on issues requiring a common approach across government (for example, the process that will be used to second staff to county governments). Capacity building 16.45 Line ministries know best what capacities county administrations will need to manage their functions. Individual ministry implementation plans should include proposals for them to conduct capacity needs assessments, and to develop capacity building plans. The plans should be costed, and the Transition Authority can help bring together donors and ministries to identify where resources for capacity building can be sourced. 16.46 There will be big differences in capacity between weaker and stronger counties. One of the main risks of devolution is that it actually exacerbates these gaps. The TA should develop a strategy for ensuring that weaker counties receive special assistance. This might involve targeting donors to give priority to supporting specific counties, or developing special programs of assistance for them. A particular risk is the possibility that skilled public servants move away from remote locations. In conjunction with the ministry responsible for public service, the TA could help to identify incentives that would attract qualified staff to work in these locations. The incentives might include financial rewards in form of remote allowances, but they could also include greater access to other benefits of public sector employment like professional development, nationally-funded scholarships, or promotional opportunities (some more detailed suggestions are included in Chapter 15). Getting on the ground 16.47 It is on the ground, in each of the forty-seven counties, that evolution will ultimately fail or succeed. It is always easy to focus on what is happening in the national capital, at the exclusion of the real ‘boots- on-the-ground’ work of devolution out in the counties. A great deal will be expected of the forty-seven new governors, and they should have the basic systems in place that allow them to achieve some early wins. The TA proposes to appoint teams for each county, to help them get these systems established. The approach to local government transition adopted in South Africa, where transition committees were appointed to provide feedback on how the transition arrangements were working, might prove useful for Kenya (see Box 16-7). 16.48 With so many different steps involved in transition, sequencing will be important. It is helpful to think about what decisions the county governments will want to make first, and focus on putting in place the systems they will need for these. For example, the county assemblies will need clerks and other staff to make the work of the assembly and its committees effective. Under the CGA, staff can only be appointed by the County Assembly Service Board (a sister organisation to the County Public Service Board). Even if many national government staff are seconded to county governments initially, most county governments will want to appoint some staff of their own. This means that appointing the county Public Service Boards and establishing the systems to make them functionally effective are first-order priorities. Establishing PFM systems will also be very important. Box 16-8 sets out some of the priorities for establishing county systems before county governments are elected. 214 Devolution without disruption – Pathways to a successful new Kenya Box 16-7: South Africa’s Local Government Transition Guidelines. South Africa undertook a major consolidation and reorganization of municipalities in the late 1990s. Elections for the new councils were held in 2000. In preparation for the commencement of newly restructured municipalities, the Department of Provincial and Local Government issued guidelines. They provided for: • Establishment of a Transition Facilitation Committee with specific terms of reference. • Defining responsibilities of each administrative head, including for assets and bank accounts. • List of “Day one issues� to be addressed immediately after disestablishment of the old municipality and establishment of the new one. • Requirement for financial reports by the old council, transfer of votes to the new council, and revision of existing budgets, including provision for the new municipal manager to issue budget delegations • Requirements in relation to billing and debtors, transfer of insurance policies, covering assets to the new municipality, safeguarding of documents, stationery and stores. • Directions in respect of payroll administration. Source: World Bank staff analysis. 215 Chapter 16: Managing a complex transition Box 16-8: Checklist for establishment of county systems. Transition Team • Identify different roles needed within the team. • Develop ToRs for each position and identify skills needed. • Strategy to second from existing district or LA staff or second from national level. • Ensure team has necessary legal authority to undertaken their roles (for example, authority to spend funds). Human Resource Systems • Ensure framework of uniform national standards for county public service are in place. • Design procedures and systems needed to manage HR locally in each county. • Develop suggested procedure for recruitment/appointment of county Public Service Board members. • Action plan for county Public Service Board to follow in establishing procedures and systems. • Identify staff being seconded from local authorities and consolidate records. Function assignments • Agree a process for negotiating function assignments. • Identify key function issues by sector (which functions are clearly to be devolved; what are functions should or should not be devolved; what considerations should be taken into account in deciding). • Ensure function assignment papers identify a) criteria for transfer of functions, b) systems that counties need to put in place to manage a function. • Gazette determination of first functions being transferred. Develop first county budgets • Based on initial assigned function list, and staff identified for secondment, develop budget for known costs based on existing expenditure on functions. • Clarify scope for county governments to reallocate from original budget. Public Financial Management • Establish IFMIS in one district per county, open bank accounts. • Design and implement links to existing LAIFOMS. • Develop internal procedures for county budget. Focus initially procedures for budget execution and regular review, then on procedures for budget preparation. • Establish payroll system. Local authority staff should be the first to be migrated to new county payroll. Disaggregate national ministry payrolls to county level in preparation for transfer both to county payroll system. • Design revenue module for IFMIS. Transfer local authority revenue collection systems to new IFMIS. If this cannot happen immediately, develop a strategy for managing manual integration of revenue systems. • Appoint AIE holders for county programs. Development plans • Adapt/consolidate existing plans. Ensure they meet requirements of the PFM Act. Policy and legislation development • Ensure transition legal gaps filled. • Develop models for laws that counties will need to pass immediately (including County Assembly Standing Orders, procedures for county Public Service Board, General Provisions Act etc.). • Develop model procedures for county executive. Performance and oversight • Design interim performance oversight system. • Design and implement model procedures for public participation during preparation phase. Source: World Bank staff analysis. 216 Devolution without disruption – Pathways to a successful new Kenya CHAPTER 16: KEY INSIGHTS / RECOMMENDATIONS. Kenya’s devolution is highly ambitious even by global standards and the challenges implied in the transition to devolved government are daunting. There are three immediate risks including (i) service delivery interruption, (ii) an unsustainable fiscal burden at the center and (iii) failed devolution with national administrations continuing to exist alongside the new counties There is a legal framework in Kenya to manage the transition to devolved government but no ‘champion’ or entity that is empowered to lead the process Kenya’s choice to adopt a secondment model is original and potentially risky from a fiscal point of view. And there are also gaps with respect to the fate of civil servants carrying out non-devolved functions or the responsibility for paying seconded staff emoluments PFM systems and the budget process will need to be revised under the new dispensation: the transition years will be particularly delicate as the budget cycle is not aligned to the creation of counties and as PFM systems will need to be built from scratch Local authorities will be legally abolished the day of the election but counties will not be ready to take over their functions, staff and assets . *** The Transition Authority will need to be empowered financially and statutorily to lead the transition process with clear authority over other organs of government The TA should develop an overarching strategy for implementing devolution and provide uniform guidelines for line ministries to follow when preparing their implementation plans The TA should appoint teams in each county to help establish basic systems in a sequence prioritizing the first decisions county governments will have to make The TA should develop a strategy for filling major capacity gaps in Kenya’s most disadvantaged counties so as to pre-empt a widening of these gaps under devolution As a matter of priority the GoK should decide which level will be responsible for paying the salaries of seconded staff and if these will be deducted from the counties’ equitable share transfers During the transition year the national budget could include forty-seven county votes showing county allocations irrespective of which level actually carries out the corresponding functions A plan should be developed to absorb redundant staff at the national level should a significant number of county secondees seek to return to the national civil service It may be prudent to allow the administrative structures of LAs to remain for a period after county governments are established and county powers to collect revenues should be legally assured before the election. 217 MAIN CONCLUSION AND RECOMMENDATIONS In coming years, Kenya will experience a massive and complex transformation―administratively as well as politically. Devolution, enshrined in the Constitution, represents an extraordinary opportunity for Kenyans to hold their leaders to account at every level of Government, and to take an active part in the policy decision making process. This report has focused on ways of making the new system work for Kenya and on ensuring that the transition to devolved government can proceed smoothly in the critical early years of implementation. This concluding section summarizes key insights and recommendations from the report. The hope is that they will add to the existing menu of policy options that Kenyan citizens and decision makers will be able to draw upon when confronting the tough choices ahead. PART I: Seeing the future from the perspective of the past Implementing devolution, following the broad canvas provided by the Constitution, will be a tall order because of: (i) the radical scale of the proposed transformation―unique by international standards; (ii) the very short time frame for design and implementation; and, (iii) the fundamental heterogeneity, in starting points and in needs, of Kenya’s future counties. Kenya’s devolution is dramatic because it involves transferring substantial powers and resources to entirely new units of government, and it implies major changes to political and administrative institutions at the same time. Kenya’s public sector has been highly centralized for several decades, and devolution will imply both a major reorganization and a ‘culture change’. The new counties will need to set up institutions from scratch, while taking over functions, staff and revenue raising powers from existing institutions. Moreover, Kenya’s future counties are at dramatically different points (and possibly potential) in terms of overall development, economic infrastructure and access to social services. The challenge will be to make the framework of devolved government (and associated intergovernmental finances) accommodate such diversity. For instance, there are clear challenges for function assignment, fiscal equalization and, capacity building. Recommendations • Public expectations should be carefully managed as to how much and how fast devolution can deliver to avoid disappointment and overloading of weak counties. • Devolution provides a unique opportunity to rationalize the service delivery framework, by making counties the hub for organizing services at the local level. But this will require strategies to blend institutional frameworks of co-existing public administrations, clarify their roles and responsibilities, and transition staff currently employed in local authorities and districts. • Specific capacity building programs are needed for ‘lagging’ counties. • Equalization policies should be focused first and foremost on equalizing opportunities to access basic social services. 218 Devolution without disruption – Pathways to a successful new Kenya PART II: Equity between levels of government O ne of main challenges facing policy-makers will be to devise a “fair split� of national resources between the future counties and the national government. This will imply resourcing counties to adequately perform their functions, and to continue to deliver on the essential services devolved to them. But it will also involve finding the fiscal space required to deliver on the promise of equalization across Kenya’s counties. This can be done in part by reallocating resources from historically privileged to formerly neglected counties, and it will also call for tough trade-offs at the national level. At the same time, expectations need to be managed on how fast equalization can be achieved in a context of overall fiscal stress. A first step, in order to make this sharing process as transparent and fair as possible, is to clearly establish (and eventually cost) what functions counties will be expected to perform, in terms of both existing services to be devolved and new functions they will be required to perform. Failure to do this could result in blurred accountability, and possibly costly mismatches between funding allocations and spending obligations. This is a key mandate of the newly established Transition Authority (TA) which has a central role to play. A second step will be to design the intergovernmental financing framework in support of devolution. By collectively guaranteeing counties a minimum transfer of 15 percent of national revenue, the Constitution has not preempted the need for further definition of the intergovernmental transfer architecture, for two essential reasons. First, all available evidence suggests that aggregate county needs will be well in excess of 15 percent, calling for additional transfers to counties. Second, the equitable share transfer is only one of several instruments for financing counties. Others include: own source revenues, transfers under the Equalization Fund, and any other transfers that the national government could decide to extend out of ‘its’ share (whether conditional and formula driven or not). The intergovernmental transfer architecture should be designed as a system, with aggregate financing matched to aggregate needs. Recommendations • The TA should accelerate the process of clarifying function assignments, and create a framework of assignment to guide this process involving line ministries and central government, from the start. • A roadmap is needed to streamline and simplify the initial asymmetric transfer of functions, possibly by opting for “phased transfers� based upon clearly defined readiness criteria. • Comprehensive function assignment and costing are prerequisites for working out the vertical, aggregate division of revenues across levels of government. • Decisions regarding resourcing of counties should consider all possible sources of revenue available to them. • The tradeoff between unconditional and conditional funding deserves particular attention, as the scope for one will constrain the use of the other. • As a matter of urgency, the GoK should decide the fate of existing earmarked programs (CDF, LATF, RMLF), which if maintained, would constitute conditional grants to be resourced above and beyond the equitable share. 219 Main conclusion and recommendations PART III: Equity across Kenya’s diverse counties M aking sure that counties, collectively, are adequately funded is crucial from a macro and fiscal point of view. It is equally important that each county receives what it needs, as funds are further split across the forty-seven new units. Counties will be able to collect own revenues, but with rare exceptions these will be modest. In the short term, it will be key that the constitutional base for county tax collection is converted into legal instruments that enable counties to collect revenues from day one after the elections. Beyond this, it will be important to broaden the array of tax instruments that counties can use. Counties will initially be highly transfer dependent and much of their future resources will come from the county equitable share. Because of its constitutionally mandated floor, the equitable share is likely to be the single largest transfer. The CRA has recommended a formula for allocating the equitable share across Kenya’s forty-seven counties. The proposed formula is highly re-distributive, in line with the Constitutional emphasis on equalization, and this may create a fiscal challenge going forward. Because the formula is so redistributive, historically privileged counties (with relatively large inherited service delivery obligations) could be facing a significant financial shortfall―unless the total transfers to counties are expanded well beyond what is currently being spent on devolved functions. In turn, doing this would require fiscal space being created out of the national budget, which may be difficult to achieve immediately without compromising other essential services. It is also critical that the intergovernmental transfer system be designed in light of explicit choices regarding conditional funding because of: (i) Kenya’s tight fiscal environment; (ii) the constitutional guarantee that the equitable share will be purely unconditional; and, (iii) specific merits of conditional funding. The need for conditional transfers―if only to extend currently earmarked programs such as RMLF, CDF and LATF―will determine the fiscal space left available for the equitable share (above 15 percent). An expansion of the county equitable share above that minimum would proportionately reduce the scope for conditional funding (a key ingredient of devolution worldwide). Finally, the Equalization Fund―given its small size―will be insufficient to address the vast needs of Kenya’s marginalized communities. Moreover, the impact could be nil, or even negative, if it crowds out existing programs whose fate under the new dispensation is still unclear. Recommendations • A new interim national law governing county own revenues could provide an opportunity to revise and update existing taxes, particularly property rates (to apply to both land and improvements, and to ensure that public land generates adequate fiscal revenue) and the Single Business Permit. • Additional local revenues should be sought among a range of possible options, including surcharges on personal income tax, taxes on the use of motor vehicles and payroll taxes. • Particular consideration should be given to the sharing of taxation related to natural resources as their potential may increase vastly in years to come, and this issue often constitutes a source of conflict both between levels of Government, and with local communities. • Given the experimental nature of CRA’s formula, and lack of a clear estimate of actual county needs and capacity, it may make sense to limit the equitable share transfer, at or close to 15 percent, during a transitional period. Unmet needs at county level could be filled using other unconditional or conditional instruments. 220 Devolution without disruption – Pathways to a successful new Kenya • Given the radical and experimental nature of Kenya’s devolution it may make sense to maximize the scope for conditional funding (and minimize that of unconditional resourcing) if only to ensure that essential programs will remain funded but also to foster inter-county competition around performance and promote a system of performance monitoring at the local level. PART IV: Translating the vision, with a smooth transition Implementing the new devolved system of government during the transition will be a massive undertaking. Below are just a few of the more pressing challenges ahead. A new system of intergovernmental coordination will be needed. The Constitution limits the National Government’s power in relation to county governments, but many service delivery functions will be effectively shared between levels of Government. Effective coordination will involve cultural change at national level, and capacity building at county level. A substantive framework is provided for in the devolution laws, but still needs to be operationalized. Getting intergovernmental bodies working effectively with forty-seven county members will be particularly challenging. A new legal framework for PFM has been adopted, but there are many remaining gaps to fill. Budgeting for transition will be particularly challenging because counties will only have very limited time to prepare budgets for 2013/14, and because they will take on their assigned functions gradually, following an asymmetric transfer process. Under the new dispensation, the national legislature will have vastly increased power in the budget process. This could potentially derail the budget process at national and county levels if mechanisms are not put in place to manage the politics of legislative budgeting. At the county level, the budget process will have to be built up from scratch as there is currently only limited experience and capacity at the local level, especially within districts. This will require central coordination and significant capacity building. Increased accountability of government is at the heart of the Constitution, but decentralization is not a silver bullet to make it happen. At the local level, weaknesses include limited capacity, poor information systems, weak checks and balances, and often poorly organized civil society. Setting up a system of social accountability implies greater fiscal transparency, effective citizen participation, and accountability mechanisms. There is great uncertainty surrounding the fate of cities under the new dispensation and Kenya may be adopting a unique approach by re-centralizing municipal services to county level. Kenya’s cities will no longer have an elected leadership, and only a handful will have a corporate body to manage them. All of the revenues currently enjoyed by Local Authorities will become county own revenues and urban service resourcing will be at the discretion of mostly rural counties. Specific arrangements will be needed to ensure that large cities are appropriately governed and it will also be crucial that urban functions are appropriately resourced. Leaving urban services and governance unattended would be dangerous as urban areas worldwide are the main engines of growth in a country. As counties come into existence, forty-seven new separate public services will be created. Kenya’s new counties will have considerable autonomy, including over public service management. The CGA provides part of the legal framework for management of county public servants, but it leaves many gaps and it is not obvious which national agency is responsible to fill them. There are important choices to be made about how far national standards should apply, and how much should be left to county governments. Too much flexibility 221 Main conclusion and recommendations in counties’ ability to make their own public service arrangements would create risks of: (i) elite capture; (ii) uncontrolled personnel spending; and, (iii) widening capacity gaps across counties. Moreover, the secondment approach chosen by Kenya’s policy-makers could create significant fiscal risks, and an uncontrolled expansion of the civil service wage bill at the national level. It is inevitable that implementing such sweeping change across so many fronts in a short time-frame will be challenging. The overriding aim should be to minimize disruption to service delivery. Kenya has put in place an independent TA, but there is much work to be done. The TA will need to gain the trust and cooperation of line ministries, since it is their staff, functions and funding that will be devolved. Resolving uncertainty about the future of public servants in districts and local authorities is crucial. An immediate task after the election will be to integrate former local authorities and new county governments, and some legal loopholes remain to be closed to minimize confusion and complication. Counties will need funding from day one, but are unlikely to have the systems in place to do this on their own. A transitional approach to financing counties remains to be developed, and will have a critical impact on how smoothly transition can be managed. Recommendations • Despite the many competing priorities, it is important to focus on getting intergovernmental coordination bodies functioning early on. • One priority is to get sector bodies working on issues of function assignment, service delivery standards, and performance monitoring. • Another priority is to focus on intergovernmental relationships at the county level, particularly by resolving the role of provincial and district administration staff, and designing the new arrangements for the remaining national staff at county level. • During the transition period, the budget could include county votes showing total county allocations, even if a portion remains executed by the National Government, on behalf of counties. • To mitigate the risk of executive-legislative gridlock during the budget process, a two-step budget process could be introduced with parliamentary votes at each stage, to generate consensus around the fiscal framework. • Building county PFM systems from scratch will require central support to capacity building, monitoring of county progress against clear benchmarks and standardized guidelines and templates developed nationally (for example to help counties develop an integrated planning and budget process). • Public financial information (including results) should be made public in a way that allows citizens to assess the efficiency and effectiveness of national and county spending. • Citizens should be actively involved in planning and budgeting at local level and equipped with the tools to make significant contributions. • More urban centers should have corporate bodies to manage them. This could be achieved either by lowering the threshold for ‘municipality’ status, or by amending the powers and functions of towns under the current law. • Functions of city and municipal boards should be clarified as well as the formal process for counties to delegate additional functions to them. • A national policy on county public services should be developed. At the very least, the framework of regulation should provide uniform procedures and a comprehensive regulatory framework to apply until the counties pass their own laws. 222 Devolution without disruption – Pathways to a successful new Kenya • National regulation of some aspects of county public service management is needed, in order to maximize career progression opportunities and encourage public service mobility. • The National Government should support the establishment of county Public Service Boards to ensure that they are fully independent and competent, and a regulation under the County Governments Act should set out the process and criteria for appointing Board members. • As a matter of urgency, the fate of Local Authorities’ employees should be clarified through a law to deal with transition issues involved in abolishing Local Authorities. • The TA will need to be empowered financially and statutorily to lead the transition process with clear authority over other organs of government. • The TA has started to develop an overarching strategy for implementing devolution, but what is also needed are uniform and authoritative guidelines for line ministries to follow when preparing their implementation plans. • The TA should appoint teams in each county, to help establish basic systems in a sequence, prioritizing the first decisions County Governments will have to make. 223 BIBLIOGRAPHY ▪ Ahmad, E. and Searle, B. (2005). On the implementation of transfers to subnational governments. IMF Working Paper 130. ▪ Bahl, R. and Wallace, S. (2004). Intergovernmental Transfers: The Vertical Sharing Dimension. Working Paper 04-19, Andrew Young School of Policy Studies. ▪ Bahl, R. and Bird, R. M. (2008). Subnational Taxes in Developing Countries: The Way Forward. Public Financial Publications Inc. ▪ Bardhan, P. and Mookherjee, D. (2006). Decentralization and Local Governance in Developing Countries: A Comparative Perspective. Cambridge, MA: MIT Press. ▪ Bird, R. M., & Slack, E. (2006). Taxing Land and Property in Emerging Economies: Raising Revenue... And More? In Ingram, G. K. and Hong, Y.-H. Land Policies and Their Outcomes. Cambridge, Mass.: Lincoln Institute of Land Policy. ▪ Bird, R. M. and Smart, M. (2009). Earmarked Grants and Accountability in Government. University of Toronto. ▪ Blunt, P., Turner, M. and Lindroth, H. (2012). Patronage’s Progress in Post-Soeharto Indonesia. Public Administration and Development, Volume 32, pages 64-81. ▪ Boadway, R. (2003). National taxation, Fiscal Federalism and Global Taxation. WIDER . ▪ Boadway, R. and Shah, A. 2007. Overview in Intergovernmental Fiscal Transfers: Principals and Practice. ▪ Branch, D. and Cheesman, N. (2007). The politics of control in Kenya: Understanding the bureaucratic-executive state, 1952-78. Review of African Political Economy, Volume 33:107. ▪ Brosio, G. (2002). Decentralisation in Africa. In Ahmad, E. and Tanzi, V. Managing Fiscal Decentralization. Routledge, London and New York. ▪ Devarajan, S. Khemani, S. and Shah, S. (2009). The politics of partial decentralization. In Ahmad, E. and Brosio, G. (eds), Does Decentralization Enhance Service Delivery and Poverty Reduction. Edward Elgar, Cheltenham. ▪ Dumas, V. and Kaizer, K.(2009). Subnational Performance Monitoring Systems. Commonwealth of Local Governments Forum Conference 2009. ▪ Eaton, K., Kaiser, K. and Smoke, P. (2010). The Political Economy of Decentralisation Reforms: Implications for Aid Effectiveness. World Bank. ▪ Evans, A. and Manning, N. (2004). Administrative Decentralization: A Review of Staffing Practices during Decentralization in Eight Countries. Unpublished manuscript prepared for the World Bank. ▪ Ferrazzi, G. and Rohdewohld, R. (2009). Functional Assignment in Multi-Level Government. Esohborn: GIZ. ▪ FIAS/World Bank. (2009). Taxation as State-Building: Reforming Tax Systems for Political Stability and Sustainable Economic Growth A Practitioner’s Guide. International Finance Corporation, Washington DC. ▪ FIAS/World Bank. (2011). Avoiding the Fiscal Pitfalls of Subnational Regulation: How to Optimize Local Regulatory Fees to Encourage Growth. World Bank, Washington DC. ▪ Fjeldstadt, O.-H., and Semboja, J. (2000). Dilemmas of Fiscal Decentralisation.
A Study of Local Government Taxation in Tanzania. Forum for Development Studies, Vol. 27, No. 1. ▪ Fjeldstadt, O.-H., Katera, L., and Ngalewa, E. et al. (2009). Outsourcing Revenue Collection to Private Agents: Experiences from Local Authorities in Tanzania. Special Paper 09.28, Dar es Salaam, Research on Poverty Alleviation (REPOA). ▪ Garcia-Milà, T. and McGuire, T. J. (2002). Fiscal Decentralization in Spain: An Asymmetric Transition to Democracy. 224 Devolution without disruption – Pathways to a successful new Kenya ▪ Gikonyo, W. (2008). The CDF Social Audit Guide: A Handbook for Communities. Open Society Initiative for East Africa, Nairobi. ▪ Government of Kenya. (1965). Sessional Paper No. 10 on African Socialism and its Application for Planning in Kenya. ▪ Government of Kenya. (2006). Kenya Integrated Household Budget Survey (KIHBS) 2005/06. ▪ Government of Kenya. (2007). Basic Report on Well-Being in Kenya (Based on KIHB Survey 2005/06). Kenya National Bureau of Statistics, 2007. ▪ Government of Kenya. (2009). Census (Kenya National Bureau of Statistics). ▪ Government of Kenya. (2011). Report of the Auditor-General on the Appropriation Accounts, other Public Accounts and the Accounts of the Funds of the Republic of Kenya for the Year 2009/10. Nairobi, Kenya. ▪ Government of Kenya. (2012). Report of the Auditor-General on the Appropriation Accounts, other Public Accounts and the Accounts of the Funds of the Republic of Kenya for the Year 2010/11. Nairobi, Kenya. ▪ Green, A. E. (2005). Managing human resources in a decentralized context. In Smoke, P. and White, R. (ed.) East Asia Decentralizes, The World Bank. ▪ International Monetary Fund. (2009). Macro Policy Lessons for a Sound Design of Fiscal Decentralization―Background Studies, IMF Fiscal Affairs Department, July 2009. ▪ Iversen, V., Fjeldstad, O.-H., Bahiigwa, G., Ellis, F., and James, R. (2006). Private Tax Collection―Remnant of the Past or a Way Forward? Evidence from Rural Uganda. Public Administration and Development 26, 317 – 328 (2006). ▪ JICA. (2007). Local-level Service Delivery, Decentralisation and Governance: A comparative study of Uganda, Kenya and Tanzania Education, Health and Agriculture sectors. Kenya Case Report. ▪ Kaiser, K. (2006). Decentralization Reforms. In Coudouel, A. and Paternostro, S. (eds). Analyzing the Distributional Impact of Reforms: A practitioner’s guide to pension, health, labor markets, public sector downsizing, taxation, decentralization, and macroeconomic modeling. Washington, DC: World Bank, pp. 313–353. ▪ Kelly, R. (2000). Property Taxation in East Africa: The Tale of Three Reforms. Lincoln Institute of Land Policy Working Paper, 2000. ▪ Kelly, R. (2002), Property Taxation in Kenya. In Bird, R. M. and Slack, E. (eds.) Land Taxation in Practice: Selected Case Studies. World Bank, Toronto. ▪ Kelly, R. (2011). Strengthening the Local Revenue Side of Intergovernmental Financing, including Property Tax Devolution-Indonesian and International Experiences, Paper prepared for the International Conference on Fiscal Decentralization in Indonesia a Decade after Big Bang. Jakarta, 13-14 September 2011. ▪ Khemani, S. (2006). Local Government Accountability for Service Delivery in Nigeria, Journal of African Economies 15 (2): 285-312. ▪ Kramon, E. and Posner, D. (2011). Kenya’s New Constitution. Journal of Democracy 22.2: 89–103. ▪ Kuteesa, F. (ed). (2010). Uganda’s Economic Reforms: Insider Accounts. Oxford University Press. ▪ Lakin, J. (2012). Drafting Regulations for Kenya’s Public Financial Management Act 2012: Key Concerns Related to Tansparency, Budget Brief Number 15, August 2012, International Budget Partnership, 2012. ▪ Letelier, L. (2006). Local Government Organization and Finance: Chile. In Shah, A. (ed.) Local Governance in Developing Countries. World Bank. Washington DC. ▪ Liu, L. and Webb, S. (2011). Laws for Fiscal Responsibility for Subnational Discipline: International Experience. World Bank, Poverty Reduction and Economic Management Policy Research Working Paper 5587, Washington DC. ▪ Mboga, H. (2009). Understanding the Local Government System in Kenya: a Citizen’s Handbook. Institute of Economic Affairs. 225 Bibliography ▪ Menon, B. (2008). Cities of Hope: Governance, Economic and Human Challenges of Kenya’s Five Largest Cities. Water and Urban Unit 1, Africa Region, World Bank, Washington DC. ▪ Mikesell, J. (2007). Developing Options for the Administration of Local Taxes. Public Budgeting and Finance 27, no. 1. ▪ Ministry of Finance. (2012). Medium Term Budget Policy Statement (BPS). April 2012. ▪ Mitullah, W. (2010). Local Government in Kenya: Negative Citizen Perception and Minimal Engagement in Local Government Affairs. Afrobarometer Briefing Paper No. 89. ▪ Murray, C. and Wehner, J. (2012). How New Laws Balance Budgetary Powers: Constitution Has Safeguards that Ensure the Executive Can no longer Unilaterally Dictate Fiscal Policy and the Legislature Cannot do as it Wishes Without Upsetting Economic and Political Stability, Nairobi Law Monthly, January 2012, pp. 38 – 43. ▪ Kenya Local Government Reform Programme. (2007). Study on the Impact of the Local Authority Service Delivery Action Plan. Ministry of Local Government, Nairobi. ▪ Ministry of State for Planning, National Development and Vision 2030. (2010). Public Expenditure Review: Policy for Prosperity 2010. Republic of Kenya. ▪ OECD. (1997). Managing Across Levels of Government. ▪ Petchey, J. D., MacDonald G. A., and Josie, J. (2004). A capital grant model for South Africa. Paper presented at the 10th Anniversary Conference of the Financial and Fiscal Commission, January 1st, 2004, Cape Town, South Africa. ▪ Petchey, J. and Macdonald, G. (2007). Financing Capital Expenditures through Grants. In Boadway, R. and Shah, A. (eds). Intergovernmental Fiscal Transfers, The World Bank. ▪ Public Service Commission of Kenya (2010). A proposal on Modalities for Implementation of the Commission's Mandate under the Constitution, December 2010 - February 2011. ▪ Rao, G. (2003). Intergovernmental finance in South Africa, Some observations. Working Papers 1. National Institute of Public Finance and Policy. ▪ Reid G. (Unpublished). Unpublished Advice on Human Resource Issues and Options (prepared for Sierra Leone). ▪ Republic of Kenya (2011). Final Report of the Task Force on Devolved Government – Volume I: A Report on the Implementation of Devolved Government in Kenya. Republic of Kenya. ▪ Reyes, J. (2001). Colombia: Decentralised Education Management. Human Development Department, LCSHD Paper Series, World Bank, Washington DC. ▪ Searle, B. and Martinez-Vasquez, J. (2007). The Nature and Functions of Tied Grants. In Searle, B. and Martinez- Vasquez, J. (eds.) Fiscal Equalization: Challenges in the Design of Intergovernmental Transfers. Springer. ▪ Shah, P. and Wagle, S. (2002). Participation in Public Expenditure Systems. Participation and Civic Engagement Group - Social Development Department. The World Bank. ▪ Shah, A. (2007). A practitioner’s guide to intergovernmental fiscal transfers. In Boadway, R. and Shah, A. Intergovernmental Fiscal Transfers, Principles and Practice, World Bank. ▪ Shantayanan, D., Khemani, S. and Shah, S. (2009). The Politicsof partial decentralization, in Etisham Ahmad and Giogio Brosio. Does Decentralization Enhance Service Delivery and Poverty Reduction. Edward Elgar. ▪ Smoke, P. (2001). Fiscal decentralization in developing countries – A review of current concepts and practice. UNRISD. ▪ Smoke, P, Muhumuza, W. and Ssewankambo, E. (2010), Comparative Assessment of Decentralization in Africa: Uganda Desk Study. USAID. ▪ Smoke, P. (2012). Recentralization in Developing Countries: Forms, Motivations and Consequences. New York. ▪ Steffensen, J. (2011), Performance-Based Grant Systems: Concept and International Experience. UNCDF. 226 Devolution without disruption – Pathways to a successful new Kenya ▪ Steytler N, and Fessha, Y. (2007). Defining Provincial and Local Government Powers and Functions. South African law Journal 124: 320. ▪ Task Force on Devolved Government (2011). Final Report of the Task Force on Devolved Government. Volume 1. A Report on the implementation of Devolved Government in Kenya. ▪ Transparency International. (2011). Devolution in Kenya: Linking discretion with accountability. ▪ Taliercio, R. T. (2005). Subnational Own Source Revenue: Getting Policy and Administration Right. In Smoke, P. and White, R. (ed.) East Asia Decentralizes, The World Bank. ▪ Turner, M. Podger, O. Sumardjono, M. and Tithayaasa, W. (2003). Decentralisation in Indonesia: Redesigning the State. Asia Pacific School of Economics and Governance. Australian National University Press . ▪ USAID (2010). Uganda Desk Study. Comparative Assessment of Decentralization in Africa. ▪ UCLGA (2010). Local Government Finance. The challenges of the 21st. Second Global Report on Decentralization and Democracy United Local Government and Cities Association in Barcelona. ▪ Wehner, J. (2001). Nigeria: Democracy brings opportunity for participation in budget process, IDASA: Africa Budget Project 12 November 2001. ▪ Wehner, J. (2010). Legislatures and the Budget Process: the Myth of Fiscal Control. Palgrave Macmillan, Basingstoke, UK. ▪ Whimp, K. (2005). Comments on Problems with the Organic Law on Provincial Governments and Local-level Governments, unpublished paper. ▪ White, R. & Smoke, P. (2005). East Asia Decentralizes, in World Bank, East Asia Decentralizes: Making local government work. World Bank, Washington DC, page 131. ▪ Williamson, T. (2010). Fiscal Decentralisation in Uganda. In Kuteesa, F. (ed). (2010). Uganda’s Economic Reforms: Insider Accounts. Oxford University Press. ▪ World Bank. (2000). World Development Report: Entering the 21st Century: The Changing Development Landscape. World Bank, Washington DC. ▪ World Bank. (2002), Kenya: An Assessment of Local Service Delivery and Local Governments in Kenya, Report no. 24383 Water and Urban 1, Africa Region, World Bank Washington DC. ▪ World Bank. (2003). Madagascar Decentralization. Report 25793, World Bank Washington DC. ▪ World Bank. (2004). 2004 World Development Report. Making Services Work for Poor People. World Bank, Washington DC. ▪ World Bank. (2008). Cities of Hope? Governance, Economic and Human Challenges of Kenya’s Five Largest Cities. Report No. 46988, World Bank. ▪ World Bank. (2008). Nigeria’s Experience Publishing Budget Allocations: A Practical Tool to Promote Demand for Better Governance. Social Development Notes, October 2008. ▪ World Bank. (2009). Local Taxes, Regulations and the Business Environment: Finding the Right Balance. Investment Climate in Practice – Business Operations and Taxation, No. 5, April 2009. ▪ World Bank. (2009). Reshaping Economic Geography. World Bank, Washington DC. ▪ World Bank. (2011a). A Bumpy Ride to Prosperity - Infrastructure for Shared Growth in Kenya World Bank Office Nairobi. ▪ World Bank. (2011b). Kenya Economic Update # 5. World Bank Office Nairobi. ▪ World Bank. (2011c). Kenya Economic Update # 4. World Bank Office Nairobi. 227 Annex 1: Overview of FDKP T he “Fiscal Decentralization Work Program� has been implemented over the past thirteen months through the Australian Aid funded Fiscal Decentralization Knowledge Program (FDKP). The programme comprises both supply- and demand-side workstreams, with major activities including TA, capacity building and analytical work, in-depth analysis and just-in-time responses to government demand, and most recently the report Devolution without disruption: Pathways to a successful new Kenya. It brings together key stakeholders around the theme of Kenya’s devolution and focuses on the following four areas: i) managing institutional change for effective decentralization; ii) adapting financial management and data systems at national and local levels; iii) designing mechanisms and formulas for vertical and horizontal revenue sharing; and iv) strengthening fiscal management capacities of national and county governments. The program operates within a remarkable window of opportunity for influencing devolution―related policy outcomes, as Kenya contemplates and prepares for the introduction of county governments in 2013. This is because the constitution implementation process itself is quite fast paced―29 pieces of game-changing legislation will need to have been passed between August 2010 and August 2012, with 19 more by August 2015, not to mention additional regulation to support the implementation of Acts. FDKP is a leading source of advice to the Government of Kenya, working across the full range of government stakeholders involved in the devolution design and implementation process. The key stakeholders with whom FDKP has been interacting are likely to determine the core elements of design of Kenya’s devolution arrangements. The stakeholders include: • The Task Force on Devolved Government: Established in October 2010, the Task Force was mandated to propose implementation mechanisms for the devolved system of government as envisaged in the Constitution of Kenya. The Task Force , which was disbanded in July 2012 after submitting its final report, worked closely with FDKP in conceptualizing the framework of Kenya’s devolution and identifying bottlenecks which needed to be mitigated very early in the process, to ensure its successful implementation. The FDKP provided support to the Task Force including background data and analysis, theoretical frameworks on fiscal decentralization and global devolution experience from countries like South Africa, Uganda, Indonesia and Papua New Guinea. • A Committee on Fiscal Decentralization established within the Ministry of Finance: The Committee was constituted to advise The National Treasury and central ministries on how to restructure Kenya’s institutional and fiscal architecture as required by the new constitution. Besides being the World Bank’s principal counterpart within the Government of Kenya, the Committee on Fiscal Decentralization, is the main point of engagement with Kenya’s policy-makers. 228 Devolution without disruption – Pathways to a successful new Kenya • The Commission on Revenue Allocation: Established in early 2011, with the responsibility of recommending the formulas for both vertical and horizontal sharing, the CRA has collaborated quite closely with FDKP. Four highlights of this engagement are: i) a joint trip by CRA and FDKP to South Africa for a peer-to-peer learning engagement with the Financial and Fiscal Commission (FFC); ii) the joint hosting of Uganda’s Local Government Finance Commission (LGFC); iii) the joint production in December 2011, of the first edition of the Kenya County Fact Sheets (with socio-economic and fiscal data aggregated at the county level for the first time); and, iv) ongoing technical assistance. • The Commission on Implementation of the Constitution: The CIC was established to monitor, facilitate, coordinate and oversee the development of the legislation and administrative procedures required to implement the Constitution. In fulfilling this mandate, CIC receives comments from a wide range of interested parties on various Bills and legislative drafts, before they are presented to Cabinet, Parliament and the President for approval, enactment and assent, respectively. FDKP has made several submissions to (and held various engagements with) CIC, relating to the Urban Areas and Cities Act, the Inter-governmental Relations Act and the Transition to Devolved Government Act, among others. In addition, FDKP made detailed representations to CIC regarding the criteria for the distribution of the constitutionally-prescribed 1,450 wards across the forty-seven counties The programme’s engagements can be categorized broadly along the following lines: Capacity Building: ▪ Organization and facilitation of a peer-to-peer learning visit by members of the Commission on Revenue Allocation to the Fiscal Finance Commission in South Africa ▪ Organization and facilitation of a peer-to-peer learning visit by Uganda’s Local Government Finance Commission (LGFC) to Nairobi ▪ Facilitation of learning for individual CRA commissioners at Georgia State University on Fiscal Decentralization ▪ Participation / presentation in a workshop on urban governance and service delivery under devolution hosted by the Ministry of Local Government ▪ Participation / presentation in a national workshop on ‘Alignment of Water Act 2002 and Water Sector with the Constitution of Kenya’ Technical Assistance: ▪ Regular engagement with CRA around their mandate to recommend the equitable share and the horizontal formula―with brief policy notes and presentations (options for urban grants, civil service transitions, and simulations of horizontal sharing formulas). ▪ In collaboration with the Bank’s health sector team, building a comprehensive and geographically disaggregated dataset on health costs and conducting related consultations with the Health sector ministries―the Ministry of Medical Services, and the Ministry of Public Health―to shed light on how to manage the transition from the existing distribution of costs to a more equitable one. The Program has been requested to continue providing technical support using the health sector as a pilot for the preparation of implementation plans (covering function assignment, staff transitions etc). Further consultations are also being planned with ministries from other highly devolved sectors―Agriculture, and Water and Irrigation. 229 Annex 1: Overview of FDKP ▪ Participation as standing observer members to the Ministry of Finance Committee on Fiscal Decentralization. ▪ Participation in the retreat held by the Task Force on Devolved Government to prepare their interim report. ▪ Various presentations to Parliamentary working groups on main devolution challenges (PFM, Civil Service transition, Intergovernmental Coordination). ▪ Facilitation of donor coordination around devolution by convening regular meetings of an Interim Donor Steering Forum to synchronize support to the GoK on devolution. ▪ Comments on draft devolution legislation provided to the Ministry of Finance (PFM Bill), Ministry of Local Government (UAC Act, County Governments Bill) and Commission on Implementation of the Constitution (all major devolution legislation) as well as public consultation fora. Analytical and statistical support ▪ Developed and maintained the first county level subnational database (including data on revenues, expenditure, poverty, demography and service delivery). ▪ Building on the county level database, support to the CRA in designing the Kenya County Factsheets― published by the CRA with World Bank support and widely distributed at national and local level. ▪ Timely and in-depth analysis of devolution policy challenges in the Special Focus section of the 4th edition of the Kenya Economic Update (launched at a high level workshop including leading Kenya policy-makers and civil society representatives). ▪ Regular contributions to the national debate on devolution through media articles on key issues, including a eight page insert in the East African newspaper with regional EAC coverage in March 2012. ▪ Collaborating with the broader PREM team in Nairobi on the preparation of a series of district public expenditure reviews that will explore inter-regional inequity and inefficiency in the provision of government services in the past. ▪ Initiated, in conjunction with the Ministry of Local Government, the CRA and the Association of Local Governments of Kenya (ALGAK), a series of county case studies covering four counties. The case studies are intended to identify major transition issues at the local level in the areas of financial management, ervice delivery and social accountability. ▪ Within the Bank, raised awareness of sector teams (Urban, Water, Health) on the major implications of devolution for policy and projects in sectors. The ESW report reflects lessons accumulated through FDKP’s close engagement with key stakeholders in the devolution implementation process, backed-up by qualitative and quantitative analysis from different World Bank teams as well as contributions from leading local and international experts on design and implementation of fiscal decentralization. 230 Produced by Poverty Reduction and Economic Management Unit Africa Region Photo credit © Tobin Jones - Demotix Design by Robert Waiharo Second print run | November 2012